.codex/skills/release-workflow/SKILL.md

@@ -0,0 +1,78 @@
+---
+name: release-workflow
+description: Use when the user asks to release, bump version, update changelog/version files, or commit/push a repository release for the Planet repo. Applies the repo's versioning rules, updates all required version-bearing files, updates changelog/version-history, runs minimal relevant validation, and then commits/pushes when requested.
+---
+
+# Release Workflow
+
+Use this skill for release-oriented work in this repository.
+
+## When To Use
+
+- The user asks to `发版`
+- The user asks to bump a version
+- The user asks to update `CHANGELOG`, `version-history`, or version files as part of a release
+- The user asks to commit/push a release or a publishable bugfix/feature bundle
+
+Do not use this skill for ordinary commits that are not being released.
+
+## Versioning Rules
+
+- `feature` -> bump `+0.1.0`
+- `bugfix` -> bump `+0.0.1`
+- `docs`, `maintenance`, and `refactor` do not bump by default unless the user explicitly wants a release
+
+When intent is mixed, prefer the user’s stated release intent. If they ask to release a bugfix bundle, use a patch bump.
+
+## Required Files
+
+Every release bump must update these files together:
+
+- `/home/ray/dev/linkong/planet/VERSION`
+- `/home/ray/dev/linkong/planet/frontend/package.json`
+- `/home/ray/dev/linkong/planet/pyproject.toml`
+- `/home/ray/dev/linkong/planet/uv.lock`
+- `/home/ray/dev/linkong/planet/docs/CHANGELOG.md`
+- `/home/ray/dev/linkong/planet/docs/version-history.md`
+
+## Workflow
+
+1. Inspect the current worktree and current version.
+2. Decide the release type from the user request:
+   - feature
+   - bugfix
+   - release without code changes
+3. Compute the next version.
+4. Update all required version-bearing files.
+5. Add a concise but specific changelog entry:
+   - highlights
+   - important added/improved/fixed items
+   - mention the highest-signal files only
+6. Update `docs/version-history.md`:
+   - current dev version
+   - new timeline row with summary
+7. Run the smallest relevant validation available.
+8. Before commit, verify the target version is present in all required files.
+9. If the user asked for commit/push:
+   - stage the release files and code changes
+   - commit with a conventional message
+   - push to the requested branch, usually `dev`
+
+## Validation Guidance
+
+- Prefer scope-matched validation over broad expensive checks
+- Typical examples:
+  - Python backend edits: `python3 -m py_compile ...`
+  - Frontend edits: use the project-standard frontend build/check if available
+- If the environment prevents a check, say that explicitly in the final summary
+
+## Release Checklist
+
+Before closing the task, confirm:
+
+- version bump applied consistently
+- changelog updated
+- version history updated
+- generated/runtime artifacts are not accidentally staged
+- validation status recorded
+- commit and push completed if requested

.gitignore

@@ -145,3 +145,8 @@ docs/.venv/
 *.temp
 tmp/
 temp/
+
+# ----------------------
+# Runtime Data
+# ----------------------
+data/ai/bgp-briefs/

README.md

@@ -102,6 +102,13 @@
 | Axios | HTTP 客户端 |
 | Socket.io-client | WebSocket 客户端 |
 | ECharts | 统计图表 |
+| Bun | 前端包管理与脚本运行 |
+
+前端工程统一使用 Bun：
+
+- 安装依赖使用 `bun install`
+- 运行脚本使用 `bun run <script>`
+- 不使用 `npm`、`pnpm`、`yarn`
 
 ### 虚幻引擎客户端
 
@@ -205,10 +212,47 @@
 ./planet.sh health
 ```
 
+前端命令约定：
+
+```bash
+cd frontend
+bun install
+bun run dev
+bun run build
+```
+
+不要使用 `npm run ...`，避免在 WSL/Windows 混合环境里触发 `cmd.exe` 路径兼容问题。
+
 ## API 文档
 
 启动服务后访问: `http://localhost:8000/docs`
 
+## 启动容错参数
+
+`planet.sh` 现在为依赖安装、数据库、AI Provider 启动加入了有限次重试，并会在数据库与 `aiprovider` 启动后额外等待 Docker healthcheck。
+
+可通过环境变量临时调整:
+
+```bash
+# 例: 放宽 AI Provider 与数据库在网络抖动下的自愈次数
+AI_PROVIDER_START_MAX_RETRIES=5 \
+AI_PROVIDER_RETRY_INTERVAL=10 \
+DATABASE_START_MAX_RETRIES=5 \
+DATABASE_RETRY_INTERVAL=10 \
+./planet.sh restart
+```
+
+常用参数:
+
+- `DEPENDENCY_INSTALL_MAX_RETRIES` / `DEPENDENCY_INSTALL_RETRY_INTERVAL`: 控制 `uv sync`、`bun install` 的重试次数与间隔，默认 `3` 次、`5` 秒
+- `DATABASE_START_MAX_RETRIES` / `DATABASE_RETRY_INTERVAL`: 控制 `postgres`、`redis` 的启动/重启与健康检查自愈，默认 `3` 次、`5` 秒
+- `AI_PROVIDER_START_MAX_RETRIES` / `AI_PROVIDER_RETRY_INTERVAL`: 控制 `aiprovider` 的构建/启动与容器重启自愈，默认 `3` 次、`5` 秒
+- `BACKEND_MAX_RETRIES`: 控制后端进程启动重试次数，默认 `3`
+- `FRONTEND_MAX_RETRIES`: 控制前端 dev server 启动重试次数，默认 `3`
+- `BACKEND_HEALTH_CHECK_ATTEMPTS` / `BACKEND_HEALTH_CHECK_INTERVAL`: 控制后端 HTTP 健康检查等待次数与间隔，默认 `10` 次、`2` 秒
+- `FRONTEND_HEALTH_CHECK_ATTEMPTS` / `FRONTEND_HEALTH_CHECK_INTERVAL`: 控制前端 HTTP 可访问检查等待次数与间隔，默认 `10` 次、`2` 秒
+- `AI_PROVIDER_HEALTH_CHECK_ATTEMPTS` / `AI_PROVIDER_HEALTH_CHECK_INTERVAL`: 控制 `aiprovider` HTTP 健康检查等待次数与间隔，默认 `10` 次、`2` 秒
+
 ## AI 接口预留
 
 项目现在采用“两层”设计:
@@ -286,6 +330,23 @@ AI_PROVIDER_SERVICE_TOKEN=change_me
 
 - [docs/aiprovider.md](/home/ray/dev/linkong/planet/docs/aiprovider.md)
 - [aiprovider/README.md](/home/ray/dev/linkong/planet/aiprovider/README.md)
+- [docs/frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/frontend-layout-guidelines.md)
+- [docs/ai-playground-development-plan.md](/home/ray/dev/linkong/planet/docs/ai-playground-development-plan.md)
+- [docs/situational-awareness-foundation-plan.md](/home/ray/dev/linkong/planet/docs/situational-awareness-foundation-plan.md)
+
+## 前端页面布局规范
+
+管理后台页面默认遵循“单屏工作区”原则：
+
+- 页头、摘要区、主工作区应在一屏内形成稳定结构
+- 主表格 / 主图表 / 主分析区应占据页面主要可视空间
+- 模块内容超出时优先在卡片、表格、标签页内部滚动
+- 不依赖整页纵向撑开来容纳主要工作区
+
+当前推荐参考实现：
+
+- [frontend/src/pages/BGP/BGP.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/BGP/BGP.tsx)
+- [docs/frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/frontend-layout-guidelines.md)
 
 ## License
 

TODO.md

@@ -13,6 +13,9 @@
 - [x] 接入 `IPtoASN / IPtoCountry` 作为 prefix-centric geography 的主数据源
 - [x] 接入 `OpenGeoFeed` 作为 prefix geography 的高质量覆盖/override 数据源
 - [x] 把 RIR delegated 设计成 prefix geography 的 fallback，而不是主来源
+- [ ] 为 `aiprovider` 建立 `provider -> api adapter -> compat policy` 的配置中心，优先落成 `json` 或 `yaml` 文件，运行时按 `provider/model` 读取兼容设置，而不是把专项兼容继续散落在 Python 分支里
+- [ ] 为市面上主流 AI 服务补专项兼容配置并固化到配置文件中，至少覆盖 `OpenAI / Anthropic / MiniMax / Ollama / Moonshot / DeepSeek / Qwen / GLM / Gemini / OpenRouter / vLLM / LM Studio / One API`
+- [ ] 在兼容配置中补齐可声明项：`api adapter`、`base_url pattern`、`auth header`、`thinking default`、`reasoning block mapping`、`stream path`、`tool-call capability`、`multimodal capability`、`provider-specific request patch`
 - [ ] 接入 `inetnum` / `inet6num` whois 作为比 RIR 更细粒度的后备层
 - [x] 在 activity layer 之后继续补 `route leak` 和 `path instability / flap` detector
 - [ ] 对 [frontend/public/earth/js/bgp.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/bgp.js) 做按职责拆分的小重构，拆成 data / markers / overlays / animation，降低后续维护复杂度

VERSION

@@ -1 +1 @@
-0.23.0
+0.27.0

aiprovider/.env.example

@@ -6,29 +6,51 @@ AI_TIMEOUT_SECONDS=60
 AI_HTTP_RETRY_ATTEMPTS=2
 AI_ANALYSIS_SYSTEM_PROMPT=你是态势感知分析助手。请基于输入的上下文、观测与约束，输出结构化、克制、可执行的分析。
 
-# Select one provider mode:
-# - openai_compatible
-# - claude_compatible
+# Provider identity. Recommended values:
+# - minimax
+# - openai
 # - ollama
-AI_PROVIDER=ollama
+# Compatibility aliases still accepted:
+# - openai_compatible
+# - anthropic_compatible
+# - claude_compatible
+AI_PROVIDER=minimax
+
+# Request adapter style, following OpenClaw's API-seam pattern:
+# - auto
+# - openai-completions
+# - anthropic-messages
+# - ollama-generate
+AI_PROVIDER_API=anthropic-messages
 
 # Common model selection
-AI_MODEL=qwen2.5:7b
+AI_MODEL=MiniMax-M2.7
+
+# MiniMax CN Anthropic-compatible example
+AI_BASE_URL=https://api.minimaxi.com/anthropic
+AI_API_KEY=sk-cp-change-me
+AI_MAX_TOKENS=1200
+AI_ANTHROPIC_VERSION=2023-06-01
 
 # OpenAI-compatible example (vLLM / LM Studio / One API / local gateway)
-# AI_PROVIDER=openai_compatible
+# AI_PROVIDER=openai
+# AI_PROVIDER_API=openai-completions
 # AI_BASE_URL=http://127.0.0.1:8001/v1
 # AI_API_KEY=local-key
+# AI_MODEL=your-local-model
 
-# Claude-compatible example (Anthropic / MiniMax / Claude-compatible gateway)
-# AI_PROVIDER=claude_compatible
-# AI_BASE_URL=http://127.0.0.1:8002
+# Anthropic-compatible example (Claude-compatible gateway)
+# AI_PROVIDER=anthropic
+# AI_PROVIDER_API=anthropic-messages
+# AI_BASE_URL=http://127.0.0.1:8002/anthropic
 # AI_API_KEY=local-key
+# AI_MODEL=your-model
 # AI_MAX_TOKENS=1200
 # AI_ANTHROPIC_VERSION=2023-06-01
 
 # Ollama native example
-AI_BASE_URL=http://127.0.0.1:11434
-AI_API_KEY=
-AI_MAX_TOKENS=1200
-AI_ANTHROPIC_VERSION=2023-06-01
+# AI_PROVIDER=ollama
+# AI_PROVIDER_API=ollama-generate
+# AI_BASE_URL=http://127.0.0.1:11434
+# AI_API_KEY=
+# AI_MODEL=qwen2.5:7b

aiprovider/README.md

@@ -8,17 +8,27 @@
 
 当前支持:
 
-- `AI_PROVIDER=openai`
-- `AI_PROVIDER=openai_compatible`
-- `AI_PROVIDER=anthropic`
-- `AI_PROVIDER=anthropic_compatible`
-- `AI_PROVIDER=claude_compatible`
-- `AI_PROVIDER=ollama`
+- provider identity:
+  - `AI_PROVIDER=openai`
+  - `AI_PROVIDER=anthropic`
+  - `AI_PROVIDER=minimax`
+  - `AI_PROVIDER=ollama`
+- request adapter:
+  - `AI_PROVIDER_API=openai-completions`
+  - `AI_PROVIDER_API=anthropic-messages`
+  - `AI_PROVIDER_API=ollama-generate`
+
+兼容别名仍然保留:
+
+- `openai_compatible`
+- `anthropic_compatible`
+- `claude_compatible`
 
 典型配置:
 
 ```env
-AI_PROVIDER=openai_compatible
+AI_PROVIDER=openai
+AI_PROVIDER_API=openai-completions
 AI_BASE_URL=https://api.openai.com/v1
 AI_API_KEY=your_api_key
 AI_MODEL=gpt-4o-mini
@@ -26,13 +36,14 @@ AI_TIMEOUT_SECONDS=60
 AI_PROVIDER_SERVICE_TOKEN=change_me
 ```
 
-Claude 兼容供应商示例:
+MiniMax 中国大陆节点示例:
 
 ```env
-AI_PROVIDER=claude_compatible
-AI_BASE_URL=https://your-claude-compatible-endpoint.example.com
-AI_API_KEY=your_api_key
-AI_MODEL=your-claude-compatible-model
+AI_PROVIDER=minimax
+AI_PROVIDER_API=anthropic-messages
+AI_BASE_URL=https://api.minimaxi.com/anthropic
+AI_API_KEY=sk-cp-xxxxx
+AI_MODEL=MiniMax-M2.7
 AI_TIMEOUT_SECONDS=60
 AI_MAX_TOKENS=1200
 AI_ANTHROPIC_VERSION=2023-06-01
@@ -43,12 +54,15 @@ AI_PROVIDER_SERVICE_TOKEN=change_me
 
 - Anthropic 官方 Claude API
 - Claude 兼容网关
-- MiniMax 等提供 Claude/Anthropic 风格消息接口的服务
+- MiniMax 等提供 Anthropic Messages 风格接口的服务
+
+这套命名方式参考了 OpenClaw 的接入模式: provider 负责标识供应商, `AI_PROVIDER_API` 负责标识协议适配层, 避免把“供应商”和“协议”绑死在一起。
 
 Ollama 原生示例:
 
 ```env
 AI_PROVIDER=ollama
+AI_PROVIDER_API=ollama-generate
 AI_BASE_URL=http://127.0.0.1:11434
 AI_API_KEY=
 AI_MODEL=qwen2.5:7b
@@ -58,8 +72,8 @@ AI_PROVIDER_SERVICE_TOKEN=change_me
 
 本地模型接入建议:
 
-- `vLLM`、`LM Studio`、`One API`：优先使用 `openai_compatible`
-- `MiniMax`、Claude 兼容网关：使用 `claude_compatible`
+- `vLLM`、`LM Studio`、`One API`：`AI_PROVIDER=openai` + `AI_PROVIDER_API=openai-completions`
+- `MiniMax`、Claude 兼容网关：`AI_PROVIDER=minimax|anthropic` + `AI_PROVIDER_API=anthropic-messages`
 - `Ollama`：可直接使用 `ollama`
 
 启动模板:

aiprovider/config.py

@@ -9,6 +9,7 @@ class Settings(BaseSettings):
     SERVICE_VERSION: str = "0.1.0"
 
     AI_PROVIDER: str = "disabled"
+    AI_PROVIDER_API: str = "auto"
     AI_BASE_URL: str = "https://api.openai.com/v1"
     AI_API_KEY: str = ""
     AI_MODEL: str = ""

aiprovider/provider_service.py

@@ -8,6 +8,7 @@ from fastapi import HTTPException, status
 
 from aiprovider.config import settings
 from aiprovider.schemas import (
+    AIContentBlock,
     AIProviderStatusResponse,
     SituationalAnalysisRequest,
     SituationalAnalysisResponse,
@@ -18,9 +19,39 @@ def _normalize_provider(value: str) -> str:
     return (value or "disabled").strip().lower()
 
 
+def _normalize_provider_api(value: str) -> str:
+    return (value or "auto").strip().lower().replace("_", "-")
+
+
+def _resolve_provider_api(provider: str, configured_api: str) -> str:
+    if configured_api and configured_api != "auto":
+        return configured_api
+
+    if provider in {"openai", "openai-compatible", "openai_compatible"}:
+        return "openai-completions"
+    if provider in {
+        "anthropic",
+        "anthropic-compatible",
+        "anthropic_compatible",
+        "claude-compatible",
+        "claude_compatible",
+        "minimax",
+        "kimi-coding",
+        "moonshot-anthropic",
+    }:
+        return "anthropic-messages"
+    if provider == "ollama":
+        return "ollama-generate"
+    return "disabled"
+
+
 class ProviderService:
     def __init__(self) -> None:
         self.provider = _normalize_provider(settings.AI_PROVIDER)
+        self.provider_api = _resolve_provider_api(
+            self.provider,
+            _normalize_provider_api(settings.AI_PROVIDER_API),
+        )
         self.base_url = settings.AI_BASE_URL.rstrip("/")
         self.api_key = settings.AI_API_KEY
         self.default_model = settings.AI_MODEL
@@ -32,9 +63,11 @@ class ProviderService:
 
     def get_status(self) -> AIProviderStatusResponse:
         enabled = self.provider != "disabled"
-        configured = enabled and bool(self.base_url and self.api_key and self.default_model)
+        has_credentials = bool(self.api_key) if self._requires_api_key() else True
+        configured = enabled and bool(self.base_url and has_credentials and self.default_model)
         return AIProviderStatusResponse(
             provider=self.provider,
+            api=self.provider_api if enabled else None,
             enabled=enabled,
             configured=configured,
             model=self.default_model or None,
@@ -49,7 +82,8 @@ class ProviderService:
             )
 
         model = payload.preferred_model or self.default_model
-        if not self.base_url or not self.api_key or not model:
+        has_credentials = bool(self.api_key) if self._requires_api_key() else True
+        if not self.base_url or not has_credentials or not model:
             raise HTTPException(
                 status_code=status.HTTP_503_SERVICE_UNAVAILABLE,
                 detail="AI provider is not fully configured. Check AI_BASE_URL, AI_API_KEY, and AI_MODEL.",
@@ -57,28 +91,40 @@ class ProviderService:
 
         prompt = self._build_prompt(payload)
 
-        if self.provider in {"openai", "openai_compatible"}:
+        if self.provider_api == "openai-completions":
             data = await self._request_openai_compatible(model, prompt)
             content = self._extract_openai_content(data)
-        elif self.provider in {"anthropic", "anthropic_compatible", "claude_compatible"}:
-            data = await self._request_anthropic_compatible(model, prompt)
+            content_blocks = self._extract_openai_blocks(data)
+        elif self.provider_api == "anthropic-messages":
+            data = await self._request_anthropic_messages(model, prompt, payload.thinking)
             content = self._extract_anthropic_content(data)
-        elif self.provider == "ollama":
+            content_blocks = self._extract_anthropic_blocks(data)
+        elif self.provider_api == "ollama-generate":
             data = await self._request_ollama(model, prompt)
             content = self._extract_ollama_content(data)
+            content_blocks = self._extract_ollama_blocks(data)
         else:
             raise HTTPException(
                 status_code=status.HTTP_400_BAD_REQUEST,
-                detail=f"Unsupported AI provider: {self.provider}",
+                detail=f"Unsupported AI provider API: {self.provider_api}",
             )
 
+        text_blocks = [block.text for block in content_blocks if block.text]
+        thinking_blocks = [block.thinking for block in content_blocks if block.thinking]
+
         return SituationalAnalysisResponse(
             provider=self.provider,
             model=model,
             content=content,
+            content_blocks=content_blocks,
+            text_blocks=text_blocks,
+            thinking_blocks=thinking_blocks,
             raw_response=data,
         )
 
+    def _requires_api_key(self) -> bool:
+        return self.provider_api != "ollama-generate"
+
     def _build_prompt(self, payload: SituationalAnalysisRequest) -> str:
         sections = [
             f"任务标题:\n{payload.title}",
@@ -113,7 +159,12 @@ class ProviderService:
             request_body=request_body,
         )
 
-    async def _request_anthropic_compatible(self, model: str, prompt: str) -> dict[str, Any]:
+    async def _request_anthropic_messages(
+        self,
+        model: str,
+        prompt: str,
+        thinking: dict[str, Any] | None = None,
+    ) -> dict[str, Any]:
         request_body = {
             "model": model,
             "system": self.system_prompt,
@@ -131,8 +182,15 @@ class ProviderService:
             "max_tokens": self.max_tokens,
             "temperature": 0.2,
         }
+        resolved_thinking = self._resolve_anthropic_thinking(thinking)
+        if resolved_thinking:
+            request_body["thinking"] = resolved_thinking
+        if self.provider == "minimax" and self.base_url.endswith("/anthropic"):
+            path = "/v1/messages"
+        else:
+            path = "/messages"
         return await self._post(
-            path="/messages",
+            path=path,
             headers={
                 "x-api-key": self.api_key,
                 "anthropic-version": self.anthropic_version,
@@ -141,6 +199,25 @@ class ProviderService:
             request_body=request_body,
         )
 
+    def _resolve_anthropic_thinking(self, thinking: dict[str, Any] | None) -> dict[str, Any] | None:
+        if thinking:
+            return thinking
+
+        # OpenClaw treats MiniMax's Anthropic-compatible path specially:
+        # disable thinking by default unless the caller explicitly opts in.
+        if self.provider == "minimax":
+            return {"type": "disabled"}
+
+        return None
+
+    async def _request_anthropic_compatible(
+        self,
+        model: str,
+        prompt: str,
+        thinking: dict[str, Any] | None = None,
+    ) -> dict[str, Any]:
+        return await self._request_anthropic_messages(model, prompt, thinking)
+
     async def _request_ollama(self, model: str, prompt: str) -> dict[str, Any]:
         request_body = {
             "model": model,
@@ -218,6 +295,30 @@ class ProviderService:
             )
         return ""
 
+    def _extract_openai_blocks(self, payload: dict[str, Any]) -> list[AIContentBlock]:
+        choices = payload.get("choices") or []
+        if not choices:
+            return []
+
+        message = choices[0].get("message") or {}
+        content = message.get("content")
+        if isinstance(content, str):
+            return [AIContentBlock(type="text", text=content)]
+        if not isinstance(content, list):
+            return []
+
+        blocks: list[AIContentBlock] = []
+        for item in content:
+            if not isinstance(item, dict):
+                continue
+            blocks.append(
+                AIContentBlock(
+                    type=str(item.get("type", "text")),
+                    text=item.get("text") if isinstance(item.get("text"), str) else None,
+                    metadata={k: v for k, v in item.items() if k not in {"type", "text"}},
+                )
+            )
+        return blocks
     def _extract_anthropic_content(self, payload: dict[str, Any]) -> str:
         content = payload.get("content")
         if isinstance(content, str):
@@ -233,8 +334,39 @@ class ProviderService:
                 fragments.append(item["text"])
         return "".join(fragments)
 
+    def _extract_anthropic_blocks(self, payload: dict[str, Any]) -> list[AIContentBlock]:
+        content = payload.get("content")
+        if not isinstance(content, list):
+            return []
+
+        blocks: list[AIContentBlock] = []
+        for item in content:
+            if not isinstance(item, dict):
+                continue
+            blocks.append(
+                AIContentBlock(
+                    type=str(item.get("type", "unknown")),
+                    text=item.get("text") if isinstance(item.get("text"), str) else None,
+                    thinking=item.get("thinking") if isinstance(item.get("thinking"), str) else None,
+                    signature=item.get("signature") if isinstance(item.get("signature"), str) else None,
+                    metadata={
+                        k: v
+                        for k, v in item.items()
+                        if k not in {"type", "text", "thinking", "signature"}
+                    },
+                )
+            )
+        return blocks
+
+
     def _extract_ollama_content(self, payload: dict[str, Any]) -> str:
         response = payload.get("response")
         if isinstance(response, str):
             return response
         return ""
+
+    def _extract_ollama_blocks(self, payload: dict[str, Any]) -> list[AIContentBlock]:
+        response = payload.get("response")
+        if isinstance(response, str) and response:
+            return [AIContentBlock(type="text", text=response)]
+        return []

aiprovider/schemas.py

@@ -3,6 +3,14 @@ from typing import Any
 from pydantic import BaseModel, Field
 
 
+class AIContentBlock(BaseModel):
+    type: str
+    text: str | None = None
+    thinking: str | None = None
+    signature: str | None = None
+    metadata: dict[str, Any] = Field(default_factory=dict)
+
+
 class SituationalAnalysisRequest(BaseModel):
     title: str = Field(..., min_length=1, max_length=200)
     objective: str = Field(..., min_length=1, max_length=1000)
@@ -10,17 +18,22 @@ class SituationalAnalysisRequest(BaseModel):
     observations: list[str] = Field(default_factory=list)
     constraints: list[str] = Field(default_factory=list)
     preferred_model: str | None = Field(default=None, max_length=200)
+    thinking: dict[str, Any] | None = None
 
 
 class SituationalAnalysisResponse(BaseModel):
     provider: str
     model: str
     content: str
+    content_blocks: list[AIContentBlock] = Field(default_factory=list)
+    text_blocks: list[str] = Field(default_factory=list)
+    thinking_blocks: list[str] = Field(default_factory=list)
     raw_response: dict[str, Any] = Field(default_factory=dict)
 
 
 class AIProviderStatusResponse(BaseModel):
     provider: str
+    api: str | None = None
     enabled: bool
     configured: bool
     model: str | None = None

backend/app/api/main.py

@@ -14,6 +14,7 @@ from app.api.v1 import (
     visualization,
     bgp,
     system_control,
+    tv,
 )
 
 api_router = APIRouter()
@@ -33,3 +34,4 @@ api_router.include_router(settings.router, prefix="/settings", tags=["settings"]
 api_router.include_router(system_control.router, prefix="/system", tags=["system"])
 api_router.include_router(visualization.router, prefix="/visualization", tags=["visualization"])
 api_router.include_router(bgp.router, prefix="/bgp", tags=["bgp"])
+api_router.include_router(tv.router, prefix="/tv", tags=["tv"])

backend/app/api/v1/ai.py

@@ -1,15 +1,52 @@
 from uuid import uuid4
 
-from fastapi import APIRouter, Depends, Request, Response
+from fastapi import APIRouter, Depends, HTTPException, Request, Response
+from sqlalchemy.ext.asyncio import AsyncSession
 
 from app.core.security import get_current_user
+from app.db.session import get_db
 from app.models.user import User
 from app.schemas.ai import (
     AIProviderStatusResponse,
+    AlertBriefRequest,
+    AlertBriefResponse,
+    BGPBriefRequest,
+    BGPBriefRecordResponse,
+    BGPBriefRecordSummary,
+    PlaygroundMessageActionResponse,
+    PlaygroundMessageCreateRequest,
+    PlaygroundMessageEditRequest,
+    PlaygroundMessageResendRequest,
+    PlaygroundMessageStopRequest,
+    PlaygroundSessionResponse,
+    PlaygroundSessionUpsertRequest,
+    PlaygroundThreadResponse,
+    SituationalAlertBriefRequest,
+    SituationalAlertBriefResponse,
     SituationalAnalysisRequest,
     SituationalAnalysisResponse,
 )
+from app.services.alert_ai_brief import build_alert_brief_request
 from app.services.ai_client import AIProviderClient, get_ai_provider_client
+from app.services.bgp_ai_brief import build_bgp_brief_request
+from app.services.bgp_ai_brief_store import (
+    get_bgp_brief_record,
+    get_latest_bgp_brief_record,
+    list_bgp_brief_records,
+    save_bgp_brief_record,
+)
+from app.services.playground_session_store import (
+    get_playground_session,
+    upsert_playground_session,
+)
+from app.services.playground_chat_service import (
+    create_turn,
+    edit_user_message,
+    get_thread,
+    resend_turn,
+    stop_message,
+)
+from app.services.situational_alert_ai_brief import build_situational_alert_brief_request
 
 router = APIRouter()
 
@@ -37,3 +74,208 @@ async def analyze_situational_awareness(
     request_id = request.headers.get("X-Request-ID") or str(uuid4())
     response.headers["X-Request-ID"] = request_id
     return await provider_client.analyze(payload, request_id=request_id)
+
+
+@router.get("/playground/thread", response_model=PlaygroundThreadResponse | None)
+async def get_playground_thread(
+    session_key: str = "default",
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    return await get_thread(
+        db,
+        user_id=current_user.id,
+        session_key=session_key,
+    )
+
+
+@router.get("/playground/session", response_model=PlaygroundSessionResponse | None)
+async def get_saved_playground_session(
+    session_key: str = "default",
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    return await get_playground_session(
+        db,
+        user_id=current_user.id,
+        session_key=session_key,
+    )
+
+
+@router.put("/playground/session", response_model=PlaygroundSessionResponse)
+async def save_playground_session(
+    payload: PlaygroundSessionUpsertRequest,
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    return await upsert_playground_session(
+        db,
+        user_id=current_user.id,
+        payload=payload,
+    )
+
+
+@router.post("/playground/messages", response_model=PlaygroundMessageActionResponse)
+async def create_playground_message(
+    payload: PlaygroundMessageCreateRequest,
+    current_user: User = Depends(get_current_user),
+    provider_client: AIProviderClient = Depends(get_ai_provider_client),
+    db: AsyncSession = Depends(get_db),
+):
+    return await create_turn(
+        db,
+        user_id=current_user.id,
+        payload=payload,
+        provider_client=provider_client,
+    )
+
+
+@router.post("/playground/messages/stop", response_model=PlaygroundMessageActionResponse)
+async def stop_playground_message(
+    payload: PlaygroundMessageStopRequest,
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    return await stop_message(
+        db,
+        user_id=current_user.id,
+        payload=payload,
+    )
+
+
+@router.post("/playground/messages/resend", response_model=PlaygroundMessageActionResponse)
+async def resend_playground_message(
+    payload: PlaygroundMessageResendRequest,
+    current_user: User = Depends(get_current_user),
+    provider_client: AIProviderClient = Depends(get_ai_provider_client),
+    db: AsyncSession = Depends(get_db),
+):
+    return await resend_turn(
+        db,
+        user_id=current_user.id,
+        payload=payload,
+        provider_client=provider_client,
+    )
+
+
+@router.post("/playground/messages/edit", response_model=PlaygroundMessageActionResponse)
+async def edit_playground_message(
+    payload: PlaygroundMessageEditRequest,
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    return await edit_user_message(
+        db,
+        user_id=current_user.id,
+        payload=payload,
+    )
+
+
+@router.get("/bgp/briefs", response_model=list[BGPBriefRecordSummary])
+async def list_saved_bgp_briefs(
+    current_user: User = Depends(get_current_user),
+):
+    return list_bgp_brief_records()
+
+
+@router.get("/bgp/briefs/latest", response_model=BGPBriefRecordResponse | None)
+async def get_latest_saved_bgp_brief(
+    current_user: User = Depends(get_current_user),
+):
+    return get_latest_bgp_brief_record()
+
+
+@router.get("/bgp/briefs/{brief_id}", response_model=BGPBriefRecordResponse)
+async def get_saved_bgp_brief(
+    brief_id: str,
+    current_user: User = Depends(get_current_user),
+):
+    record = get_bgp_brief_record(brief_id)
+    if record is None:
+        raise HTTPException(status_code=404, detail="BGP brief not found")
+    return record
+
+
+@router.post("/bgp/brief", response_model=BGPBriefRecordResponse)
+async def analyze_bgp_brief(
+    payload: BGPBriefRequest,
+    request: Request,
+    response: Response,
+    current_user: User = Depends(get_current_user),
+    provider_client: AIProviderClient = Depends(get_ai_provider_client),
+    db: AsyncSession = Depends(get_db),
+):
+    request_id = request.headers.get("X-Request-ID") or str(uuid4())
+    response.headers["X-Request-ID"] = request_id
+
+    brief_request, facts, context = await build_bgp_brief_request(
+        db,
+        incident_limit=payload.incident_limit,
+        anomaly_limit=payload.anomaly_limit,
+        collector_limit=payload.collector_limit,
+    )
+    brief_request.preferred_model = payload.preferred_model
+    brief_request.thinking = payload.thinking
+
+    analysis = await provider_client.analyze(brief_request, request_id=request_id)
+    return save_bgp_brief_record(
+        analysis,
+        request_id=request_id,
+        facts=facts,
+        context=context,
+    )
+
+
+@router.post("/alerts/brief", response_model=AlertBriefResponse)
+async def analyze_alert_brief(
+    payload: AlertBriefRequest,
+    request: Request,
+    response: Response,
+    current_user: User = Depends(get_current_user),
+    provider_client: AIProviderClient = Depends(get_ai_provider_client),
+    db: AsyncSession = Depends(get_db),
+):
+    request_id = request.headers.get("X-Request-ID") or str(uuid4())
+    response.headers["X-Request-ID"] = request_id
+
+    brief_request, facts, context = await build_alert_brief_request(
+        db,
+        alert_limit=payload.alert_limit,
+    )
+    brief_request.preferred_model = payload.preferred_model
+    brief_request.thinking = payload.thinking
+
+    analysis = await provider_client.analyze(brief_request, request_id=request_id)
+    return AlertBriefResponse(
+        **analysis.model_dump(),
+        title=brief_request.title,
+        objective=brief_request.objective,
+        facts=facts,
+        context=context,
+    )
+
+
+@router.post("/situational-alerts/brief", response_model=SituationalAlertBriefResponse)
+async def analyze_situational_alert_brief(
+    payload: SituationalAlertBriefRequest,
+    request: Request,
+    response: Response,
+    current_user: User = Depends(get_current_user),
+    provider_client: AIProviderClient = Depends(get_ai_provider_client),
+    db: AsyncSession = Depends(get_db),
+):
+    request_id = request.headers.get("X-Request-ID") or str(uuid4())
+    response.headers["X-Request-ID"] = request_id
+
+    brief_request, facts, context = await build_situational_alert_brief_request(db)
+    brief_request.preferred_model = payload.preferred_model
+    brief_request.thinking = payload.thinking
+
+    analysis = await provider_client.analyze(brief_request, request_id=request_id)
+    return SituationalAlertBriefResponse(
+        **analysis.model_dump(),
+        title=brief_request.title,
+        objective=brief_request.objective,
+        facts=facts,
+        context=context,
+    )

backend/app/api/v1/alerts.py

@@ -1,7 +1,7 @@
 from datetime import UTC, datetime
 from typing import Optional
 
-from fastapi import APIRouter, Depends
+from fastapi import APIRouter, Depends, HTTPException
 from sqlalchemy import select, func, case
 from sqlalchemy.ext.asyncio import AsyncSession
 
@@ -9,6 +9,7 @@ from app.db.session import get_db
 from app.models.user import User
 from app.core.security import get_current_user
 from app.models.alert import Alert, AlertSeverity, AlertStatus
+from app.schemas.alert import AlertResolutionRequest
 
 
 router = APIRouter()
@@ -77,7 +78,7 @@ async def acknowledge_alert(
 @router.post("/{alert_id}/resolve")
 async def resolve_alert(
     alert_id: int,
-    resolution: str,
+    payload: AlertResolutionRequest,
     current_user: User = Depends(get_current_user),
     db: AsyncSession = Depends(get_db),
 ):
@@ -85,12 +86,12 @@ async def resolve_alert(
     alert = result.scalar_one_or_none()
 
     if not alert:
-        return {"error": "Alert not found"}
+        raise HTTPException(status_code=404, detail="Alert not found")
 
     alert.status = AlertStatus.RESOLVED
     alert.resolved_by = current_user.id
     alert.resolved_at = datetime.now(UTC)
-    alert.resolution_notes = resolution
+    alert.resolution_notes = payload.resolution
     await db.commit()
 
     return {"message": "Alert resolved", "alert": alert.to_dict()}
@@ -101,25 +102,44 @@ async def get_alert_stats(
     current_user: User = Depends(get_current_user),
     db: AsyncSession = Depends(get_db),
 ):
-    critical_query = select(func.count(Alert.id)).where(
-        Alert.severity == AlertSeverity.CRITICAL,
-        Alert.status == AlertStatus.ACTIVE,
+    result = await db.execute(
+        select(
+            func.sum(
+                case(
+                    (
+                        (Alert.severity == AlertSeverity.CRITICAL)
+                        & (Alert.status == AlertStatus.ACTIVE),
+                        1,
+                    ),
+                    else_=0,
+                )
+            ).label("critical"),
+            func.sum(
+                case(
+                    (
+                        (Alert.severity == AlertSeverity.WARNING)
+                        & (Alert.status == AlertStatus.ACTIVE),
+                        1,
+                    ),
+                    else_=0,
+                )
+            ).label("warning"),
+            func.sum(
+                case(
+                    (
+                        (Alert.severity == AlertSeverity.INFO)
+                        & (Alert.status == AlertStatus.ACTIVE),
+                        1,
+                    ),
+                    else_=0,
+                )
+            ).label("info"),
+        )
     )
-    warning_query = select(func.count(Alert.id)).where(
-        Alert.severity == AlertSeverity.WARNING,
-        Alert.status == AlertStatus.ACTIVE,
-    )
-    info_query = select(func.count(Alert.id)).where(
-        Alert.severity == AlertSeverity.INFO,
-        Alert.status == AlertStatus.ACTIVE,
-    )
-
-    critical_result = await db.execute(critical_query)
-    warning_result = await db.execute(warning_query)
-    info_result = await db.execute(info_query)
+    row = result.one()
 
     return {
-        "critical": critical_result.scalar() or 0,
-        "warning": warning_result.scalar() or 0,
-        "info": info_result.scalar() or 0,
+        "critical": row.critical or 0,
+        "warning": row.warning or 0,
+        "info": row.info or 0,
     }

backend/app/api/v1/bgp.py

@@ -22,16 +22,161 @@ def _parse_dt(value: Optional[str]) -> Optional[datetime]:
     if not value:
         return None
     return datetime.fromisoformat(value.replace("Z", "+00:00"))
+def _event_filters(
+    *,
+    prefix: Optional[str],
+    origin_asn: Optional[int],
+    peer_asn: Optional[int],
+    collector: Optional[str],
+    event_type: Optional[str],
+    source: Optional[str],
+    time_from: Optional[datetime],
+    time_to: Optional[datetime],
+):
+    filters = [BGPObservation.source.in_(BGP_SOURCES)]
+    if source:
+        filters.append(BGPObservation.source == source)
+    if prefix:
+        filters.append(BGPObservation.prefix == prefix)
+    if origin_asn is not None:
+        filters.append(BGPObservation.origin_asn == origin_asn)
+    if peer_asn is not None:
+        filters.append(BGPObservation.peer_asn == peer_asn)
+    if collector:
+        filters.append(BGPObservation.collector == collector)
+    if event_type:
+        filters.append(BGPObservation.event_type == event_type)
+    if time_from:
+        filters.append(BGPObservation.observed_at >= time_from)
+    if time_to:
+        filters.append(BGPObservation.observed_at <= time_to)
+    return filters
 
 
-def _matches_time(value: Optional[datetime], time_from: Optional[datetime], time_to: Optional[datetime]) -> bool:
-    if value is None:
-        return False
-    if time_from and value < time_from:
-        return False
-    if time_to and value > time_to:
-        return False
-    return True
+def _anomaly_filters(
+    *,
+    severity: Optional[str],
+    anomaly_type: Optional[str],
+    status: Optional[str],
+    prefix: Optional[str],
+    origin_asn: Optional[int],
+    time_from: Optional[datetime],
+    time_to: Optional[datetime],
+):
+    filters = []
+    if severity:
+        filters.append(BGPAnomaly.severity == severity)
+    if anomaly_type:
+        filters.append(BGPAnomaly.anomaly_type == anomaly_type)
+    if status:
+        filters.append(BGPAnomaly.status == status)
+    if prefix:
+        filters.append(BGPAnomaly.prefix == prefix)
+    if origin_asn is not None:
+        filters.append(BGPAnomaly.origin_asn == origin_asn)
+    if time_from:
+        filters.append(BGPAnomaly.created_at >= time_from)
+    if time_to:
+        filters.append(BGPAnomaly.created_at <= time_to)
+    return filters
+
+
+def _incident_filters(
+    *,
+    severity: Optional[str],
+    incident_type: Optional[str],
+    status: Optional[str],
+):
+    filters = []
+    if severity:
+        filters.append(BGPIncident.severity == severity)
+    if incident_type:
+        filters.append(BGPIncident.incident_type == incident_type)
+    if status:
+        filters.append(BGPIncident.status == status)
+    return filters
+
+
+async def _build_event_summary_payload(db: AsyncSession) -> dict:
+    base_filters = [BGPObservation.source.in_(BGP_SOURCES)]
+
+    total_result = await db.execute(
+        select(func.count(BGPObservation.id)).where(*base_filters)
+    )
+    collectors_result = await db.execute(
+        select(func.count(func.distinct(BGPObservation.collector))).where(
+            *base_filters, BGPObservation.collector.isnot(None)
+        )
+    )
+    prefixes_result = await db.execute(
+        select(func.count(func.distinct(BGPObservation.prefix))).where(
+            *base_filters, BGPObservation.prefix.isnot(None)
+        )
+    )
+    type_result = await db.execute(
+        select(BGPObservation.event_type, func.count(BGPObservation.id))
+        .where(*base_filters)
+        .group_by(BGPObservation.event_type)
+    )
+
+    return {
+        "total": total_result.scalar() or 0,
+        "collector_count": collectors_result.scalar() or 0,
+        "prefix_count": prefixes_result.scalar() or 0,
+        "by_type": {row[0]: row[1] for row in type_result.fetchall()},
+    }
+
+
+async def _build_anomaly_summary_payload(db: AsyncSession) -> dict:
+    total_result = await db.execute(select(func.count(BGPAnomaly.id)))
+    type_result = await db.execute(
+        select(BGPAnomaly.anomaly_type, func.count(BGPAnomaly.id))
+        .group_by(BGPAnomaly.anomaly_type)
+        .order_by(func.count(BGPAnomaly.id).desc())
+    )
+    severity_result = await db.execute(
+        select(BGPAnomaly.severity, func.count(BGPAnomaly.id))
+        .group_by(BGPAnomaly.severity)
+        .order_by(func.count(BGPAnomaly.id).desc())
+    )
+    status_result = await db.execute(
+        select(BGPAnomaly.status, func.count(BGPAnomaly.id))
+        .group_by(BGPAnomaly.status)
+        .order_by(func.count(BGPAnomaly.id).desc())
+    )
+
+    return {
+        "total": total_result.scalar() or 0,
+        "by_type": {row[0]: row[1] for row in type_result.fetchall()},
+        "by_severity": {row[0]: row[1] for row in severity_result.fetchall()},
+        "by_status": {row[0]: row[1] for row in status_result.fetchall()},
+    }
+
+
+async def _build_incident_summary_payload(db: AsyncSession) -> dict:
+    total_result = await db.execute(select(func.count(BGPIncident.id)))
+    type_result = await db.execute(
+        select(BGPIncident.incident_type, func.count(BGPIncident.id))
+        .group_by(BGPIncident.incident_type)
+        .order_by(func.count(BGPIncident.id).desc())
+    )
+    severity_result = await db.execute(
+        select(BGPIncident.severity, func.count(BGPIncident.id))
+        .group_by(BGPIncident.severity)
+        .order_by(func.count(BGPIncident.id).desc())
+    )
+    status_result = await db.execute(
+        select(BGPIncident.status, func.count(BGPIncident.id))
+        .group_by(BGPIncident.status)
+        .order_by(func.count(BGPIncident.id).desc())
+    )
+
+    return {
+        "total": total_result.scalar() or 0,
+        "by_type": {row[0]: row[1] for row in type_result.fetchall()},
+        "by_severity": {row[0]: row[1] for row in severity_result.fetchall()},
+        "by_status": {row[0]: row[1] for row in status_result.fetchall()},
+    }
 
 
 @router.get("/events")
@@ -49,41 +194,36 @@ async def list_bgp_events(
     current_user: User = Depends(get_current_user),
     db: AsyncSession = Depends(get_db),
 ):
-    stmt = (
-        select(BGPObservation)
-        .where(BGPObservation.source.in_(BGP_SOURCES))
-        .order_by(BGPObservation.observed_at.desc(), BGPObservation.id.desc())
-    )
-    if source:
-        stmt = stmt.where(BGPObservation.source == source)
-
-    result = await db.execute(stmt)
-    records = result.scalars().all()
     dt_from = _parse_dt(time_from)
     dt_to = _parse_dt(time_to)
-
-    filtered = []
-    for record in records:
-        if prefix and record.prefix != prefix:
-            continue
-        if origin_asn is not None and record.origin_asn != origin_asn:
-            continue
-        if peer_asn is not None and record.peer_asn != peer_asn:
-            continue
-        if collector and record.collector != collector:
-            continue
-        if event_type and record.event_type != event_type:
-            continue
-        if (dt_from or dt_to) and not _matches_time(record.observed_at, dt_from, dt_to):
-            continue
-        filtered.append(record)
-
+    filters = _event_filters(
+        prefix=prefix,
+        origin_asn=origin_asn,
+        peer_asn=peer_asn,
+        collector=collector,
+        event_type=event_type,
+        source=source,
+        time_from=dt_from,
+        time_to=dt_to,
+    )
     offset = (page - 1) * page_size
+    count_result = await db.execute(
+        select(func.count(BGPObservation.id)).where(*filters)
+    )
+    data_result = await db.execute(
+        select(BGPObservation)
+        .where(*filters)
+        .order_by(BGPObservation.observed_at.desc(), BGPObservation.id.desc())
+        .offset(offset)
+        .limit(page_size)
+    )
+    records = data_result.scalars().all()
+
     return {
-        "total": len(filtered),
+        "total": count_result.scalar() or 0,
         "page": page,
         "page_size": page_size,
-        "data": [record.to_dict() for record in filtered[offset : offset + page_size]],
+        "data": [record.to_dict() for record in records],
     }
 
 
@@ -92,21 +232,7 @@ async def get_bgp_event_summary(
     current_user: User = Depends(get_current_user),
     db: AsyncSession = Depends(get_db),
 ):
-    result = await db.execute(select(BGPObservation).where(BGPObservation.source.in_(BGP_SOURCES)))
-    records = result.scalars().all()
-
-    collectors = sorted({record.collector for record in records if record.collector})
-    prefixes = sorted({record.prefix for record in records if record.prefix})
-    by_type: dict[str, int] = {}
-    for record in records:
-        by_type[record.event_type] = by_type.get(record.event_type, 0) + 1
-
-    return {
-        "total": len(records),
-        "collector_count": len(collectors),
-        "prefix_count": len(prefixes),
-        "by_type": by_type,
-    }
+    return await _build_event_summary_payload(db)
 
 
 @router.get("/collectors")
@@ -138,6 +264,32 @@ async def get_bgp_collector_summary(
     }
 
 
+@router.get("/overview/summary")
+async def get_bgp_overview_summary(
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    event_summary = await _build_event_summary_payload(db)
+    anomaly_summary = await _build_anomaly_summary_payload(db)
+    incident_summary = await _build_incident_summary_payload(db)
+    collectors = await build_bgp_collector_coverage(db, source_filter=BGP_SOURCES)
+    active_collectors = [item for item in collectors if item["observation_count"] > 0]
+
+    return {
+        "incidentSummary": incident_summary,
+        "anomalySummary": anomaly_summary,
+        "eventSummary": event_summary,
+        "collectorSummary": {
+            "total": len(collectors),
+            "active_collectors": len(active_collectors),
+            "observed_prefixes": sum(item["prefix_count"] for item in active_collectors),
+            "observed_origins": sum(item["origin_asn_count"] for item in active_collectors),
+            "recent_24h_events": sum(item["recent_24h_observation_count"] for item in active_collectors),
+            "recent_7d_events": sum(item["recent_7d_observation_count"] for item in active_collectors),
+        },
+    }
+
+
 @router.get("/events/{event_id}")
 async def get_bgp_event(
     event_id: int,
@@ -164,31 +316,35 @@ async def list_bgp_anomalies(
     current_user: User = Depends(get_current_user),
     db: AsyncSession = Depends(get_db),
 ):
-    stmt = select(BGPAnomaly).order_by(BGPAnomaly.created_at.desc(), BGPAnomaly.id.desc())
-    if severity:
-        stmt = stmt.where(BGPAnomaly.severity == severity)
-    if anomaly_type:
-        stmt = stmt.where(BGPAnomaly.anomaly_type == anomaly_type)
-    if status:
-        stmt = stmt.where(BGPAnomaly.status == status)
-    if prefix:
-        stmt = stmt.where(BGPAnomaly.prefix == prefix)
-    if origin_asn is not None:
-        stmt = stmt.where(BGPAnomaly.origin_asn == origin_asn)
-
-    result = await db.execute(stmt)
-    records = result.scalars().all()
     dt_from = _parse_dt(time_from)
     dt_to = _parse_dt(time_to)
-    if dt_from or dt_to:
-        records = [record for record in records if _matches_time(record.created_at, dt_from, dt_to)]
-
+    filters = _anomaly_filters(
+        severity=severity,
+        anomaly_type=anomaly_type,
+        status=status,
+        prefix=prefix,
+        origin_asn=origin_asn,
+        time_from=dt_from,
+        time_to=dt_to,
+    )
     offset = (page - 1) * page_size
+    total_result = await db.execute(
+        select(func.count(BGPAnomaly.id)).where(*filters)
+    )
+    data_result = await db.execute(
+        select(BGPAnomaly)
+        .where(*filters)
+        .order_by(BGPAnomaly.created_at.desc(), BGPAnomaly.id.desc())
+        .offset(offset)
+        .limit(page_size)
+    )
+    records = data_result.scalars().all()
+
     return {
-        "total": len(records),
+        "total": total_result.scalar() or 0,
         "page": page,
         "page_size": page_size,
-        "data": [record.to_dict() for record in records[offset : offset + page_size]],
+        "data": [record.to_dict() for record in records],
     }
 
 
@@ -197,29 +353,7 @@ async def get_bgp_anomaly_summary(
     current_user: User = Depends(get_current_user),
     db: AsyncSession = Depends(get_db),
 ):
-    total_result = await db.execute(select(func.count(BGPAnomaly.id)))
-    type_result = await db.execute(
-        select(BGPAnomaly.anomaly_type, func.count(BGPAnomaly.id))
-        .group_by(BGPAnomaly.anomaly_type)
-        .order_by(func.count(BGPAnomaly.id).desc())
-    )
-    severity_result = await db.execute(
-        select(BGPAnomaly.severity, func.count(BGPAnomaly.id))
-        .group_by(BGPAnomaly.severity)
-        .order_by(func.count(BGPAnomaly.id).desc())
-    )
-    status_result = await db.execute(
-        select(BGPAnomaly.status, func.count(BGPAnomaly.id))
-        .group_by(BGPAnomaly.status)
-        .order_by(func.count(BGPAnomaly.id).desc())
-    )
-
-    return {
-        "total": total_result.scalar() or 0,
-        "by_type": {row[0]: row[1] for row in type_result.fetchall()},
-        "by_severity": {row[0]: row[1] for row in severity_result.fetchall()},
-        "by_status": {row[0]: row[1] for row in status_result.fetchall()},
-    }
+    return await _build_anomaly_summary_payload(db)
 
 
 @router.get("/anomalies/{anomaly_id}")
@@ -244,22 +378,29 @@ async def list_bgp_incidents(
     current_user: User = Depends(get_current_user),
     db: AsyncSession = Depends(get_db),
 ):
-    stmt = select(BGPIncident).order_by(BGPIncident.created_at.desc(), BGPIncident.id.desc())
-    if severity:
-        stmt = stmt.where(BGPIncident.severity == severity)
-    if incident_type:
-        stmt = stmt.where(BGPIncident.incident_type == incident_type)
-    if status:
-        stmt = stmt.where(BGPIncident.status == status)
-
-    result = await db.execute(stmt)
-    records = result.scalars().all()
+    filters = _incident_filters(
+        severity=severity,
+        incident_type=incident_type,
+        status=status,
+    )
     offset = (page - 1) * page_size
+    total_result = await db.execute(
+        select(func.count(BGPIncident.id)).where(*filters)
+    )
+    data_result = await db.execute(
+        select(BGPIncident)
+        .where(*filters)
+        .order_by(BGPIncident.created_at.desc(), BGPIncident.id.desc())
+        .offset(offset)
+        .limit(page_size)
+    )
+    records = data_result.scalars().all()
+
     return {
-        "total": len(records),
+        "total": total_result.scalar() or 0,
         "page": page,
         "page_size": page_size,
-        "data": [record.to_dict() for record in records[offset : offset + page_size]],
+        "data": [record.to_dict() for record in records],
     }
 
 
@@ -268,29 +409,7 @@ async def get_bgp_incident_summary(
     current_user: User = Depends(get_current_user),
     db: AsyncSession = Depends(get_db),
 ):
-    total_result = await db.execute(select(func.count(BGPIncident.id)))
-    type_result = await db.execute(
-        select(BGPIncident.incident_type, func.count(BGPIncident.id))
-        .group_by(BGPIncident.incident_type)
-        .order_by(func.count(BGPIncident.id).desc())
-    )
-    severity_result = await db.execute(
-        select(BGPIncident.severity, func.count(BGPIncident.id))
-        .group_by(BGPIncident.severity)
-        .order_by(func.count(BGPIncident.id).desc())
-    )
-    status_result = await db.execute(
-        select(BGPIncident.status, func.count(BGPIncident.id))
-        .group_by(BGPIncident.status)
-        .order_by(func.count(BGPIncident.id).desc())
-    )
-
-    return {
-        "total": total_result.scalar() or 0,
-        "by_type": {row[0]: row[1] for row in type_result.fetchall()},
-        "by_severity": {row[0]: row[1] for row in severity_result.fetchall()},
-        "by_status": {row[0]: row[1] for row in status_result.fetchall()},
-    }
+    return await _build_incident_summary_payload(db)
 
 
 @router.get("/incidents/{incident_id}")

backend/app/api/v1/dashboard.py

@@ -2,7 +2,7 @@
 
 from datetime import UTC, datetime, timedelta
 from fastapi import APIRouter, Depends
-from sqlalchemy import select, func, text
+from sqlalchemy import case, select, func, text
 from sqlalchemy.ext.asyncio import AsyncSession
 
 from app.db.session import get_db
@@ -118,58 +118,77 @@ async def get_stats(
     built_in_count = len(COLLECTOR_INFO)
     built_in_active = built_in_count  # Built-in are always "active" for counting purposes
 
-    # Count custom configs from database
-    result = await db.execute(select(func.count(DataSourceConfig.id)))
-    custom_count = result.scalar() or 0
-
     result = await db.execute(
-        select(func.count(DataSourceConfig.id)).where(DataSourceConfig.is_active == True)
+        select(
+            func.count(DataSourceConfig.id).label("custom_count"),
+            func.sum(
+                case((DataSourceConfig.is_active == True, 1), else_=0)
+            ).label("custom_active"),
+        )
     )
-    custom_active = result.scalar() or 0
+    datasource_stats = result.one()
+    custom_count = datasource_stats.custom_count or 0
+    custom_active = datasource_stats.custom_active or 0
 
     # Total datasources
     total_datasources = built_in_count + custom_count
     active_datasources = built_in_active + custom_active
 
-    # Tasks today (from database)
     result = await db.execute(
-        select(func.count(CollectionTask.id)).where(CollectionTask.started_at >= today_start)
-    )
-    tasks_today = result.scalar() or 0
-
-    result = await db.execute(
-        select(func.count(CollectionTask.id)).where(
-            CollectionTask.status == "success",
-            CollectionTask.started_at >= today_start,
+        select(
+            func.count(CollectionTask.id).label("tasks_today"),
+            func.sum(
+                case(
+                    (CollectionTask.status == "success", 1),
+                    else_=0,
+                )
+            ).label("success_tasks"),
         )
+        .where(CollectionTask.started_at >= today_start)
     )
-    success_tasks = result.scalar() or 0
+    task_stats = result.one()
+    tasks_today = task_stats.tasks_today or 0
+    success_tasks = task_stats.success_tasks or 0
     success_rate = (success_tasks / tasks_today * 100) if tasks_today > 0 else 0
 
-    # Alerts
     result = await db.execute(
-        select(func.count(Alert.id)).where(
-            Alert.severity == AlertSeverity.CRITICAL,
-            Alert.status == "active",
+        select(
+            func.sum(
+                case(
+                    (
+                        (Alert.severity == AlertSeverity.CRITICAL)
+                        & (Alert.status == "active"),
+                        1,
+                    ),
+                    else_=0,
+                )
+            ).label("critical_alerts"),
+            func.sum(
+                case(
+                    (
+                        (Alert.severity == AlertSeverity.WARNING)
+                        & (Alert.status == "active"),
+                        1,
+                    ),
+                    else_=0,
+                )
+            ).label("warning_alerts"),
+            func.sum(
+                case(
+                    (
+                        (Alert.severity == AlertSeverity.INFO)
+                        & (Alert.status == "active"),
+                        1,
+                    ),
+                    else_=0,
+                )
+            ).label("info_alerts"),
         )
     )
-    critical_alerts = result.scalar() or 0
-
-    result = await db.execute(
-        select(func.count(Alert.id)).where(
-            Alert.severity == AlertSeverity.WARNING,
-            Alert.status == "active",
-        )
-    )
-    warning_alerts = result.scalar() or 0
-
-    result = await db.execute(
-        select(func.count(Alert.id)).where(
-            Alert.severity == AlertSeverity.INFO,
-            Alert.status == "active",
-        )
-    )
-    info_alerts = result.scalar() or 0
+    alert_stats = result.one()
+    critical_alerts = alert_stats.critical_alerts or 0
+    warning_alerts = alert_stats.warning_alerts or 0
+    info_alerts = alert_stats.info_alerts or 0
 
     response = {
         "total_datasources": total_datasources,

backend/app/api/v1/datasources.py

@@ -3,7 +3,7 @@ from datetime import datetime, timedelta, timezone
 from typing import Optional
 
 from fastapi import APIRouter, Depends, HTTPException, Query
-from sqlalchemy import func, select
+from sqlalchemy import func, select, text
 from sqlalchemy.ext.asyncio import AsyncSession
 
 from app.core.time import to_iso8601_utc
@@ -11,10 +11,17 @@ from app.core.security import get_current_user
 from app.core.data_sources import get_data_sources_config
 from app.db.session import get_db
 from app.models.collected_data import CollectedData
+from app.models.data_snapshot import DataSnapshot
 from app.models.datasource import DataSource
+from app.models.datasource_config import DataSourceConfig
 from app.models.task import CollectionTask
 from app.models.user import User
-from app.services.scheduler import get_latest_task_id_for_datasource, run_collector_now, sync_datasource_job
+from app.services.scheduler import (
+    cancel_running_collector_now,
+    get_latest_task_id_for_datasource,
+    run_collector_now,
+    sync_datasource_job,
+)
 
 router = APIRouter()
 STALE_RUNNING_TASK_TIMEOUT_MINUTES = 90
@@ -34,6 +41,156 @@ def is_due_for_collection(datasource: DataSource, now: datetime) -> bool:
     return datasource.last_run_at + timedelta(minutes=datasource.frequency_minutes) <= now
 
 
+def _task_rank_column(order_column):
+    return func.row_number().over(
+        partition_by=CollectionTask.datasource_id,
+        order_by=(order_column.desc().nullslast(), CollectionTask.id.desc()),
+    ).label("row_num")
+
+
+async def _load_latest_running_tasks(
+    db: AsyncSession,
+    datasource_ids: list[int],
+) -> dict[int, CollectionTask]:
+    if not datasource_ids:
+        return {}
+
+    ranked_tasks = (
+        select(
+            CollectionTask.id.label("task_id"),
+            _task_rank_column(CollectionTask.started_at),
+        )
+        .where(CollectionTask.datasource_id.in_(datasource_ids))
+        .where(CollectionTask.status == "running")
+        .subquery()
+    )
+    result = await db.execute(
+        select(CollectionTask)
+        .join(ranked_tasks, CollectionTask.id == ranked_tasks.c.task_id)
+        .where(ranked_tasks.c.row_num == 1)
+    )
+    return {task.datasource_id: task for task in result.scalars().all()}
+
+
+async def _load_latest_completed_tasks(
+    db: AsyncSession,
+    datasource_ids: list[int],
+) -> dict[int, CollectionTask]:
+    if not datasource_ids:
+        return {}
+
+    ranked_tasks = (
+        select(
+            CollectionTask.id.label("task_id"),
+            _task_rank_column(CollectionTask.completed_at),
+        )
+        .where(CollectionTask.datasource_id.in_(datasource_ids))
+        .where(CollectionTask.completed_at.isnot(None))
+        .where(CollectionTask.status.in_(("success", "failed", "cancelled")))
+        .subquery()
+    )
+    result = await db.execute(
+        select(CollectionTask)
+        .join(ranked_tasks, CollectionTask.id == ranked_tasks.c.task_id)
+        .where(ranked_tasks.c.row_num == 1)
+    )
+    return {task.datasource_id: task for task in result.scalars().all()}
+
+
+async def _load_latest_task_ids(
+    db: AsyncSession,
+    datasource_ids: list[int],
+) -> dict[int, int]:
+    if not datasource_ids:
+        return {}
+
+    ranked_tasks = (
+        select(
+            CollectionTask.id.label("task_id"),
+            CollectionTask.datasource_id.label("datasource_id"),
+            func.row_number().over(
+                partition_by=CollectionTask.datasource_id,
+                order_by=CollectionTask.id.desc(),
+            ).label("row_num"),
+        )
+        .where(CollectionTask.datasource_id.in_(datasource_ids))
+        .subquery()
+    )
+    result = await db.execute(
+        select(ranked_tasks.c.datasource_id, ranked_tasks.c.task_id)
+        .where(ranked_tasks.c.row_num == 1)
+    )
+    return {datasource_id: task_id for datasource_id, task_id in result.all()}
+
+
+async def _load_datasource_data_counts(
+    db: AsyncSession,
+    sources: list[str],
+) -> dict[str, int]:
+    if not sources:
+        return {}
+
+    result = await db.execute(
+        select(CollectedData.source, func.count(CollectedData.id))
+        .where(CollectedData.source.in_(sources))
+        .group_by(CollectedData.source)
+    )
+    return {source: count for source, count in result.all()}
+
+
+async def _load_datasource_endpoint_overrides(
+    db: AsyncSession,
+    sources: list[str],
+) -> dict[str, str]:
+    if not sources:
+        return {}
+
+    result = await db.execute(
+        select(DataSourceConfig.name, DataSourceConfig.endpoint)
+        .where(DataSourceConfig.name.in_(sources))
+        .where(DataSourceConfig.is_active.is_(True))
+        .where(DataSourceConfig.endpoint.isnot(None))
+    )
+    return {
+        name: endpoint
+        for name, endpoint in result.all()
+        if endpoint
+    }
+
+
+async def _load_datasource_list_context(
+    db: AsyncSession,
+    datasources: list[DataSource],
+) -> tuple[dict[int, CollectionTask], dict[int, CollectionTask], dict[str, int], dict[str, str]]:
+    datasource_ids = [datasource.id for datasource in datasources]
+    sources = [datasource.source for datasource in datasources]
+
+    running_tasks = await _load_latest_running_tasks(db, datasource_ids)
+    datasource_by_id = {datasource.id: datasource for datasource in datasources}
+    now = datetime.now(timezone.utc)
+
+    stale_datasource_ids: list[int] = []
+    for datasource_id, task in running_tasks.items():
+        started_at = task.started_at
+        if started_at is None:
+            continue
+        if started_at.tzinfo is None:
+            started_at = started_at.replace(tzinfo=timezone.utc)
+        if now - started_at > timedelta(minutes=STALE_RUNNING_TASK_TIMEOUT_MINUTES):
+            datasource = datasource_by_id.get(datasource_id)
+            if datasource is not None:
+                await fail_and_rollback_stale_running_task(db, datasource, task)
+                stale_datasource_ids.append(datasource_id)
+
+    if stale_datasource_ids:
+        running_tasks = await _load_latest_running_tasks(db, datasource_ids)
+
+    completed_tasks = await _load_latest_completed_tasks(db, datasource_ids)
+    data_counts = await _load_datasource_data_counts(db, sources)
+    endpoint_overrides = await _load_datasource_endpoint_overrides(db, sources)
+    return running_tasks, completed_tasks, data_counts, endpoint_overrides
+
+
 async def get_datasource_record(db: AsyncSession, source_id: str) -> Optional[DataSource]:
     datasource = None
     try:
@@ -52,18 +209,6 @@ async def get_datasource_record(db: AsyncSession, source_id: str) -> Optional[Da
     return result.scalar_one_or_none()
 
 
-async def get_last_completed_task(db: AsyncSession, datasource_id: int) -> Optional[CollectionTask]:
-    result = await db.execute(
-        select(CollectionTask)
-        .where(CollectionTask.datasource_id == datasource_id)
-        .where(CollectionTask.completed_at.isnot(None))
-        .where(CollectionTask.status.in_(("success", "failed", "cancelled")))
-        .order_by(CollectionTask.completed_at.desc())
-        .limit(1)
-    )
-    return result.scalar_one_or_none()
-
-
 async def get_running_task(db: AsyncSession, datasource_id: int) -> Optional[CollectionTask]:
     result = await db.execute(
         select(CollectionTask)
@@ -87,17 +232,152 @@ async def get_running_task(db: AsyncSession, datasource_id: int) -> Optional[Col
     if now - started_at <= timedelta(minutes=STALE_RUNNING_TASK_TIMEOUT_MINUTES):
         return task
 
-    existing_error = (task.error_message or "").strip()
+    datasource = await db.get(DataSource, datasource_id)
+    if datasource is not None:
+        await fail_and_rollback_stale_running_task(db, datasource, task)
+    else:
+        existing_error = (task.error_message or "").strip()
+        stale_reason = (
+            f"Marked failed automatically after stale running timeout "
+            f"({STALE_RUNNING_TASK_TIMEOUT_MINUTES}m)"
+        )
+        task.status = "failed"
+        task.phase = "failed"
+        task.completed_at = now
+        task.error_message = f"{existing_error}\n{stale_reason}".strip() if existing_error else stale_reason
+        await db.commit()
+    return None
+
+
+async def rollback_orphaned_running_task(
+    db: AsyncSession,
+    datasource: DataSource,
+    running_task: CollectionTask,
+) -> None:
+    snapshot_result = await db.execute(
+        select(DataSnapshot)
+        .where(
+            DataSnapshot.datasource_id == datasource.id,
+            DataSnapshot.task_id == running_task.id,
+        )
+        .order_by(DataSnapshot.id.desc())
+        .limit(1)
+    )
+    snapshot = snapshot_result.scalar_one_or_none()
+
+    await db.execute(CollectedData.__table__.delete().where(CollectedData.task_id == running_task.id))
+
+    await db.execute(
+        text(
+            """
+            UPDATE collected_data
+            SET is_current = FALSE
+            WHERE source = :source
+            """
+        ),
+        {"source": datasource.source},
+    )
+
+    if snapshot is not None:
+        snapshot.status = "cancelled"
+        snapshot.is_current = False
+        snapshot.completed_at = datetime.now(timezone.utc)
+        summary = dict(snapshot.summary or {})
+        summary["rollback"] = True
+        summary["rollback_reason"] = "orphaned_running_task_after_backend_restart"
+        snapshot.summary = summary
+
+        if snapshot.parent_snapshot_id is not None:
+            parent_snapshot = await db.get(DataSnapshot, snapshot.parent_snapshot_id)
+            if parent_snapshot:
+                parent_snapshot.is_current = True
+            await db.execute(
+                text(
+                    """
+                    UPDATE collected_data
+                    SET is_current = TRUE
+                    WHERE snapshot_id = :snapshot_id
+                    """
+                ),
+                {"snapshot_id": snapshot.parent_snapshot_id},
+            )
+
+    running_task.status = "cancelled"
+    running_task.phase = "cancelled"
+    running_task.completed_at = datetime.now(timezone.utc)
+    existing_error = (running_task.error_message or "").strip()
+    cancel_reason = "Cancelled after backend restart because the running task handle was lost; incomplete writes rolled back"
+    running_task.error_message = f"{existing_error}\n{cancel_reason}".strip() if existing_error else cancel_reason
+    datasource.last_status = "cancelled"
+    datasource.last_run_at = datetime.now(timezone.utc)
+    await db.commit()
+
+
+async def fail_and_rollback_stale_running_task(
+    db: AsyncSession,
+    datasource: DataSource,
+    running_task: CollectionTask,
+) -> None:
+    snapshot_result = await db.execute(
+        select(DataSnapshot)
+        .where(
+            DataSnapshot.datasource_id == datasource.id,
+            DataSnapshot.task_id == running_task.id,
+        )
+        .order_by(DataSnapshot.id.desc())
+        .limit(1)
+    )
+    snapshot = snapshot_result.scalar_one_or_none()
+
+    await db.execute(CollectedData.__table__.delete().where(CollectedData.task_id == running_task.id))
+
+    await db.execute(
+        text(
+            """
+            UPDATE collected_data
+            SET is_current = FALSE
+            WHERE source = :source
+            """
+        ),
+        {"source": datasource.source},
+    )
+
+    if snapshot is not None:
+        snapshot.status = "failed"
+        snapshot.is_current = False
+        snapshot.completed_at = datetime.now(timezone.utc)
+        summary = dict(snapshot.summary or {})
+        summary["rollback"] = True
+        summary["rollback_reason"] = "stale_running_task_timeout"
+        snapshot.summary = summary
+
+        if snapshot.parent_snapshot_id is not None:
+            parent_snapshot = await db.get(DataSnapshot, snapshot.parent_snapshot_id)
+            if parent_snapshot:
+                parent_snapshot.is_current = True
+            await db.execute(
+                text(
+                    """
+                    UPDATE collected_data
+                    SET is_current = TRUE
+                    WHERE snapshot_id = :snapshot_id
+                    """
+                ),
+                {"snapshot_id": snapshot.parent_snapshot_id},
+            )
+
+    existing_error = (running_task.error_message or "").strip()
     stale_reason = (
         f"Marked failed automatically after stale running timeout "
-        f"({STALE_RUNNING_TASK_TIMEOUT_MINUTES}m)"
+        f"({STALE_RUNNING_TASK_TIMEOUT_MINUTES}m); incomplete writes rolled back"
     )
-    task.status = "failed"
-    task.phase = "failed"
-    task.completed_at = now
-    task.error_message = f"{existing_error}\n{stale_reason}".strip() if existing_error else stale_reason
+    running_task.status = "failed"
+    running_task.phase = "failed"
+    running_task.completed_at = datetime.now(timezone.utc)
+    running_task.error_message = f"{existing_error}\n{stale_reason}".strip() if existing_error else stale_reason
+    datasource.last_status = "failed"
+    datasource.last_run_at = datetime.now(timezone.utc)
     await db.commit()
-    return None
 
 
 @router.get("")
@@ -121,14 +401,17 @@ async def list_datasources(
 
     collector_list = []
     config = get_data_sources_config()
+    running_tasks, completed_tasks, data_counts, endpoint_overrides = await _load_datasource_list_context(
+        db,
+        datasources,
+    )
     for datasource in datasources:
-        running_task = await get_running_task(db, datasource.id)
-        last_task = await get_last_completed_task(db, datasource.id)
-        endpoint = await config.get_url(datasource.source, db)
-        data_count_result = await db.execute(
-            select(func.count(CollectedData.id)).where(CollectedData.source == datasource.source)
+        running_task = running_tasks.get(datasource.id)
+        last_task = completed_tasks.get(datasource.id)
+        endpoint = endpoint_overrides.get(datasource.source) or config.get_yaml_url(
+            datasource.source,
         )
-        data_count = data_count_result.scalar() or 0
+        data_count = data_counts.get(datasource.source, 0)
 
         last_run_at = datasource.last_run_at or (last_task.completed_at if last_task else None)
         last_run = to_iso8601_utc(last_run_at)
@@ -189,9 +472,13 @@ async def trigger_all_datasources(
     skipped_sources: list[dict] = []
     failed_sources: list[dict] = []
     now = datetime.now(timezone.utc)
+    running_tasks = await _load_latest_running_tasks(
+        db,
+        [datasource.id for datasource in datasources],
+    )
 
     for datasource in datasources:
-        running_task = await get_running_task(db, datasource.id)
+        running_task = running_tasks.get(datasource.id)
         if running_task is not None:
             skipped_sources.append(
                 {
@@ -219,7 +506,7 @@ async def trigger_all_datasources(
             )
             continue
 
-        previous_task_ids[datasource.id] = await get_latest_task_id_for_datasource(datasource.id)
+        previous_task_ids[datasource.id] = None
         success = run_collector_now(datasource.source)
         if not success:
             failed_sources.append(
@@ -241,13 +528,24 @@ async def trigger_all_datasources(
             }
         )
 
+    latest_task_ids = await _load_latest_task_ids(
+        db,
+        [datasource.id for datasource in datasources],
+    )
+    for datasource_id in previous_task_ids:
+        previous_task_ids[datasource_id] = latest_task_ids.get(datasource_id)
+
     for _ in range(20):
         await asyncio.sleep(0.1)
         pending = [item for item in triggered_sources if item["task_id"] is None]
         if not pending:
             break
+        latest_task_ids = await _load_latest_task_ids(
+            db,
+            [item["id"] for item in pending],
+        )
         for item in pending:
-            task_id = await get_latest_task_id_for_datasource(item["id"])
+            task_id = latest_task_ids.get(item["id"])
             if task_id is not None and task_id != previous_task_ids.get(item["id"]):
                 item["task_id"] = task_id
 
@@ -346,6 +644,7 @@ async def get_datasource_stats(
 @router.post("/{source_id}/trigger")
 async def trigger_datasource(
     source_id: str,
+    force: bool = Query(False),
     current_user: User = Depends(get_current_user),
     db: AsyncSession = Depends(get_db),
 ):
@@ -356,6 +655,26 @@ async def trigger_datasource(
     if not datasource.is_active:
         raise HTTPException(status_code=400, detail="Data source is disabled")
 
+    running_task = await get_running_task(db, datasource.id)
+    if running_task is not None and not force:
+        raise HTTPException(
+            status_code=409,
+            detail={
+                "reason": "running_task_in_progress",
+                "message": "当前采集任务尚未完成，重新触发会丢失本次未完成进度。是否强制重新采集？",
+                "task_id": running_task.id,
+                "phase": running_task.phase,
+                "progress": running_task.progress,
+                "records_processed": running_task.records_processed,
+                "total_records": running_task.total_records,
+            },
+        )
+
+    if running_task is not None and force:
+        cancelled = await cancel_running_collector_now(datasource.source)
+        if not cancelled:
+            await rollback_orphaned_running_task(db, datasource, running_task)
+
     previous_task_id = await get_latest_task_id_for_datasource(datasource.id)
     success = run_collector_now(datasource.source)
     if not success:
@@ -375,6 +694,7 @@ async def trigger_datasource(
         "source_id": datasource.id,
         "task_id": task_id,
         "collector_name": datasource.source,
+        "force": force,
         "message": f"Collector '{datasource.source}' has been triggered",
     }
 
@@ -412,6 +732,7 @@ async def clear_datasource_data(
 async def get_task_status(
     source_id: str,
     task_id: Optional[int] = None,
+    current_user: User = Depends(get_current_user),
     db: AsyncSession = Depends(get_db),
 ):
     datasource = await get_datasource_record(db, source_id)

backend/app/api/v1/settings.py

@@ -1,3 +1,4 @@
+from copy import deepcopy
 from datetime import UTC, datetime
 from typing import Optional
 
@@ -13,6 +14,7 @@ from app.models.datasource import DataSource
 from app.models.system_setting import SystemSetting
 from app.models.user import User
 from app.services.scheduler import sync_datasource_job
+from app.services.tv_streams import DEFAULT_TV_SETTINGS, get_tv_settings_payload, normalize_tv_settings
 
 router = APIRouter()
 
@@ -36,6 +38,7 @@ DEFAULT_SETTINGS = {
         "max_login_attempts": 5,
         "password_policy": "medium",
     },
+    "tv": DEFAULT_TV_SETTINGS,
 }
 
 
@@ -67,8 +70,34 @@ class CollectorSettingsUpdate(BaseModel):
     frequency_minutes: int = Field(default=60, ge=1, le=10080)
 
 
+class TVStreamSourceUpdate(BaseModel):
+    id: str = Field(min_length=1, max_length=100)
+    name: str = Field(min_length=1, max_length=200)
+    provider: str = Field(default="Unknown", max_length=100)
+    region: str = Field(default="Global", max_length=100)
+    language: str = Field(default="und", max_length=32)
+    source_type: str = Field(default="iframe", pattern="^(iframe|hls|video|external|youtube)$")
+    embed_url: str = ""
+    stream_url: str = ""
+    homepage_url: str = ""
+    poster_url: str = ""
+    youtube_video_id: str = ""
+    youtube_channel: str = ""
+    is_enabled: bool = True
+    is_fallback: bool = False
+    sort_order: int = Field(default=10, ge=0, le=9999)
+    collector_source: Optional[str] = None
+    notes: str = ""
+
+
+class TVSettingsUpdate(BaseModel):
+    default_source_id: str = Field(default=DEFAULT_TV_SETTINGS["default_source_id"], min_length=1)
+    auto_fallback: bool = True
+    sources: list[TVStreamSourceUpdate] = Field(default_factory=list)
+
+
 def merge_with_defaults(category: str, payload: Optional[dict]) -> dict:
-    merged = DEFAULT_SETTINGS[category].copy()
+    merged = deepcopy(DEFAULT_SETTINGS[category])
     if payload:
         merged.update(payload)
     return merged
@@ -79,6 +108,26 @@ async def get_setting_record(db: AsyncSession, category: str) -> Optional[System
     return result.scalar_one_or_none()
 
 
+async def get_setting_payloads(db: AsyncSession, categories: list[str]) -> dict[str, dict]:
+    if not categories:
+        return {}
+
+    result = await db.execute(
+        select(SystemSetting).where(SystemSetting.category.in_(categories))
+    )
+    records_by_category = {
+        record.category: record
+        for record in result.scalars().all()
+    }
+    return {
+        category: merge_with_defaults(
+            category,
+            records_by_category.get(category).payload if records_by_category.get(category) else None,
+        )
+        for category in categories
+    }
+
+
 async def get_setting_payload(db: AsyncSession, category: str) -> dict:
     record = await get_setting_record(db, category)
     return merge_with_defaults(category, record.payload if record else None)
@@ -175,6 +224,25 @@ async def update_security_settings(
     return {"status": "updated", "security": payload}
 
 
+@router.get("/tv")
+async def get_tv_settings(
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    return {"tv": await get_tv_settings_payload(db)}
+
+
+@router.put("/tv")
+async def update_tv_settings(
+    settings: TVSettingsUpdate,
+    current_user: User = Depends(get_current_user),
+    db: AsyncSession = Depends(get_db),
+):
+    payload = normalize_tv_settings(settings.model_dump())
+    saved = await save_setting_payload(db, "tv", payload)
+    return {"status": "updated", "tv": normalize_tv_settings(saved)}
+
+
 @router.get("/collectors")
 async def get_collector_settings(
     current_user: User = Depends(get_current_user),
@@ -212,10 +280,15 @@ async def get_all_settings(
 ):
     result = await db.execute(select(DataSource).order_by(DataSource.module, DataSource.id))
     datasources = result.scalars().all()
+    setting_payloads = await get_setting_payloads(
+        db,
+        ["system", "notifications", "security"],
+    )
     return {
-        "system": await get_setting_payload(db, "system"),
-        "notifications": await get_setting_payload(db, "notifications"),
-        "security": await get_setting_payload(db, "security"),
+        "system": setting_payloads["system"],
+        "notifications": setting_payloads["notifications"],
+        "security": setting_payloads["security"],
+        "tv": await get_tv_settings_payload(db),
         "collectors": [serialize_collector(datasource) for datasource in datasources],
         "generated_at": to_iso8601_utc(datetime.now(UTC)),
     }

backend/app/api/v1/tv.py

@@ -0,0 +1,70 @@
+from urllib.parse import quote, urljoin
+
+import httpx
+from fastapi import APIRouter, Depends, HTTPException, Query
+from fastapi.responses import Response
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.db.session import get_db
+from app.services.tv_streams import get_public_tv_payload, is_allowed_tv_proxy_url
+
+router = APIRouter()
+
+
+@router.get("/streams")
+async def list_public_tv_streams(
+    db: AsyncSession = Depends(get_db),
+):
+    return await get_public_tv_payload(db)
+
+
+@router.get("/proxy")
+async def proxy_tv_stream(
+    url: str = Query(..., description="Upstream TV stream or manifest URL"),
+    db: AsyncSession = Depends(get_db),
+):
+    payload = await get_public_tv_payload(db)
+    if not is_allowed_tv_proxy_url(url, payload.get("sources", [])):
+        raise HTTPException(status_code=403, detail="TV proxy target is not allowed")
+
+    try:
+        async with httpx.AsyncClient(follow_redirects=True, timeout=20.0) as client:
+            upstream = await client.get(
+                url,
+                headers={
+                    "User-Agent": "Mozilla/5.0",
+                    "Referer": "https://tv.cctv.com/live/cctv4/",
+                },
+            )
+            upstream.raise_for_status()
+    except httpx.HTTPError as exc:
+        raise HTTPException(status_code=502, detail=f"Failed to fetch TV stream: {exc}") from exc
+
+    content_type = upstream.headers.get("content-type", "application/octet-stream")
+    raw_content = upstream.content
+    response_url = str(upstream.url)
+    is_manifest = (
+        response_url.endswith(".m3u8")
+        or "mpegurl" in content_type.lower()
+        or raw_content.lstrip().startswith(b"#EXTM3U")
+    )
+
+    headers = {"Cache-Control": "no-store"}
+
+    if is_manifest:
+        manifest_text = upstream.text
+        rewritten_lines: list[str] = []
+        for line in manifest_text.splitlines():
+            stripped = line.strip()
+            if not stripped or stripped.startswith("#"):
+                rewritten_lines.append(line)
+                continue
+            absolute_url = urljoin(response_url, stripped)
+            rewritten_lines.append(f"/api/v1/tv/proxy?url={quote(absolute_url, safe='')}")
+        return Response(
+            content="\n".join(rewritten_lines),
+            media_type="application/vnd.apple.mpegurl",
+            headers=headers,
+        )
+
+    return Response(content=raw_content, media_type=content_type, headers=headers)

backend/app/api/v1/visualization.py

@@ -205,42 +205,84 @@ def convert_satellite_to_geojson(records: List[CollectedData]) -> Dict[str, Any]
     return {"type": "FeatureCollection", "features": features}
 
 
-def dedupe_satellite_records(records: List[CollectedData]) -> List[CollectedData]:
-    """Keep only the newest record for each satellite identity."""
-    latest_by_key: Dict[str, CollectedData] = {}
-
-    for record in records:
-        metadata = record.extra_data or {}
-        norad_id = metadata.get("norad_cat_id")
-        dedupe_key = (
-            str(norad_id)
-            if norad_id not in (None, "")
-            else str(record.source_id or record.entity_key or record.name or record.id)
-        )
-
-        existing = latest_by_key.get(dedupe_key)
-        if existing is None or (record.id or 0) > (existing.id or 0):
-            latest_by_key[dedupe_key] = record
-
-    return sorted(latest_by_key.values(), key=lambda item: item.id or 0, reverse=True)
+def _current_collected_data_stmt(source: str):
+    return (
+        select(CollectedData)
+        .where(CollectedData.source == source)
+        .where(CollectedData.is_current.is_(True))
+        .order_by(CollectedData.id.desc())
+    )
 
 
-def dedupe_collected_records(records: List[CollectedData]) -> List[CollectedData]:
-    """Keep only the newest record for each collected entity."""
-    latest_by_key: Dict[str, CollectedData] = {}
+async def _load_current_collected_data(
+    db: AsyncSession,
+    source: str,
+    *,
+    exclude_unknown_name: bool = False,
+    limit: Optional[int] = None,
+) -> List[CollectedData]:
+    stmt = _current_collected_data_stmt(source)
+    if exclude_unknown_name:
+        stmt = stmt.where(CollectedData.name != "Unknown")
+    if limit is not None:
+        stmt = stmt.limit(limit)
 
-    for record in records:
-        dedupe_key = str(
-            record.source_id
-            or record.entity_key
-            or record.name
-            or record.id
-        )
-        existing = latest_by_key.get(dedupe_key)
-        if existing is None or (record.id or 0) > (existing.id or 0):
-            latest_by_key[dedupe_key] = record
+    result = await db.execute(stmt)
+    return list(result.scalars().all())
 
-    return sorted(latest_by_key.values(), key=lambda item: item.id or 0, reverse=True)
+
+async def _load_current_collected_data_by_sources(
+    db: AsyncSession,
+    sources: List[str],
+) -> Dict[str, List[CollectedData]]:
+    if not sources:
+        return {}
+
+    stmt = (
+        select(CollectedData)
+        .where(CollectedData.source.in_(sources))
+        .where(CollectedData.is_current.is_(True))
+        .order_by(CollectedData.source.asc(), CollectedData.id.desc())
+    )
+    result = await db.execute(stmt)
+
+    grouped_records: Dict[str, List[CollectedData]] = {source: [] for source in sources}
+    for record in result.scalars().all():
+        grouped_records.setdefault(record.source, []).append(record)
+
+    return grouped_records
+
+
+def _build_landing_point_cable_maps(
+    relation_records: List[CollectedData],
+    cable_records: List[CollectedData],
+) -> tuple[Dict[int, List[int]], Dict[int, str]]:
+    city_to_cable_ids_map: Dict[int, List[int]] = {}
+    for relation_record in relation_records:
+        if not relation_record.extra_data:
+            continue
+        city_id = relation_record.extra_data.get("city_id")
+        cable_id = relation_record.extra_data.get("cable_id")
+        if city_id is None or cable_id is None:
+            continue
+        city_to_cable_ids_map.setdefault(city_id, [])
+        if cable_id not in city_to_cable_ids_map[city_id]:
+            city_to_cable_ids_map[city_id].append(cable_id)
+
+    cable_id_to_name_map: Dict[int, str] = {}
+    for cable_record in cable_records:
+        if not cable_record.extra_data:
+            continue
+        cable_id = cable_record.extra_data.get("cable_id")
+        cable_name = cable_record.name
+        if cable_id and cable_name:
+            cable_id_to_name_map[cable_id] = cable_name
+
+    return city_to_cable_ids_map, cable_id_to_name_map
+
+
+def _filter_known_records(records: List[CollectedData]) -> List[CollectedData]:
+    return [record for record in records if record.name != "Unknown"]
 
 
 def convert_supercomputer_to_geojson(records: List[CollectedData]) -> Dict[str, Any]:
@@ -722,9 +764,7 @@ def convert_bgp_incidents_to_geojson(
 async def get_cables_geojson(db: AsyncSession = Depends(get_db)):
     """获取海底电缆 GeoJSON 数据 (LineString)"""
     try:
-        stmt = select(CollectedData).where(CollectedData.source == "arcgis_cables")
-        result = await db.execute(stmt)
-        records = dedupe_collected_records(list(result.scalars().all()))
+        records = await _load_current_collected_data(db, "arcgis_cables")
 
         if not records:
             raise HTTPException(
@@ -742,36 +782,14 @@ async def get_cables_geojson(db: AsyncSession = Depends(get_db)):
 @router.get("/geo/landing-points")
 async def get_landing_points_geojson(db: AsyncSession = Depends(get_db)):
     try:
-        landing_stmt = select(CollectedData).where(CollectedData.source == "arcgis_landing_points")
-        landing_result = await db.execute(landing_stmt)
-        records = dedupe_collected_records(list(landing_result.scalars().all()))
-        
-        relation_stmt = select(CollectedData).where(CollectedData.source == "arcgis_cable_landing_relation")
-        relation_result = await db.execute(relation_stmt)
-        relation_records = dedupe_collected_records(list(relation_result.scalars().all()))
-        
-        cable_stmt = select(CollectedData).where(CollectedData.source == "arcgis_cables")
-        cable_result = await db.execute(cable_stmt)
-        cable_records = dedupe_collected_records(list(cable_result.scalars().all()))
-        
-        city_to_cable_ids_map = {}
-        for rel in relation_records:
-            if rel.extra_data:
-                city_id = rel.extra_data.get("city_id")
-                cable_id = rel.extra_data.get("cable_id")
-                if city_id is not None and cable_id is not None:
-                    if city_id not in city_to_cable_ids_map:
-                        city_to_cable_ids_map[city_id] = []
-                    if cable_id not in city_to_cable_ids_map[city_id]:
-                        city_to_cable_ids_map[city_id].append(cable_id)
-        
-        cable_id_to_name_map = {}
-        for cable in cable_records:
-            if cable.extra_data:
-                cable_id = cable.extra_data.get("cable_id")
-                cable_name = cable.name
-                if cable_id and cable_name:
-                    cable_id_to_name_map[cable_id] = cable_name
+        records = await _load_current_collected_data(db, "arcgis_landing_points")
+        relation_records = await _load_current_collected_data(db, "arcgis_cable_landing_relation")
+        cable_records = await _load_current_collected_data(db, "arcgis_cables")
+
+        city_to_cable_ids_map, cable_id_to_name_map = _build_landing_point_cable_maps(
+            relation_records,
+            cable_records,
+        )
         
         if not records:
             raise HTTPException(
@@ -788,36 +806,21 @@ async def get_landing_points_geojson(db: AsyncSession = Depends(get_db)):
 
 @router.get("/geo/all")
 async def get_all_geojson(db: AsyncSession = Depends(get_db)):
-    cables_stmt = select(CollectedData).where(CollectedData.source == "arcgis_cables")
-    cables_result = await db.execute(cables_stmt)
-    cables_records = dedupe_collected_records(list(cables_result.scalars().all()))
-
-    points_stmt = select(CollectedData).where(CollectedData.source == "arcgis_landing_points")
-    points_result = await db.execute(points_stmt)
-    points_records = dedupe_collected_records(list(points_result.scalars().all()))
-    
-    relation_stmt = select(CollectedData).where(CollectedData.source == "arcgis_cable_landing_relation")
-    relation_result = await db.execute(relation_stmt)
-    relation_records = dedupe_collected_records(list(relation_result.scalars().all()))
-    
-    city_to_cable_ids_map = {}
-    for rel in relation_records:
-        if rel.extra_data:
-            city_id = rel.extra_data.get("city_id")
-            cable_id = rel.extra_data.get("cable_id")
-            if city_id is not None and cable_id is not None:
-                if city_id not in city_to_cable_ids_map:
-                    city_to_cable_ids_map[city_id] = []
-                if cable_id not in city_to_cable_ids_map[city_id]:
-                    city_to_cable_ids_map[city_id].append(cable_id)
-    
-    cable_id_to_name_map = {}
-    for cable in cables_records:
-        if cable.extra_data:
-            cable_id = cable.extra_data.get("cable_id")
-            cable_name = cable.name
-            if cable_id and cable_name:
-                cable_id_to_name_map[cable_id] = cable_name
+    records_by_source = await _load_current_collected_data_by_sources(
+        db,
+        [
+            "arcgis_cables",
+            "arcgis_landing_points",
+            "arcgis_cable_landing_relation",
+        ],
+    )
+    cables_records = records_by_source.get("arcgis_cables", [])
+    points_records = records_by_source.get("arcgis_landing_points", [])
+    relation_records = records_by_source.get("arcgis_cable_landing_relation", [])
+    city_to_cable_ids_map, cable_id_to_name_map = _build_landing_point_cable_maps(
+        relation_records,
+        cables_records,
+    )
 
     cables = (
         convert_cable_to_geojson(cables_records)
@@ -850,17 +853,12 @@ async def get_satellites_geojson(
     db: AsyncSession = Depends(get_db),
 ):
     """获取卫星 TLE GeoJSON 数据"""
-    stmt = (
-        select(CollectedData)
-        .where(CollectedData.source == "celestrak_tle")
-        .where(CollectedData.name != "Unknown")
-        .order_by(CollectedData.id.desc())
+    records = await _load_current_collected_data(
+        db,
+        "celestrak_tle",
+        exclude_unknown_name=True,
+        limit=limit,
     )
-    result = await db.execute(stmt)
-    records = dedupe_satellite_records(list(result.scalars().all()))
-
-    if limit is not None:
-        records = records[:limit]
 
     if not records:
         return {"type": "FeatureCollection", "features": [], "count": 0}
@@ -878,15 +876,12 @@ async def get_supercomputers_geojson(
     db: AsyncSession = Depends(get_db),
 ):
     """获取 TOP500 超算中心 GeoJSON 数据"""
-    stmt = (
-        select(CollectedData)
-        .where(CollectedData.source == "top500")
-        .where(CollectedData.name != "Unknown")
-        .order_by(CollectedData.id.desc())
+    records = await _load_current_collected_data(
+        db,
+        "top500",
+        exclude_unknown_name=True,
+        limit=limit,
     )
-    result = await db.execute(stmt)
-    records = dedupe_collected_records(list(result.scalars().all()))
-    records = records[:limit]
 
     if not records:
         return {"type": "FeatureCollection", "features": [], "count": 0}
@@ -904,15 +899,12 @@ async def get_gpu_clusters_geojson(
     db: AsyncSession = Depends(get_db),
 ):
     """获取 GPU 集群 GeoJSON 数据"""
-    stmt = (
-        select(CollectedData)
-        .where(CollectedData.source == "epoch_ai_gpu")
-        .where(CollectedData.name != "Unknown")
-        .order_by(CollectedData.id.desc())
+    records = await _load_current_collected_data(
+        db,
+        "epoch_ai_gpu",
+        exclude_unknown_name=True,
+        limit=limit,
     )
-    result = await db.execute(stmt)
-    records = dedupe_collected_records(list(result.scalars().all()))
-    records = records[:limit]
 
     if not records:
         return {"type": "FeatureCollection", "features": [], "count": 0}
@@ -990,37 +982,27 @@ async def get_all_visualization_data(db: AsyncSession = Depends(get_db)):
     - supercomputers: TOP500 超算
     - gpu_clusters: GPU 集群
     """
-    cables_stmt = select(CollectedData).where(CollectedData.source == "arcgis_cables")
-    cables_result = await db.execute(cables_stmt)
-    cables_records = dedupe_collected_records(list(cables_result.scalars().all()))
-
-    points_stmt = select(CollectedData).where(CollectedData.source == "arcgis_landing_points")
-    points_result = await db.execute(points_stmt)
-    points_records = dedupe_collected_records(list(points_result.scalars().all()))
-
-    satellites_stmt = (
-        select(CollectedData)
-        .where(CollectedData.source == "celestrak_tle")
-        .where(CollectedData.name != "Unknown")
+    records_by_source = await _load_current_collected_data_by_sources(
+        db,
+        [
+            "arcgis_cables",
+            "arcgis_landing_points",
+            "celestrak_tle",
+            "top500",
+            "epoch_ai_gpu",
+        ],
     )
-    satellites_result = await db.execute(satellites_stmt)
-    satellites_records = dedupe_satellite_records(list(satellites_result.scalars().all()))
-
-    supercomputers_stmt = (
-        select(CollectedData)
-        .where(CollectedData.source == "top500")
-        .where(CollectedData.name != "Unknown")
+    cables_records = records_by_source.get("arcgis_cables", [])
+    points_records = records_by_source.get("arcgis_landing_points", [])
+    satellites_records = _filter_known_records(
+        records_by_source.get("celestrak_tle", []),
     )
-    supercomputers_result = await db.execute(supercomputers_stmt)
-    supercomputers_records = dedupe_collected_records(list(supercomputers_result.scalars().all()))
-
-    gpu_stmt = (
-        select(CollectedData)
-        .where(CollectedData.source == "epoch_ai_gpu")
-        .where(CollectedData.name != "Unknown")
+    supercomputers_records = _filter_known_records(
+        records_by_source.get("top500", []),
+    )
+    gpu_records = _filter_known_records(
+        records_by_source.get("epoch_ai_gpu", []),
     )
-    gpu_result = await db.execute(gpu_stmt)
-    gpu_records = dedupe_collected_records(list(gpu_result.scalars().all()))
 
     cables = (
         convert_cable_to_geojson(cables_records)
@@ -1084,13 +1066,8 @@ async def get_cable_graph(db: AsyncSession) -> CableGraph:
     global _cable_graph
 
     if _cable_graph is None:
-        cables_stmt = select(CollectedData).where(CollectedData.source == "arcgis_cables")
-        cables_result = await db.execute(cables_stmt)
-        cables_records = list(cables_result.scalars().all())
-
-        points_stmt = select(CollectedData).where(CollectedData.source == "arcgis_landing_points")
-        points_result = await db.execute(points_stmt)
-        points_records = list(points_result.scalars().all())
+        cables_records = await _load_current_collected_data(db, "arcgis_cables")
+        points_records = await _load_current_collected_data(db, "arcgis_landing_points")
 
         cables_data = convert_cable_to_geojson(cables_records)
         points_data = convert_landing_point_to_geojson(points_records)

backend/app/core/data_sources.py

@@ -11,6 +11,7 @@ COLLECTOR_URL_KEYS = {
     "fao_landing_points": "fao.landing_point_url",
     "telegeography_cables": "telegeography.cable_url",
     "telegeography_landing": "telegeography.landing_point_url",
+    "telegeography_systems": "telegeography.cable_url",
     "huggingface_models": "huggingface.models_url",
     "huggingface_datasets": "huggingface.datasets_url",
     "huggingface_spaces": "huggingface.spaces_url",
@@ -23,6 +24,7 @@ COLLECTOR_URL_KEYS = {
     "top500": "top500.url",
     "epoch_ai_gpu": "epoch_ai.gpu_clusters_url",
     "spacetrack_tle": "spacetrack.tle_query_url",
+    "celestrak_tle": "celestrak.base_url",
     "ris_live_bgp": "ris_live.url",
     "bgpstream_bgp": "bgpstream.url",
     "iptoasn_prefix_geo": "iptoasn.combined_url",
@@ -41,18 +43,22 @@ class DataSourcesConfig:
             with open(config_path, "r") as f:
                 self._yaml_config = yaml.safe_load(f) or {}
 
-    def get_yaml_url(self, collector_name: str) -> str:
-        key = COLLECTOR_URL_KEYS.get(collector_name, "")
+    def get_yaml_value(self, key: str):
         if not key:
-            return ""
+            return None
 
         parts = key.split(".")
         value = self._yaml_config
         for part in parts:
             if isinstance(value, dict):
-                value = value.get(part, "")
+                value = value.get(part)
             else:
-                return ""
+                return None
+        return value
+
+    def get_yaml_url(self, collector_name: str) -> str:
+        key = COLLECTOR_URL_KEYS.get(collector_name, "")
+        value = self.get_yaml_value(key)
         return value if isinstance(value, str) else ""
 
     async def get_url(self, collector_name: str, db) -> str:

backend/app/core/data_sources.yaml

@@ -2,53 +2,87 @@
 # All external data source URLs should be configured here
 
 arcgis:
+  # ArcGIS 海缆 GeoJSON 查询接口
   cable_url: "https://services.arcgis.com/6DIQcwlPy8knb6sg/ArcGIS/rest/services/SubmarineCables/FeatureServer/2/query"
+  # ArcGIS 登陆点 GeoJSON 查询接口
   landing_point_url: "https://services.arcgis.com/6DIQcwlPy8knb6sg/ArcGIS/rest/services/SubmarineCables/FeatureServer/1/query"
+  # ArcGIS 海缆与登陆点关联关系查询接口
   cable_landing_relation_url: "https://services.arcgis.com/6DIQcwlPy8knb6sg/ArcGIS/rest/services/SubmarineCables/FeatureServer/3/query"
 
 fao:
+  # FAO 登陆点 CSV 下载地址
   landing_point_url: "https://data.apps.fao.org/catalog/dataset/1b75ff21-92f2-4b96-9b7b-98e8aa65ad5d/resource/b6071077-d1d4-4e97-aa00-42e902847c87/download/landing-point-geo.csv"
 
 telegeography:
+  # TeleGeography 海缆/系统主数据源，当前使用 GitHub 镜像 JSON
   cable_url: "https://raw.githubusercontent.com/lintaojlu/submarine_cable_information/main/cable.json"
+  # TeleGeography 登陆点主数据源，当前使用 GitHub 镜像 JSON
   landing_point_url: "https://raw.githubusercontent.com/lintaojlu/submarine_cable_information/main/landing_point.json"
+  # TeleGeography 历史 API 存档，用于 cable collector 的 fallback
+  archived_cable_url: "https://web.archive.org/web/2024/https://www.submarinecablemap.com/api/v3/cable"
+  # TeleGeography 官网页面，用于 cable collector 的最终 HTML 抓取 fallback
+  live_map_url: "https://www.submarinecablemap.com"
 
 huggingface:
+  # Hugging Face 模型目录 API
   models_url: "https://huggingface.co/api/models"
+  # Hugging Face 数据集目录 API
   datasets_url: "https://huggingface.co/api/datasets"
+  # Hugging Face Spaces 目录 API
   spaces_url: "https://huggingface.co/api/spaces"
 
 cloudflare:
+  # Cloudflare Radar 设备类型摘要接口
   radar_device_url: "https://api.cloudflare.com/client/v4/radar/http/summary/device_type"
+  # Cloudflare Radar 请求量时间序列接口
   radar_traffic_url: "https://api.cloudflare.com/client/v4/radar/http/timeseries/requests"
+  # Cloudflare Radar 热点地理位置接口
   radar_top_locations_url: "https://api.cloudflare.com/client/v4/radar/http/top/locations"
 
 peeringdb:
+  # PeeringDB IXP API
   ixp_url: "https://www.peeringdb.com/api/ix"
+  # PeeringDB Network API
   network_url: "https://www.peeringdb.com/api/net"
+  # PeeringDB Facility API
   facility_url: "https://www.peeringdb.com/api/fac"
 
 top500:
+  # TOP500 榜单页面，用于主表抓取
   url: "https://top500.org/lists/top500/list/2025/11/"
+  # TOP500 站点根地址，用于拼详情页链接
+  base_url: "https://top500.org"
 
 epoch_ai:
+  # Epoch AI GPU Cluster 页面
   gpu_clusters_url: "https://epoch.ai/data/gpu-clusters"
 
 spacetrack:
+  # Space-Track 站点根地址，用于首页访问和登录地址推导
   base_url: "https://www.space-track.org"
+  # Space-Track TLE 主查询接口
   tle_query_url: "https://www.space-track.org/basicspacedata/query/class/gp/orderby/EPOCH%20desc/limit/1000/format/json"
 
+celestrak:
+  # CelesTrak TLE 基础接口，collector 会在其后拼接 GROUP / FORMAT 参数
+  base_url: "https://celestrak.org/NORAD/elements/gp.php"
+
 ris_live:
+  # RIPE RIS Live 流式订阅地址
   url: "https://ris-live.ripe.net/v1/stream/?format=json&client=planet-ris-live"
 
 bgpstream:
+  # CAIDA BGPStream Broker API
   url: "https://broker.bgpstream.caida.org/v2"
 
 iptoasn:
+  # IPtoASN prefix geography 合并数据下载地址
   combined_url: "https://iptoasn.com/data/ip2asn-combined.tsv.gz"
 
 opengeofeed:
+  # OpenGeoFeed 公共 geofeed CSV
   public_csv_url: "https://opengeofeed.org/feed/public.csv"
 
 nro:
+  # NRO delegated stats 下载地址
   delegated_stats_url: "https://ftp.ripe.net/pub/stats/ripencc/nro-stats/latest/nro-delegated-stats"

backend/app/core/datasource_defaults.py

@@ -155,6 +155,13 @@ DEFAULT_DATASOURCES = {
         "priority": "P1",
         "frequency_minutes": 1440,
     },
+    "news_live_streams": {
+        "id": 26,
+        "name": "News Live Streams",
+        "module": "L4",
+        "priority": "P2",
+        "frequency_minutes": 720,
+    },
 }
 
 ID_TO_COLLECTOR = {info["id"]: name for name, info in DEFAULT_DATASOURCES.items()}

backend/app/db/session.py

@@ -95,6 +95,8 @@ async def init_db():
     import app.models.bgp_observation  # noqa: F401
     import app.models.collected_data  # noqa: F401
     import app.models.system_setting  # noqa: F401
+    import app.models.playground_session  # noqa: F401
+    import app.models.playground_message  # noqa: F401
 
     async with engine.begin() as conn:
         await conn.run_sync(Base.metadata.create_all)

backend/app/models/__init__.py

@@ -9,6 +9,8 @@ from app.models.bgp_anomaly import BGPAnomaly
 from app.models.bgp_incident import BGPIncident
 from app.models.bgp_observation import BGPObservation
 from app.models.system_setting import SystemSetting
+from app.models.playground_session import PlaygroundSession
+from app.models.playground_message import PlaygroundMessage
 
 __all__ = [
     "User",

backend/app/models/playground_message.py

@@ -0,0 +1,40 @@
+from sqlalchemy import JSON, Boolean, Column, DateTime, ForeignKey, Integer, String, Text
+from sqlalchemy.sql import func
+
+from app.db.session import Base
+
+
+class PlaygroundMessage(Base):
+    __tablename__ = "playground_messages"
+
+    id = Column(Integer, primary_key=True, autoincrement=True)
+    public_id = Column(String(64), unique=True, index=True, nullable=False)
+    session_id = Column(Integer, ForeignKey("playground_sessions.id", ondelete="CASCADE"), nullable=False, index=True)
+    user_id = Column(Integer, ForeignKey("users.id", ondelete="CASCADE"), nullable=False, index=True)
+    parent_message_id = Column(Integer, ForeignKey("playground_messages.id", ondelete="SET NULL"), nullable=True)
+    role = Column(String(20), nullable=False)
+    kind = Column(String(20), nullable=False, default="message")
+    status = Column(String(20), nullable=False, default="done")
+    title = Column(String(255), nullable=True)
+    content = Column(Text, nullable=False, default="")
+    thinking_content = Column(Text, nullable=False, default="")
+    meta = Column(JSON, nullable=False, default=list)
+    provider = Column(String(100), nullable=True)
+    model = Column(String(200), nullable=True)
+    request_id = Column(String(100), nullable=True)
+    raw_response = Column(JSON, nullable=False, default=dict)
+    content_blocks = Column(JSON, nullable=False, default=list)
+    text_blocks = Column(JSON, nullable=False, default=list)
+    thinking_blocks = Column(JSON, nullable=False, default=list)
+    sort_order = Column(Integer, nullable=False, default=0, index=True)
+    is_visible = Column(Boolean, nullable=False, default=True)
+    created_at = Column(DateTime(timezone=True), server_default=func.now(), nullable=False)
+    updated_at = Column(
+        DateTime(timezone=True),
+        server_default=func.now(),
+        onupdate=func.now(),
+        nullable=False,
+    )
+
+    def __repr__(self):
+        return f"<PlaygroundMessage public_id={self.public_id} role={self.role} status={self.status}>"

backend/app/models/playground_session.py

@@ -0,0 +1,27 @@
+from sqlalchemy import JSON, Column, DateTime, ForeignKey, Integer, String, UniqueConstraint
+from sqlalchemy.sql import func
+
+from app.db.session import Base
+
+
+class PlaygroundSession(Base):
+    __tablename__ = "playground_sessions"
+    __table_args__ = (
+        UniqueConstraint("user_id", "session_key", name="uq_playground_sessions_user_session_key"),
+    )
+
+    id = Column(Integer, primary_key=True, autoincrement=True)
+    user_id = Column(Integer, ForeignKey("users.id", ondelete="CASCADE"), nullable=False, index=True)
+    session_key = Column(String(100), nullable=False, default="default")
+    title = Column(String(200), nullable=False, default="Playground 会话")
+    state = Column(JSON, nullable=False, default={})
+    created_at = Column(DateTime(timezone=True), server_default=func.now(), nullable=False)
+    updated_at = Column(
+        DateTime(timezone=True),
+        server_default=func.now(),
+        onupdate=func.now(),
+        nullable=False,
+    )
+
+    def __repr__(self):
+        return f"<PlaygroundSession user_id={self.user_id} session_key={self.session_key}>"

backend/app/schemas/ai.py

@@ -3,6 +3,14 @@ from typing import Any
 from pydantic import BaseModel, Field
 
 
+class AIContentBlock(BaseModel):
+    type: str
+    text: str | None = None
+    thinking: str | None = None
+    signature: str | None = None
+    metadata: dict[str, Any] = Field(default_factory=dict)
+
+
 class SituationalAnalysisRequest(BaseModel):
     title: str = Field(..., min_length=1, max_length=200)
     objective: str = Field(..., min_length=1, max_length=1000)
@@ -10,18 +18,158 @@ class SituationalAnalysisRequest(BaseModel):
     observations: list[str] = Field(default_factory=list)
     constraints: list[str] = Field(default_factory=list)
     preferred_model: str | None = Field(default=None, max_length=200)
+    thinking: dict[str, Any] | None = None
+
+
+class BGPBriefRequest(BaseModel):
+    incident_limit: int = Field(default=5, ge=1, le=10)
+    anomaly_limit: int = Field(default=6, ge=1, le=12)
+    collector_limit: int = Field(default=5, ge=1, le=10)
+    preferred_model: str | None = Field(default=None, max_length=200)
+    thinking: dict[str, Any] | None = None
+
+
+class AlertBriefRequest(BaseModel):
+    alert_limit: int = Field(default=8, ge=1, le=20)
+    preferred_model: str | None = Field(default=None, max_length=200)
+    thinking: dict[str, Any] | None = None
+
+
+class SituationalAlertBriefRequest(BaseModel):
+    preferred_model: str | None = Field(default=None, max_length=200)
+    thinking: dict[str, Any] | None = None
 
 
 class SituationalAnalysisResponse(BaseModel):
     provider: str
     model: str
     content: str
+    content_blocks: list[AIContentBlock] = Field(default_factory=list)
+    text_blocks: list[str] = Field(default_factory=list)
+    thinking_blocks: list[str] = Field(default_factory=list)
     raw_response: dict[str, Any] = Field(default_factory=dict)
 
 
+class BGPBriefRecordSummary(BaseModel):
+    id: str
+    title: str
+    provider: str
+    model: str
+    request_id: str | None = None
+    generated_at: str
+
+
+class BGPBriefRecordResponse(BGPBriefRecordSummary):
+    content_markdown: str
+    facts: list[str] = Field(default_factory=list)
+    context: dict[str, Any] = Field(default_factory=dict)
+
+
+class AlertBriefResponse(SituationalAnalysisResponse):
+    title: str
+    objective: str
+    facts: list[str] = Field(default_factory=list)
+    context: dict[str, Any] = Field(default_factory=dict)
+
+
+class SituationalAlertBriefResponse(SituationalAnalysisResponse):
+    title: str
+    objective: str
+    facts: list[str] = Field(default_factory=list)
+    context: dict[str, Any] = Field(default_factory=dict)
+
+
 class AIProviderStatusResponse(BaseModel):
     provider: str
+    api: str | None = None
     enabled: bool
     configured: bool
     model: str | None = None
     base_url: str | None = None
+
+
+class PlaygroundSessionState(BaseModel):
+    messages: list[dict[str, Any]] = Field(default_factory=list)
+    selectedPresetKey: str = Field(default="bgp-brief", max_length=100)
+    title: str = Field(default="", max_length=200)
+    objective: str = Field(default="", max_length=1000)
+    constraints: str = Field(default="")
+    inputValue: str = Field(default="")
+    analysis: dict[str, Any] | None = None
+    latestAnalysisMessageId: str | None = Field(default=None, max_length=200)
+    analysisMeta: dict[str, Any] = Field(default_factory=dict)
+    helpExpanded: bool = True
+
+
+class PlaygroundSessionUpsertRequest(BaseModel):
+    session_key: str = Field(default="default", min_length=1, max_length=100)
+    title: str | None = Field(default=None, max_length=200)
+    state: PlaygroundSessionState
+
+
+class PlaygroundMessageRecord(BaseModel):
+    id: str
+    role: str
+    kind: str = "message"
+    status: str = "done"
+    title: str | None = None
+    content: str = ""
+    thinking_content: str = ""
+    meta: list[str] = Field(default_factory=list)
+    markdown: bool = True
+    provider: str | None = None
+    model: str | None = None
+    request_id: str | None = None
+    raw_response: dict[str, Any] = Field(default_factory=dict)
+    content_blocks: list[dict[str, Any]] = Field(default_factory=list)
+    text_blocks: list[str] = Field(default_factory=list)
+    thinking_blocks: list[str] = Field(default_factory=list)
+    parent_message_id: str | None = None
+    created_at: str
+    updated_at: str
+
+
+class PlaygroundSessionResponse(BaseModel):
+    id: str
+    session_key: str
+    title: str
+    state: PlaygroundSessionState
+    created_at: str
+    updated_at: str
+
+
+class PlaygroundThreadResponse(BaseModel):
+    session: PlaygroundSessionResponse
+    messages: list[PlaygroundMessageRecord] = Field(default_factory=list)
+
+
+class PlaygroundMessageCreateRequest(BaseModel):
+    session_key: str = Field(default="default", min_length=1, max_length=100)
+    title: str = Field(..., min_length=1, max_length=200)
+    objective: str = Field(..., min_length=1, max_length=1000)
+    constraints: str = Field(default="")
+    input: str = Field(..., min_length=1)
+    selected_preset_key: str = Field(default="bgp-brief", max_length=100)
+    help_expanded: bool = True
+
+
+class PlaygroundMessageActionResponse(BaseModel):
+    session: PlaygroundSessionResponse
+    messages: list[PlaygroundMessageRecord] = Field(default_factory=list)
+    active_message_id: str | None = None
+
+
+class PlaygroundMessageStopRequest(BaseModel):
+    session_key: str = Field(default="default", min_length=1, max_length=100)
+    message_id: str = Field(..., min_length=1, max_length=64)
+
+
+class PlaygroundMessageResendRequest(BaseModel):
+    session_key: str = Field(default="default", min_length=1, max_length=100)
+    user_message_id: str = Field(..., min_length=1, max_length=64)
+
+
+class PlaygroundMessageEditRequest(BaseModel):
+    session_key: str = Field(default="default", min_length=1, max_length=100)
+    user_message_id: str = Field(..., min_length=1, max_length=64)
+    content: str = Field(..., min_length=1)

backend/app/schemas/alert.py

@@ -0,0 +1,5 @@
+from pydantic import BaseModel, Field
+
+
+class AlertResolutionRequest(BaseModel):
+    resolution: str = Field(..., min_length=1, max_length=1000)

backend/app/services/alert_ai_brief.py

@@ -0,0 +1,103 @@
+from __future__ import annotations
+
+from collections import Counter
+from typing import Any
+
+from sqlalchemy import func, select
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.models.alert import Alert, AlertSeverity, AlertStatus
+from app.schemas.ai import AlertBriefRequest, SituationalAnalysisRequest
+
+
+def _format_counter(counter: Counter[str], empty_text: str = "无") -> str:
+    if not counter:
+        return empty_text
+    return "，".join(f"{key} {value}" for key, value in counter.items())
+
+
+async def build_alert_brief_request(
+    db: AsyncSession,
+    *,
+    alert_limit: int = 8,
+) -> tuple[SituationalAnalysisRequest, list[str], dict[str, Any]]:
+    recent_alerts_result = await db.execute(
+        select(Alert)
+        .order_by(Alert.created_at.desc(), Alert.id.desc())
+        .limit(max(alert_limit, 1))
+    )
+    total_result = await db.execute(select(func.count(Alert.id)))
+    active_result = await db.execute(select(func.count(Alert.id)).where(Alert.status == AlertStatus.ACTIVE))
+    acknowledged_result = await db.execute(
+        select(func.count(Alert.id)).where(Alert.status == AlertStatus.ACKNOWLEDGED)
+    )
+    resolved_result = await db.execute(select(func.count(Alert.id)).where(Alert.status == AlertStatus.RESOLVED))
+
+    recent_alerts = recent_alerts_result.scalars().all()
+    total_alerts = total_result.scalar() or 0
+    active_alerts = active_result.scalar() or 0
+    acknowledged_alerts = acknowledged_result.scalar() or 0
+    resolved_alerts = resolved_result.scalar() or 0
+
+    severity_counts = Counter((item.severity.value if item.severity else "unknown") for item in recent_alerts)
+    status_counts = Counter((item.status.value if item.status else "unknown") for item in recent_alerts)
+    datasource_counts = Counter((item.datasource_name or "未命名数据源") for item in recent_alerts)
+    active_datasource_counts = Counter(
+        (item.datasource_name or "未命名数据源")
+        for item in recent_alerts
+        if item.status == AlertStatus.ACTIVE
+    )
+
+    facts = [
+        f"告警总量 {total_alerts} 条，其中 active {active_alerts} 条、acknowledged {acknowledged_alerts} 条、resolved {resolved_alerts} 条。",
+        f"最近告警严重度分布：{_format_counter(severity_counts)}。",
+        f"最近告警状态分布：{_format_counter(status_counts)}。",
+        f"最近告警数据源分布：{_format_counter(Counter(dict(datasource_counts.most_common(6))))}。",
+    ]
+
+    if active_datasource_counts:
+        facts.append(
+            "当前待处理告警主要集中在："
+            + _format_counter(Counter(dict(active_datasource_counts.most_common(5))))
+            + "。"
+        )
+
+    if recent_alerts:
+        facts.append(
+            "最近告警摘录："
+            + "；".join(
+                [
+                    f"{item.datasource_name or '未命名数据源'} / {item.severity.value if item.severity else '-'} / {item.status.value if item.status else '-'} / {item.message or '-'}"
+                    for item in recent_alerts[:6]
+                ]
+            )
+        )
+
+    context = {
+        "source": "alerts",
+        "total_alerts": total_alerts,
+        "active_alerts": active_alerts,
+        "acknowledged_alerts": acknowledged_alerts,
+        "resolved_alerts": resolved_alerts,
+        "severity_distribution": dict(severity_counts),
+        "status_distribution": dict(status_counts),
+        "top_datasources": dict(datasource_counts.most_common(6)),
+        "top_active_datasources": dict(active_datasource_counts.most_common(5)),
+    }
+
+    return (
+        SituationalAnalysisRequest(
+            title="告警态势 AI 简报",
+            objective="基于当前告警总量、严重度、状态、数据源分布与最近告警摘录，生成一份面向值班人员的简明告警态势简报，突出待处理风险、告警集中点和优先动作。",
+            observations=facts,
+            constraints=[
+                "明确区分事实、推断与建议。",
+                "优先指出仍处于 active 状态且高严重度的告警簇。",
+                "不要把 acknowledged 或 resolved 告警误判成当前仍在扩大。",
+                "如果证据不足，请明确指出缺失的上下文。",
+            ],
+            context=context,
+        ),
+        facts,
+        context,
+    )

backend/app/services/bgp_ai_brief.py

@@ -0,0 +1,259 @@
+from __future__ import annotations
+
+from collections import Counter
+from typing import Any
+
+from sqlalchemy import func, select
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.api.v1.bgp import BGP_SOURCES
+from app.models.bgp_anomaly import BGPAnomaly
+from app.models.bgp_incident import BGPIncident
+from app.models.bgp_observation import BGPObservation
+from app.schemas.ai import SituationalAnalysisRequest
+from app.services.bgp_collectors import build_bgp_collector_coverage
+from app.services.bgp_enrichment import lookup_prefix_geography
+
+
+def _format_counter(counter: dict[str, int], empty_text: str = "无") -> str:
+    if not counter:
+        return empty_text
+    return "，".join(f"{key} {value}" for key, value in counter.items())
+
+
+def _severity_rank(value: str | None) -> int:
+    order = {
+        "critical": 0,
+        "high": 1,
+        "medium": 2,
+        "low": 3,
+        "info": 4,
+    }
+    return order.get((value or "").lower(), 99)
+
+
+def _normalize_geo_key(country: str | None, city: str | None) -> str:
+    if city and country:
+        return f"{city}, {country}"
+    return city or country or "未知区域"
+
+
+def _top_counter_items(counter: Counter[str], limit: int = 5) -> dict[str, int]:
+    return {name: count for name, count in counter.most_common(limit) if name}
+
+
+def _collect_incident_regions(incidents: list[BGPIncident]) -> Counter[str]:
+    counter: Counter[str] = Counter()
+    for item in incidents:
+        for region in item.affected_regions or []:
+            if not isinstance(region, dict):
+                continue
+            counter[_normalize_geo_key(region.get("country"), region.get("city"))] += 1
+    return counter
+
+
+def _collect_collector_regions(collectors: list[dict[str, Any]]) -> Counter[str]:
+    counter: Counter[str] = Counter()
+    for item in collectors:
+        counter[_normalize_geo_key(item.get("country"), item.get("city"))] += int(item.get("recent_24h_observation_count") or 0)
+    return counter
+
+
+def _format_geo_evidence(prefix_geographies: dict[str, dict[str, Any]], limit: int = 6) -> str:
+    if not prefix_geographies:
+        return "没有命中 prefix geography 证据。"
+
+    rows = []
+    for prefix, item in list(prefix_geographies.items())[:limit]:
+        region = _normalize_geo_key(item.get("country"), item.get("city"))
+        source = item.get("source") or item.get("geography_mode") or "unknown"
+        as_hint = item.get("asn")
+        as_name = item.get("as_name")
+        as_text = ""
+        if as_hint:
+            as_text = f" / ASN AS{as_hint}"
+            if as_name:
+                as_text += f" ({as_name})"
+        rows.append(f"{prefix} -> {region} / 来源 {source}{as_text}")
+    return "；".join(rows)
+
+
+async def build_bgp_brief_request(
+    db: AsyncSession,
+    *,
+    incident_limit: int = 5,
+    anomaly_limit: int = 6,
+    collector_limit: int = 5,
+) -> tuple[SituationalAnalysisRequest, list[str], dict[str, int | str | dict[str, int]]]:
+    incidents_result = await db.execute(
+        select(BGPIncident)
+        .order_by(BGPIncident.created_at.desc(), BGPIncident.id.desc())
+        .limit(max(incident_limit, 1))
+    )
+    anomalies_result = await db.execute(
+        select(BGPAnomaly)
+        .order_by(BGPAnomaly.created_at.desc(), BGPAnomaly.id.desc())
+        .limit(max(anomaly_limit, 1))
+    )
+    observations_result = await db.execute(
+        select(BGPObservation).where(BGPObservation.source.in_(BGP_SOURCES))
+    )
+    incident_count_result = await db.execute(select(func.count(BGPIncident.id)))
+    anomaly_count_result = await db.execute(select(func.count(BGPAnomaly.id)))
+
+    incidents = incidents_result.scalars().all()
+    anomalies = anomalies_result.scalars().all()
+    observations = observations_result.scalars().all()
+    collectors = await build_bgp_collector_coverage(db, source_filter=BGP_SOURCES)
+
+    total_incidents = incident_count_result.scalar() or 0
+    total_anomalies = anomaly_count_result.scalar() or 0
+    total_observations = len(observations)
+    active_collectors = [item for item in collectors if item["observation_count"] > 0]
+
+    incident_status_counts = Counter((item.status or "unknown") for item in incidents)
+    incident_severity_counts = Counter((item.severity or "unknown") for item in incidents)
+    incident_type_counts = Counter((item.incident_type or "unknown") for item in incidents)
+    anomaly_type_counts = Counter((item.anomaly_type or "unknown") for item in anomalies)
+    event_type_counts = Counter((item.event_type or "unknown") for item in observations)
+    incident_region_counts = _collect_incident_regions(incidents)
+
+    top_collectors = sorted(
+        active_collectors,
+        key=lambda item: (
+            -int(item["recent_24h_observation_count"]),
+            -int(item["observation_count"]),
+            str(item["collector"]),
+        ),
+    )[: max(collector_limit, 1)]
+    collector_region_counts = _collect_collector_regions(top_collectors)
+
+    prefix_candidates = sorted(
+        {
+            prefix
+            for item in incidents
+            for prefix in (item.affected_prefixes or [])
+            if prefix
+        }
+        | {item.prefix for item in anomalies if item.prefix}
+    )
+    prefix_geographies = await lookup_prefix_geography(db, prefix_candidates) if prefix_candidates else {}
+    geography_region_counts = Counter(
+        _normalize_geo_key(item.get("country"), item.get("city"))
+        for item in prefix_geographies.values()
+        if item.get("country") or item.get("city")
+    )
+    hotspot_region_counts = geography_region_counts + incident_region_counts
+    collector_bias_regions = [
+        region
+        for region, count in collector_region_counts.most_common(3)
+        if count > hotspot_region_counts.get(region, 0)
+    ]
+
+    observations_lines: list[str] = [
+        f"当前共有 {total_incidents} 起 BGP incidents、{total_anomalies} 条 anomalies、{total_observations} 条原始观测事件。",
+        f"活跃观测站 {len(active_collectors)} 个；近 24 小时事件数合计 {sum(int(item['recent_24h_observation_count']) for item in active_collectors)}。",
+        f"最近 incidents 严重度分布：{_format_counter(dict(sorted(incident_severity_counts.items(), key=lambda item: _severity_rank(item[0]))))}。",
+        f"最近 incidents 状态分布：{_format_counter(dict(incident_status_counts))}。",
+        f"最近 incidents 类型分布：{_format_counter(dict(incident_type_counts.most_common(5)))}。",
+        f"最近 anomalies 类型分布：{_format_counter(dict(anomaly_type_counts.most_common(6)))}。",
+        f"观测事件类型分布：{_format_counter(dict(event_type_counts.most_common(6)))}。",
+    ]
+
+    if hotspot_region_counts:
+        observations_lines.append(
+            "区域热点事实层："
+            + _format_counter(_top_counter_items(hotspot_region_counts, limit=5), empty_text="无明显区域聚集")
+            + "。"
+        )
+
+    if prefix_geographies:
+        observations_lines.append("Prefix geography 证据：" + _format_geo_evidence(prefix_geographies))
+
+    if collector_bias_regions:
+        observations_lines.append(
+            "观测偏差提示：重点观测站最近 24h 活跃度更集中在 "
+            + "、".join(collector_bias_regions)
+            + "，这些区域的事件升温结论需要结合 prefix geography 与 affected regions 交叉验证。"
+        )
+    elif top_collectors:
+        observations_lines.append(
+            "观测偏差提示：当前未发现明显高于区域热点事实层的单一观测站集中区域，但仍需区分 collector coverage 与真实区域风险。"
+        )
+
+    if incidents:
+        observations_lines.append(
+            "最近 incident 摘要：" + "；".join(
+                [
+                    f"{item.incident_type} / {item.severity} / {item.status}"
+                    f" / 前缀 {', '.join(item.affected_prefixes[:2]) if item.affected_prefixes else '-'}"
+                    f" / 观测站 {len(item.affected_collectors or [])} 个"
+                    for item in incidents
+                ]
+            )
+        )
+
+    if anomalies:
+        observations_lines.append(
+            "最近 anomaly 摘要：" + "；".join(
+                [
+                    f"{item.anomaly_type} / {item.severity}"
+                    f" / 前缀 {item.prefix or '-'}"
+                    f" / ASN {item.new_origin_asn or item.origin_asn or '-'}"
+                    for item in anomalies
+                ]
+            )
+        )
+
+    if top_collectors:
+        observations_lines.append(
+            "重点观测站：" + "；".join(
+                [
+                    f"{item['collector']} ({', '.join([part for part in [item.get('city'), item.get('country')] if part]) or '未知位置'})"
+                    f" / 近24h {item['recent_24h_observation_count']} 条"
+                    f" / 前缀 {item['prefix_count']} 个"
+                    for item in top_collectors
+                ]
+            )
+        )
+
+    context = {
+        "source": "bgp-overview",
+        "incident_total": total_incidents,
+        "anomaly_total": total_anomalies,
+        "observation_total": total_observations,
+        "active_collectors": len(active_collectors),
+        "top_incident_types": dict(incident_type_counts.most_common(5)),
+        "top_anomaly_types": dict(anomaly_type_counts.most_common(6)),
+        "top_event_types": dict(event_type_counts.most_common(6)),
+        "region_hotspots": _top_counter_items(hotspot_region_counts, limit=6),
+        "incident_regions": _top_counter_items(incident_region_counts, limit=6),
+        "collector_bias_regions": collector_bias_regions,
+        "prefix_geography_sources": dict(
+            Counter(str(item.get("source") or "unknown") for item in prefix_geographies.values()).most_common(5)
+        ),
+        "prefix_geography_sample": {
+            prefix: {
+                "country": item.get("country"),
+                "city": item.get("city"),
+                "source": item.get("source"),
+                "asn": item.get("asn"),
+                "as_name": item.get("as_name"),
+            }
+            for prefix, item in list(prefix_geographies.items())[:8]
+        },
+    }
+
+    return SituationalAnalysisRequest(
+        title="BGP 态势 AI 简报",
+        objective="基于当前 BGP incidents、anomalies、原始观测事件、观测站覆盖与 prefix geography 证据，生成一份面向操作员的简明态势简报，突出区域热点、观测偏差、当前风险、证据和优先动作。",
+        observations=observations_lines,
+        constraints=[
+            "明确区分事实、推断与建议。",
+            "优先指出需要立即关注的高严重度 incident 或异常模式。",
+            "需要单独指出哪些区域结论来自 prefix geography / affected regions，哪些可能受 collector coverage 偏差影响。",
+            "结论应服务值班排障，不要写成泛泛的模型演示文案。",
+            "如果证据不足，要明确指出缺失数据。",
+        ],
+        context=context,
+    ), observations_lines, context

backend/app/services/bgp_ai_brief_store.py

@@ -0,0 +1,160 @@
+from __future__ import annotations
+
+import json
+from dataclasses import dataclass
+from datetime import UTC, datetime
+from pathlib import Path
+from typing import Any
+from uuid import uuid4
+
+from app.core.config import ROOT_DIR
+from app.schemas.ai import BGPBriefRecordResponse, BGPBriefRecordSummary, SituationalAnalysisResponse
+
+
+_BRIEF_STORAGE_DIR = ROOT_DIR / "data" / "ai" / "bgp-briefs"
+_METADATA_PREFIX = "<!-- planet-bgp-brief-meta "
+_METADATA_SUFFIX = " -->"
+_BRIEF_TITLE = "BGP AI 简报"
+
+
+@dataclass(slots=True)
+class _StoredBrief:
+    id: str
+    title: str
+    provider: str
+    model: str
+    request_id: str | None
+    generated_at: str
+    content_markdown: str
+    facts: list[str]
+    context: dict[str, Any]
+    path: Path
+
+
+def _ensure_storage_dir() -> Path:
+    _BRIEF_STORAGE_DIR.mkdir(parents=True, exist_ok=True)
+    return _BRIEF_STORAGE_DIR
+
+
+def _build_metadata_line(metadata: dict[str, Any]) -> str:
+    return f"{_METADATA_PREFIX}{json.dumps(metadata, ensure_ascii=False)}{_METADATA_SUFFIX}"
+
+
+def _parse_brief_file(path: Path) -> _StoredBrief | None:
+    try:
+        raw_text = path.read_text(encoding="utf-8")
+    except OSError:
+        return None
+
+    first_line, separator, remainder = raw_text.partition("\n")
+    if not separator or not first_line.startswith(_METADATA_PREFIX) or not first_line.endswith(_METADATA_SUFFIX):
+        return None
+
+    metadata_payload = first_line[len(_METADATA_PREFIX) : -len(_METADATA_SUFFIX)]
+
+    try:
+        metadata = json.loads(metadata_payload)
+    except json.JSONDecodeError:
+        return None
+
+    return _StoredBrief(
+        id=str(metadata.get("id") or path.stem),
+        title=str(metadata.get("title") or _BRIEF_TITLE),
+        provider=str(metadata.get("provider") or "-"),
+        model=str(metadata.get("model") or "-"),
+        request_id=metadata.get("request_id"),
+        generated_at=str(metadata.get("generated_at") or datetime.fromtimestamp(path.stat().st_mtime, UTC).isoformat()),
+        content_markdown=remainder.lstrip("\n"),
+        facts=list(metadata.get("facts") or []),
+        context=dict(metadata.get("context") or {}),
+        path=path,
+    )
+
+
+def list_bgp_brief_records(limit: int = 50) -> list[BGPBriefRecordSummary]:
+    storage_dir = _ensure_storage_dir()
+    records: list[_StoredBrief] = []
+
+    for path in storage_dir.glob("*.md"):
+        parsed = _parse_brief_file(path)
+        if parsed is not None:
+            records.append(parsed)
+
+    records.sort(key=lambda item: item.generated_at, reverse=True)
+
+    return [
+        BGPBriefRecordSummary(
+            id=item.id,
+            title=item.title,
+            provider=item.provider,
+            model=item.model,
+            request_id=item.request_id,
+            generated_at=item.generated_at,
+        )
+        for item in records[: max(limit, 1)]
+    ]
+
+
+def get_bgp_brief_record(brief_id: str) -> BGPBriefRecordResponse | None:
+    path = _ensure_storage_dir() / f"{brief_id}.md"
+    parsed = _parse_brief_file(path)
+    if parsed is None:
+        return None
+
+    return BGPBriefRecordResponse(
+        id=parsed.id,
+        title=parsed.title,
+        provider=parsed.provider,
+        model=parsed.model,
+        request_id=parsed.request_id,
+        generated_at=parsed.generated_at,
+        content_markdown=parsed.content_markdown,
+        facts=parsed.facts,
+        context=parsed.context,
+    )
+
+
+def get_latest_bgp_brief_record() -> BGPBriefRecordResponse | None:
+    summaries = list_bgp_brief_records(limit=1)
+    if not summaries:
+        return None
+    return get_bgp_brief_record(summaries[0].id)
+
+
+def save_bgp_brief_record(
+    analysis: SituationalAnalysisResponse,
+    *,
+    request_id: str | None,
+    facts: list[str] | None = None,
+    context: dict[str, Any] | None = None,
+    generated_at: datetime | None = None,
+) -> BGPBriefRecordResponse:
+    created_at = generated_at or datetime.now(UTC)
+    brief_id = f"{created_at.strftime('%Y%m%dT%H%M%SZ')}-{uuid4().hex[:8]}"
+    path = _ensure_storage_dir() / f"{brief_id}.md"
+
+    metadata = {
+        "id": brief_id,
+        "title": _BRIEF_TITLE,
+        "provider": analysis.provider,
+        "model": analysis.model,
+        "request_id": request_id,
+        "generated_at": created_at.isoformat(),
+        "facts": facts or [],
+        "context": context or {},
+    }
+
+    markdown_text = f"{_build_metadata_line(metadata)}\n\n{analysis.content.rstrip()}\n"
+    path.write_text(markdown_text, encoding="utf-8")
+
+    return BGPBriefRecordResponse(
+        id=brief_id,
+        title=_BRIEF_TITLE,
+        provider=analysis.provider,
+        model=analysis.model,
+        request_id=request_id,
+        generated_at=created_at.isoformat(),
+        content_markdown=analysis.content,
+        facts=facts or [],
+        context=context or {},
+    )

backend/app/services/bgp_collectors.py

@@ -6,7 +6,7 @@ from collections import defaultdict
 from datetime import UTC, datetime, timedelta
 from typing import Any
 
-from sqlalchemy import select
+from sqlalchemy import case, distinct, func, select
 from sqlalchemy.ext.asyncio import AsyncSession
 
 from app.core.time import to_iso8601_utc
@@ -14,6 +14,16 @@ from app.models.bgp_observation import BGPObservation
 from app.services.collectors.bgp_common import RIPE_RIS_COLLECTOR_COORDS
 
 
+def _collector_base_filters(source_filter: tuple[str, ...] | None) -> list[Any]:
+    filters: list[Any] = [
+        BGPObservation.collector.isnot(None),
+        func.length(func.btrim(BGPObservation.collector)) > 0,
+    ]
+    if source_filter:
+        filters.append(BGPObservation.source.in_(source_filter))
+    return filters
+
+
 async def build_bgp_collector_coverage(
     db: AsyncSession,
     *,
@@ -24,88 +34,148 @@ async def build_bgp_collector_coverage(
     recent_24h_threshold = now - timedelta(hours=24)
     recent_7d_threshold = now - timedelta(days=7)
 
-    stmt = select(BGPObservation).order_by(BGPObservation.observed_at.desc(), BGPObservation.id.desc())
-    if source_filter:
-        stmt = stmt.where(BGPObservation.source.in_(source_filter))
+    filters = _collector_base_filters(source_filter)
+    country_expr = func.nullif(BGPObservation.collector_geo["country"].as_string(), "")
+    city_expr = func.nullif(BGPObservation.collector_geo["city"].as_string(), "")
 
-    result = await db.execute(stmt)
-    records = list(result.scalars().all())
+    aggregate_stmt = (
+        select(
+            BGPObservation.collector.label("collector"),
+            func.count(BGPObservation.id).label("observation_count"),
+            func.count(distinct(BGPObservation.prefix)).label("prefix_count"),
+            func.count(distinct(BGPObservation.origin_asn)).label("origin_asn_count"),
+            func.count(distinct(BGPObservation.peer_asn)).label("peer_asn_count"),
+            func.sum(case((BGPObservation.observed_at >= recent_15m_threshold, 1), else_=0)).label("recent_15m_observation_count"),
+            func.sum(case((BGPObservation.observed_at >= recent_24h_threshold, 1), else_=0)).label("recent_24h_observation_count"),
+            func.sum(case((BGPObservation.observed_at >= recent_7d_threshold, 1), else_=0)).label("recent_7d_observation_count"),
+            func.count(distinct(case((BGPObservation.observed_at >= recent_15m_threshold, BGPObservation.prefix), else_=None))).label("recent_15m_prefix_count"),
+            func.count(distinct(case((BGPObservation.observed_at >= recent_24h_threshold, BGPObservation.prefix), else_=None))).label("recent_24h_prefix_count"),
+            func.count(distinct(case((BGPObservation.observed_at >= recent_7d_threshold, BGPObservation.prefix), else_=None))).label("recent_7d_prefix_count"),
+            func.max(BGPObservation.observed_at).label("latest_observed_at"),
+        )
+        .where(*filters)
+        .group_by(BGPObservation.collector)
+    )
+    aggregate_rows = (await db.execute(aggregate_stmt)).all()
+
+    latest_subquery = (
+        select(
+            BGPObservation.collector.label("collector"),
+            BGPObservation.event_type.label("latest_event_type"),
+            country_expr.label("country"),
+            city_expr.label("city"),
+            func.row_number()
+            .over(
+                partition_by=BGPObservation.collector,
+                order_by=(BGPObservation.observed_at.desc(), BGPObservation.id.desc()),
+            )
+            .label("rn"),
+        )
+        .where(*filters)
+        .subquery()
+    )
+    latest_rows = (
+        await db.execute(
+            select(
+                latest_subquery.c.collector,
+                latest_subquery.c.latest_event_type,
+                latest_subquery.c.country,
+                latest_subquery.c.city,
+            ).where(latest_subquery.c.rn == 1)
+        )
+    ).all()
+
+    event_counts_subquery = (
+        select(
+            BGPObservation.collector.label("collector"),
+            BGPObservation.event_type.label("event_type"),
+            func.count(BGPObservation.id).label("count"),
+            func.row_number()
+            .over(
+                partition_by=BGPObservation.collector,
+                order_by=(func.count(BGPObservation.id).desc(), BGPObservation.event_type.asc()),
+            )
+            .label("rn"),
+        )
+        .where(*filters)
+        .group_by(BGPObservation.collector, BGPObservation.event_type)
+        .subquery()
+    )
+    top_event_rows = (
+        await db.execute(
+            select(
+                event_counts_subquery.c.collector,
+                event_counts_subquery.c.event_type,
+                event_counts_subquery.c.count,
+            ).where(event_counts_subquery.c.rn <= 3)
+        )
+    ).all()
+
+    scope_rows = (
+        await db.execute(
+            select(
+                BGPObservation.collector.label("collector"),
+                country_expr.label("country"),
+                city_expr.label("city"),
+            )
+            .where(*filters)
+            .distinct()
+        )
+    ).all()
+
+    latest_by_collector = {
+        row.collector: {
+            "latest_event_type": row.latest_event_type,
+            "country": row.country,
+            "city": row.city,
+        }
+        for row in latest_rows
+    }
+
+    scope_by_collector: dict[str, dict[str, set[str]]] = defaultdict(lambda: {"countries": set(), "cities": set()})
+    for row in scope_rows:
+        if row.country:
+            scope_by_collector[row.collector]["countries"].add(row.country)
+        if row.city:
+            scope_by_collector[row.collector]["cities"].add(row.city)
+
+    top_events_by_collector: dict[str, list[dict[str, Any]]] = defaultdict(list)
+    for row in top_event_rows:
+        top_events_by_collector[row.collector].append(
+            {"event_type": row.event_type, "count": row.count}
+        )
 
     by_collector: dict[str, dict[str, Any]] = {}
-    for record in records:
-        collector = str(record.collector or "").strip()
-        if not collector:
-            continue
+    for row in aggregate_rows:
+        collector = row.collector
+        latest = latest_by_collector.get(collector, {})
+        fallback_location = RIPE_RIS_COLLECTOR_COORDS.get(collector, {})
+        scope = scope_by_collector.get(collector, {"countries": set(), "cities": set()})
 
-        coverage = by_collector.get(collector)
-        if coverage is None:
-            location = record.collector_geo or RIPE_RIS_COLLECTOR_COORDS.get(collector, {})
-            coverage = {
-                "collector": collector,
-                "city": location.get("city"),
-                "country": location.get("country"),
-                "latitude": location.get("latitude"),
-                "longitude": location.get("longitude"),
-                "observation_count": 0,
-                "prefixes": set(),
-                "origin_asns": set(),
-                "peer_asns": set(),
-                "event_types": defaultdict(int),
-                "countries": set(),
-                "cities": set(),
-                "recent_15m_observation_count": 0,
-                "recent_24h_observation_count": 0,
-                "recent_7d_observation_count": 0,
-                "recent_15m_prefixes": set(),
-                "recent_24h_prefixes": set(),
-                "recent_7d_prefixes": set(),
-                "latest_observed_at": None,
-                "latest_event_type": None,
-            }
-            by_collector[collector] = coverage
-
-        coverage["observation_count"] += 1
-        if record.prefix:
-            coverage["prefixes"].add(record.prefix)
-        if record.origin_asn is not None:
-            coverage["origin_asns"].add(record.origin_asn)
-        if record.peer_asn is not None:
-            coverage["peer_asns"].add(record.peer_asn)
-        if record.event_type:
-            coverage["event_types"][record.event_type] += 1
-
-        observed_at = record.observed_at
-        if observed_at is not None:
-            aware_observed_at = (
-                observed_at.astimezone(UTC)
-                if observed_at.tzinfo
-                else observed_at.replace(tzinfo=UTC)
-            )
-            if aware_observed_at >= recent_15m_threshold:
-                coverage["recent_15m_observation_count"] += 1
-                if record.prefix:
-                    coverage["recent_15m_prefixes"].add(record.prefix)
-            if aware_observed_at >= recent_24h_threshold:
-                coverage["recent_24h_observation_count"] += 1
-                if record.prefix:
-                    coverage["recent_24h_prefixes"].add(record.prefix)
-            if aware_observed_at >= recent_7d_threshold:
-                coverage["recent_7d_observation_count"] += 1
-                if record.prefix:
-                    coverage["recent_7d_prefixes"].add(record.prefix)
-
-        geo = record.collector_geo or {}
-        if geo.get("country"):
-            coverage["countries"].add(geo["country"])
-        if geo.get("city"):
-            coverage["cities"].add(geo["city"])
-
-        current_latest = coverage["latest_observed_at"]
-        if current_latest is None or (
-            record.observed_at is not None and record.observed_at > current_latest
-        ):
-            coverage["latest_observed_at"] = record.observed_at
-            coverage["latest_event_type"] = record.event_type
+        by_collector[collector] = {
+            "collector": collector,
+            "city": latest.get("city") or fallback_location.get("city"),
+            "country": latest.get("country") or fallback_location.get("country"),
+            "latitude": fallback_location.get("latitude"),
+            "longitude": fallback_location.get("longitude"),
+            "observation_count": row.observation_count or 0,
+            "prefix_count": row.prefix_count or 0,
+            "origin_asn_count": row.origin_asn_count or 0,
+            "peer_asn_count": row.peer_asn_count or 0,
+            "recent_15m_observation_count": row.recent_15m_observation_count or 0,
+            "recent_24h_observation_count": row.recent_24h_observation_count or 0,
+            "recent_7d_observation_count": row.recent_7d_observation_count or 0,
+            "recent_15m_prefix_count": row.recent_15m_prefix_count or 0,
+            "recent_24h_prefix_count": row.recent_24h_prefix_count or 0,
+            "recent_7d_prefix_count": row.recent_7d_prefix_count or 0,
+            "top_event_types": top_events_by_collector.get(collector, []),
+            "latest_observed_at": to_iso8601_utc(row.latest_observed_at),
+            "latest_event_type": latest.get("latest_event_type"),
+            "baseline_scope": {
+                "countries": sorted(scope["countries"]),
+                "cities": sorted(scope["cities"]),
+            },
+        }
 
     for collector, location in RIPE_RIS_COLLECTOR_COORDS.items():
         if collector in by_collector:
@@ -117,57 +187,22 @@ async def build_bgp_collector_coverage(
             "latitude": location.get("latitude"),
             "longitude": location.get("longitude"),
             "observation_count": 0,
-            "prefixes": set(),
-            "origin_asns": set(),
-            "peer_asns": set(),
-            "event_types": defaultdict(int),
-            "countries": {location.get("country")} if location.get("country") else set(),
-            "cities": {location.get("city")} if location.get("city") else set(),
+            "prefix_count": 0,
+            "origin_asn_count": 0,
+            "peer_asn_count": 0,
             "recent_15m_observation_count": 0,
             "recent_24h_observation_count": 0,
             "recent_7d_observation_count": 0,
-            "recent_15m_prefixes": set(),
-            "recent_24h_prefixes": set(),
-            "recent_7d_prefixes": set(),
+            "recent_15m_prefix_count": 0,
+            "recent_24h_prefix_count": 0,
+            "recent_7d_prefix_count": 0,
+            "top_event_types": [],
             "latest_observed_at": None,
             "latest_event_type": None,
+            "baseline_scope": {
+                "countries": [location["country"]] if location.get("country") else [],
+                "cities": [location["city"]] if location.get("city") else [],
+            },
         }
 
-    results: list[dict[str, Any]] = []
-    for collector in sorted(by_collector.keys()):
-        item = by_collector[collector]
-        top_event_types = sorted(
-            item["event_types"].items(),
-            key=lambda pair: (-pair[1], pair[0]),
-        )
-        results.append(
-            {
-                "collector": item["collector"],
-                "city": item["city"],
-                "country": item["country"],
-                "latitude": item["latitude"],
-                "longitude": item["longitude"],
-                "observation_count": item["observation_count"],
-                "prefix_count": len(item["prefixes"]),
-                "origin_asn_count": len(item["origin_asns"]),
-                "peer_asn_count": len(item["peer_asns"]),
-                "recent_15m_observation_count": item["recent_15m_observation_count"],
-                "recent_24h_observation_count": item["recent_24h_observation_count"],
-                "recent_7d_observation_count": item["recent_7d_observation_count"],
-                "recent_15m_prefix_count": len(item["recent_15m_prefixes"]),
-                "recent_24h_prefix_count": len(item["recent_24h_prefixes"]),
-                "recent_7d_prefix_count": len(item["recent_7d_prefixes"]),
-                "top_event_types": [
-                    {"event_type": event_type, "count": count}
-                    for event_type, count in top_event_types[:3]
-                ],
-                "latest_observed_at": to_iso8601_utc(item["latest_observed_at"]),
-                "latest_event_type": item["latest_event_type"],
-                "baseline_scope": {
-                    "countries": sorted(country for country in item["countries"] if country),
-                    "cities": sorted(city for city in item["cities"] if city),
-                },
-            }
-        )
-
-    return results
+    return [by_collector[collector] for collector in sorted(by_collector.keys())]

backend/app/services/bgp_enrichment.py

@@ -7,7 +7,7 @@ from collections import defaultdict
 from datetime import UTC, datetime
 from typing import Any
 
-from sqlalchemy import select, text
+from sqlalchemy import Integer, cast, select, text
 from sqlalchemy.ext.asyncio import AsyncSession
 
 from app.core.countries import get_country_centroid, normalize_country
@@ -231,6 +231,13 @@ async def _lookup_prefix_geography(
     return results
 
 
+async def lookup_prefix_geography(
+    db: AsyncSession,
+    prefix_values: list[str],
+) -> dict[str, dict[str, Any]]:
+    return await _lookup_prefix_geography(db, prefix_values)
+
+
 async def enrich_bgp_events_for_batch(
     db: AsyncSession,
     *,
@@ -261,29 +268,40 @@ async def enrich_bgp_events_for_batch(
     historical_prefix_baseline: dict[str, dict[str, Any]] = {}
     if prefix_values:
         previous_result = await db.execute(
-            select(BGPObservation).where(
+            select(
+                BGPObservation.prefix,
+                BGPObservation.origin_asn,
+                BGPObservation.collector,
+                BGPObservation.collector_geo,
+            ).where(
                 BGPObservation.source == source,
                 BGPObservation.prefix.in_(prefix_values),
             )
         )
-        by_prefix: defaultdict[str, list[BGPObservation]] = defaultdict(list)
-        for observation in previous_result.scalars().all():
-            if observation.prefix:
-                by_prefix[observation.prefix].append(observation)
+        by_prefix: defaultdict[str, list[dict[str, Any]]] = defaultdict(list)
+        for prefix, origin_asn, collector, collector_geo in previous_result.all():
+            if prefix:
+                by_prefix[str(prefix)].append(
+                    {
+                        "origin_asn": origin_asn,
+                        "collector": collector,
+                        "collector_geo": collector_geo or {},
+                    }
+                )
 
         for prefix, observations in by_prefix.items():
             unique_origins = sorted(
                 {
-                    observation.origin_asn
+                    observation["origin_asn"]
                     for observation in observations
-                    if observation.origin_asn is not None
+                    if observation["origin_asn"] is not None
                 }
             )
             unique_collectors = sorted(
                 {
-                    observation.collector
+                    observation["collector"]
                     for observation in observations
-                    if observation.collector
+                    if observation["collector"]
                 }
             )
             historical_prefix_baseline[prefix] = {
@@ -292,9 +310,9 @@ async def enrich_bgp_events_for_batch(
                 "historical_observation_count": len(observations),
                 "historical_regions": _compact_locations(
                     [
-                        observation.collector_geo or {}
+                        observation["collector_geo"] or {}
                         for observation in observations
-                        if observation.collector_geo
+                        if observation["collector_geo"]
                     ]
                 ),
             }
@@ -303,7 +321,13 @@ async def enrich_bgp_events_for_batch(
     prefix_geographies = await _lookup_prefix_geography(db, prefix_values) if prefix_values else {}
     if origin_asns:
         peeringdb_result = await db.execute(
-            select(CollectedData).where(CollectedData.source == "peeringdb_network")
+            select(CollectedData)
+            .where(CollectedData.source == "peeringdb_network")
+            .where(CollectedData.is_current.is_(True))
+            .where(
+                cast(CollectedData.extra_data["asn"].as_string(), Integer).in_(origin_asns),
+            )
+            .order_by(CollectedData.id.desc())
         )
         for record in peeringdb_result.scalars().all():
             metadata = record.extra_data or {}

backend/app/services/bgp_incidents.py

@@ -48,14 +48,36 @@ def _collector_regions_from_anomaly(anomaly: BGPAnomaly) -> list[dict]:
     return collected
 
 
-def _dedupe_collected_records(records: list[CollectedData]) -> list[CollectedData]:
-    latest_by_key: dict[str, CollectedData] = {}
-    for record in records:
-        dedupe_key = str(record.source_id or record.entity_key or record.name or record.id)
-        existing = latest_by_key.get(dedupe_key)
-        if existing is None or (record.id or 0) > (existing.id or 0):
-            latest_by_key[dedupe_key] = record
-    return list(latest_by_key.values())
+async def _load_current_infrastructure_records(
+    db: AsyncSession,
+) -> tuple[list[CollectedData], list[CollectedData], list[CollectedData]]:
+    result = await db.execute(
+        select(CollectedData)
+        .where(
+            CollectedData.source.in_(
+                (
+                    "arcgis_landing_points",
+                    "arcgis_cable_landing_relation",
+                    "arcgis_cables",
+                )
+            )
+        )
+        .where(CollectedData.is_current.is_(True))
+        .order_by(CollectedData.source.asc(), CollectedData.id.desc())
+    )
+    grouped_records = {
+        "arcgis_landing_points": [],
+        "arcgis_cable_landing_relation": [],
+        "arcgis_cables": [],
+    }
+    for record in result.scalars().all():
+        grouped_records.setdefault(record.source, []).append(record)
+
+    return (
+        grouped_records["arcgis_landing_points"],
+        grouped_records["arcgis_cable_landing_relation"],
+        grouped_records["arcgis_cables"],
+    )
 
 
 async def infer_related_infrastructure(
@@ -75,19 +97,9 @@ async def infer_related_infrastructure(
     if not valid_regions:
         return {"related_cables": [], "related_ixps": []}
 
-    landing_result = await db.execute(
-        select(CollectedData).where(CollectedData.source == "arcgis_landing_points")
+    landing_records, relation_records, cable_records = await _load_current_infrastructure_records(
+        db,
     )
-    relation_result = await db.execute(
-        select(CollectedData).where(CollectedData.source == "arcgis_cable_landing_relation")
-    )
-    cable_result = await db.execute(
-        select(CollectedData).where(CollectedData.source == "arcgis_cables")
-    )
-
-    landing_records = _dedupe_collected_records(list(landing_result.scalars().all()))
-    relation_records = _dedupe_collected_records(list(relation_result.scalars().all()))
-    cable_records = _dedupe_collected_records(list(cable_result.scalars().all()))
 
     city_to_cable_ids: dict[int, list[int]] = {}
     for relation in relation_records:

backend/app/services/collectors/__init__.py

@@ -35,6 +35,7 @@ from app.services.collectors.bgpstream import BGPStreamBackfillCollector
 from app.services.collectors.iptoasn import IPtoASNPrefixGeoCollector
 from app.services.collectors.opengeofeed import OpenGeoFeedPrefixGeoCollector
 from app.services.collectors.nro_delegated import NRODelegatedPrefixGeoCollector
+from app.services.collectors.news_live_streams import NewsLiveStreamsCollector
 
 collector_registry.register(TOP500Collector())
 collector_registry.register(EpochAIGPUCollector())
@@ -61,3 +62,4 @@ collector_registry.register(BGPStreamBackfillCollector())
 collector_registry.register(IPtoASNPrefixGeoCollector())
 collector_registry.register(OpenGeoFeedPrefixGeoCollector())
 collector_registry.register(NRODelegatedPrefixGeoCollector())
+collector_registry.register(NewsLiveStreamsCollector())

backend/app/services/collectors/base.py

@@ -1,5 +1,6 @@
 """Base collector class for all data sources"""
 
+import asyncio
 from abc import ABC, abstractmethod
 from typing import Dict, List, Any, Optional
 from datetime import UTC, datetime
@@ -166,6 +167,59 @@ class BaseCollector(ABC):
         await db.commit()
         return snapshot.id
 
+    async def _rollback_incomplete_run(
+        self,
+        db: AsyncSession,
+        *,
+        task_id: int,
+        snapshot_id: Optional[int],
+        reason: str,
+    ) -> None:
+        from app.models.collected_data import CollectedData
+        from app.models.data_snapshot import DataSnapshot
+
+        await db.execute(CollectedData.__table__.delete().where(CollectedData.task_id == task_id))
+
+        parent_snapshot_id: Optional[int] = None
+        if snapshot_id is not None:
+            snapshot = await db.get(DataSnapshot, snapshot_id)
+            if snapshot:
+                parent_snapshot_id = snapshot.parent_snapshot_id
+                snapshot.status = "cancelled"
+                snapshot.is_current = False
+                snapshot.completed_at = datetime.now(UTC)
+                summary = dict(snapshot.summary or {})
+                summary["rollback"] = True
+                summary["rollback_reason"] = reason
+                snapshot.summary = summary
+
+        await db.execute(
+            text(
+                """
+                UPDATE collected_data
+                SET is_current = FALSE
+                WHERE source = :source
+                """
+            ),
+            {"source": self.name},
+        )
+
+        if parent_snapshot_id is not None:
+            parent_snapshot = await db.get(DataSnapshot, parent_snapshot_id)
+            if parent_snapshot:
+                parent_snapshot.is_current = True
+
+            await db.execute(
+                text(
+                    """
+                    UPDATE collected_data
+                    SET is_current = TRUE
+                    WHERE snapshot_id = :snapshot_id
+                    """
+                ),
+                {"snapshot_id": parent_snapshot_id},
+            )
+
     async def run(self, db: AsyncSession) -> Dict[str, Any]:
         """Full pipeline: fetch -> transform -> save"""
         from app.services.collectors.registry import collector_registry
@@ -227,7 +281,24 @@ class BaseCollector(ABC):
                 "records_processed": records_count,
                 "execution_time_seconds": (datetime.now(UTC) - start_time).total_seconds(),
             }
+        except asyncio.CancelledError:
+            await db.rollback()
+            task.status = "cancelled"
+            task.phase = "cancelled"
+            task.error_message = "Collection cancelled by operator and rolled back"
+            task.completed_at = datetime.now(UTC)
+            if snapshot_id is not None:
+                await self._rollback_incomplete_run(
+                    db,
+                    task_id=task_id,
+                    snapshot_id=snapshot_id,
+                    reason="cancelled_by_operator",
+                )
+            await db.commit()
+            await self._publish_task_update(force=True)
+            raise
         except Exception as e:
+            await db.rollback()
             task.status = "failed"
             task.phase = "failed"
             task.error_message = str(e)
@@ -276,20 +347,34 @@ class BaseCollector(ABC):
         updated_count = 0
         unchanged_count = 0
         seen_entity_keys: set[str] = set()
-        previous_current_keys: set[str] = set()
+        progress_commit_interval = 1000
 
         previous_current_result = await db.execute(
-            select(CollectedData.entity_key).where(
+            select(CollectedData)
+            .where(
                 CollectedData.source == self.name,
                 CollectedData.is_current == True,
             )
+            .order_by(CollectedData.entity_key.asc(), CollectedData.collected_at.desc().nullslast(), CollectedData.id.desc())
         )
-        previous_current_keys = {row[0] for row in previous_current_result.fetchall() if row[0]}
+        previous_current_records = previous_current_result.scalars().all()
+        previous_current_keys = {record.entity_key for record in previous_current_records if record.entity_key}
+        previous_current_map: dict[str, CollectedData] = {}
+        stale_previous_records: list[CollectedData] = []
+
+        for existing_record in previous_current_records:
+            entity_key = existing_record.entity_key
+            if not entity_key:
+                continue
+            if entity_key not in previous_current_map:
+                previous_current_map[entity_key] = existing_record
+                continue
+            stale_previous_records.append(existing_record)
+
+        for stale_record in stale_previous_records:
+            stale_record.is_current = False
 
         for i, item in enumerate(data):
-            print(
-                f"DEBUG: Saving item {i}: name={item.get('name')}, metadata={item.get('metadata', 'NOT FOUND')}"
-            )
             raw_metadata = item.get("metadata", {})
             extra_data = build_dynamic_metadata(
                 raw_metadata,
@@ -318,20 +403,9 @@ class BaseCollector(ABC):
             previous_record = None
 
             if entity_key and entity_key not in seen_entity_keys:
-                result = await db.execute(
-                    select(CollectedData)
-                    .where(
-                        CollectedData.source == self.name,
-                        CollectedData.entity_key == entity_key,
-                        CollectedData.is_current == True,
-                    )
-                    .order_by(CollectedData.collected_at.desc().nullslast(), CollectedData.id.desc())
-                )
-                previous_records = result.scalars().all()
-                if previous_records:
-                    previous_record = previous_records[0]
-                    for old_record in previous_records:
-                        old_record.is_current = False
+                previous_record = previous_current_map.get(entity_key)
+                if previous_record is not None:
+                    previous_record.is_current = False
 
             record = CollectedData(
                 snapshot_id=snapshot_id,
@@ -375,7 +449,7 @@ class BaseCollector(ABC):
             seen_entity_keys.add(entity_key)
             records_added += 1
 
-            if i % 100 == 0:
+            if (i + 1) % progress_commit_interval == 0:
                 await self.update_progress(i + 1, commit=True)
 
         if snapshot_id is not None:

backend/app/services/collectors/celestrak.py

@@ -21,7 +21,7 @@ class CelesTrakTLECollector(BaseCollector):
 
     @property
     def base_url(self) -> str:
-        return "https://celestrak.org/NORAD/elements/gp.php"
+        return self._resolved_url or ""
 
     async def fetch(self) -> List[Dict[str, Any]]:
         satellite_groups = [
@@ -40,7 +40,7 @@ class CelesTrakTLECollector(BaseCollector):
         async with httpx.AsyncClient(timeout=120.0) as client:
             for group in satellite_groups:
                 try:
-                    url = f"https://celestrak.org/NORAD/elements/gp.php?GROUP={group}&FORMAT=json"
+                    url = f"{self.base_url}?GROUP={group}&FORMAT=json"
                     response = await client.get(url)
 
                     if response.status_code == 200:

backend/app/services/collectors/cloudflare.py

@@ -39,6 +39,16 @@ class CloudflareRadarDeviceCollector(HTTPCollector):
         if CLOUDFLARE_API_TOKEN:
             self.headers["Authorization"] = f"Bearer {CLOUDFLARE_API_TOKEN}"
 
+    @property
+    def request_url(self) -> str:
+        return self._resolved_url or self.base_url
+
+    async def fetch(self) -> List[Dict[str, Any]]:
+        async with httpx.AsyncClient(timeout=60.0) as client:
+            response = await client.get(self.request_url, headers=self.headers)
+            response.raise_for_status()
+            return self.parse_response(response.json())
+
     def parse_response(self, response: Dict[str, Any]) -> List[Dict[str, Any]]:
         """Parse Cloudflare Radar device type response"""
         data = []
@@ -87,6 +97,16 @@ class CloudflareRadarTrafficCollector(HTTPCollector):
         if CLOUDFLARE_API_TOKEN:
             self.headers["Authorization"] = f"Bearer {CLOUDFLARE_API_TOKEN}"
 
+    @property
+    def request_url(self) -> str:
+        return self._resolved_url or self.base_url
+
+    async def fetch(self) -> List[Dict[str, Any]]:
+        async with httpx.AsyncClient(timeout=60.0) as client:
+            response = await client.get(self.request_url, headers=self.headers)
+            response.raise_for_status()
+            return self.parse_response(response.json())
+
     def parse_response(self, response: Dict[str, Any]) -> List[Dict[str, Any]]:
         """Parse Cloudflare Radar traffic timeseries response"""
         data = []
@@ -135,6 +155,16 @@ class CloudflareRadarTopASCollector(HTTPCollector):
         if CLOUDFLARE_API_TOKEN:
             self.headers["Authorization"] = f"Bearer {CLOUDFLARE_API_TOKEN}"
 
+    @property
+    def request_url(self) -> str:
+        return self._resolved_url or self.base_url
+
+    async def fetch(self) -> List[Dict[str, Any]]:
+        async with httpx.AsyncClient(timeout=60.0) as client:
+            response = await client.get(self.request_url, headers=self.headers)
+            response.raise_for_status()
+            return self.parse_response(response.json())
+
     def parse_response(self, response: Dict[str, Any]) -> List[Dict[str, Any]]:
         """Parse Cloudflare Radar top locations response"""
         data = []

backend/app/services/collectors/epoch_ai.py

@@ -23,7 +23,7 @@ class EpochAIGPUCollector(BaseCollector):
 
     async def fetch(self) -> List[Dict[str, Any]]:
         """Fetch Epoch AI GPU clusters data from webpage"""
-        url = "https://epoch.ai/data/gpu-clusters"
+        url = self._resolved_url or ""
 
         async with httpx.AsyncClient(timeout=60.0) as client:
             response = await client.get(url)

backend/app/services/collectors/fao_landing.py

@@ -18,11 +18,9 @@ class FAOLandingPointCollector(BaseCollector):
     frequency_hours = 168
     data_type = "landing_point"
 
-    csv_url = "https://data.apps.fao.org/catalog/dataset/1b75ff21-92f2-4b96-9b7b-98e8aa65ad5d/resource/b6071077-d1d4-4e97-aa00-42e902847c87/download/landing-point-geo.csv"
-
     async def fetch(self) -> List[Dict[str, Any]]:
         async with httpx.AsyncClient(timeout=60.0) as client:
-            response = await client.get(self.csv_url)
+            response = await client.get(self._resolved_url or "")
             response.raise_for_status()
             return self.parse_csv(response.text)
 

backend/app/services/collectors/huggingface.py

@@ -21,6 +21,18 @@ class HuggingFaceModelCollector(HTTPCollector):
     data_type = "model"
     base_url = "https://huggingface.co/api/models"
 
+    @property
+    def request_url(self) -> str:
+        return self._resolved_url or self.base_url
+
+    async def fetch(self) -> List[Dict[str, Any]]:
+        from httpx import AsyncClient
+
+        async with AsyncClient(timeout=60.0) as client:
+            response = await client.get(self.request_url, headers=self.headers)
+            response.raise_for_status()
+            return self.parse_response(response.json())
+
     def parse_response(self, response: Dict[str, Any]) -> List[Dict[str, Any]]:
         """Parse Hugging Face models API response"""
         data = []
@@ -63,6 +75,18 @@ class HuggingFaceDatasetCollector(HTTPCollector):
     data_type = "dataset"
     base_url = "https://huggingface.co/api/datasets"
 
+    @property
+    def request_url(self) -> str:
+        return self._resolved_url or self.base_url
+
+    async def fetch(self) -> List[Dict[str, Any]]:
+        from httpx import AsyncClient
+
+        async with AsyncClient(timeout=60.0) as client:
+            response = await client.get(self.request_url, headers=self.headers)
+            response.raise_for_status()
+            return self.parse_response(response.json())
+
     def parse_response(self, response: Dict[str, Any]) -> List[Dict[str, Any]]:
         """Parse Hugging Face datasets API response"""
         data = []
@@ -104,6 +128,18 @@ class HuggingFaceSpacesCollector(HTTPCollector):
     data_type = "space"
     base_url = "https://huggingface.co/api/spaces"
 
+    @property
+    def request_url(self) -> str:
+        return self._resolved_url or self.base_url
+
+    async def fetch(self) -> List[Dict[str, Any]]:
+        from httpx import AsyncClient
+
+        async with AsyncClient(timeout=60.0) as client:
+            response = await client.get(self.request_url, headers=self.headers)
+            response.raise_for_status()
+            return self.parse_response(response.json())
+
     def parse_response(self, response: Dict[str, Any]) -> List[Dict[str, Any]]:
         """Parse Hugging Face Spaces API response"""
         data = []

backend/app/services/collectors/news_live_streams.py

@@ -0,0 +1,79 @@
+from __future__ import annotations
+
+from datetime import UTC, datetime
+from typing import Any
+
+import httpx
+
+from app.services.collectors.base import BaseCollector
+
+
+class NewsLiveStreamsCollector(BaseCollector):
+    """Collect normalized news live-stream sources from a JSON endpoint."""
+
+    name = "news_live_streams"
+    priority = "P2"
+    module = "L4"
+    frequency_hours = 12
+    data_type = "news_live_stream"
+    fail_on_empty = False
+
+    async def fetch(self) -> list[dict[str, Any]]:
+        request_url = (self._resolved_url or "").strip()
+        if not request_url:
+            return []
+
+        async with httpx.AsyncClient(timeout=45.0, follow_redirects=True) as client:
+            response = await client.get(
+                request_url,
+                headers={
+                    "User-Agent": "Planet-Intelligence-System/1.0 (Python/collector)",
+                    "Accept": "application/json",
+                },
+            )
+            response.raise_for_status()
+            return self.parse_response(response.json())
+
+    def parse_response(self, response: Any) -> list[dict[str, Any]]:
+        if isinstance(response, dict):
+            candidates = response.get("sources") or response.get("streams") or response.get("data") or []
+        elif isinstance(response, list):
+            candidates = response
+        else:
+            candidates = []
+
+        normalized: list[dict[str, Any]] = []
+        for index, item in enumerate(candidates):
+            if not isinstance(item, dict):
+                continue
+
+            stream_id = item.get("id") or item.get("source_id") or item.get("slug") or f"news-live-{index + 1}"
+            name = str(item.get("name") or item.get("title") or f"News Live {index + 1}").strip()
+            if not name:
+                continue
+
+            metadata = {
+                "provider": item.get("provider") or item.get("publisher") or "Collector",
+                "region": item.get("region") or item.get("country") or "Global",
+                "language": item.get("language") or "und",
+                "source_type": item.get("source_type") or "iframe",
+                "embed_url": item.get("embed_url") or item.get("url") or "",
+                "stream_url": item.get("stream_url") or "",
+                "homepage_url": item.get("homepage_url") or item.get("source_url") or "",
+                "poster_url": item.get("poster_url") or "",
+                "sort_order": item.get("sort_order", 200 + index),
+                "notes": item.get("notes") or item.get("description") or "",
+                "is_enabled": item.get("is_enabled", True),
+            }
+
+            normalized.append(
+                {
+                    "source_id": str(stream_id),
+                    "name": name,
+                    "description": metadata["notes"],
+                    "metadata": metadata,
+                    "reference_date": item.get("reference_date", datetime.now(UTC).isoformat()),
+                }
+            )
+
+        return normalized

backend/app/services/collectors/peeringdb.py

@@ -16,6 +16,7 @@ from typing import Dict, Any, List
 from datetime import UTC, datetime
 
 import httpx
+from urllib.parse import urlencode
 from app.services.collectors.base import HTTPCollector
 
 
@@ -38,9 +39,13 @@ class PeeringDBIXPCollector(HTTPCollector):
             "User-Agent": "Planet-Intelligence-System/1.0 (Python/collector)",
             "Accept": "application/json",
         }
-        # API key is added to URL as query parameter
-        if PEERINGDB_API_KEY:
-            self.base_url = f"{self.base_url}?key={PEERINGDB_API_KEY}"
+    @property
+    def request_url(self) -> str:
+        base = self._resolved_url or self.base_url
+        if not PEERINGDB_API_KEY:
+            return base
+        separator = "&" if "?" in base else "?"
+        return f"{base}{separator}{urlencode({'key': PEERINGDB_API_KEY})}"
 
     async def fetch_with_retry(
         self, max_retries: int = 3, base_delay: float = 2.0
@@ -51,7 +56,7 @@ class PeeringDBIXPCollector(HTTPCollector):
         for attempt in range(max_retries):
             try:
                 async with httpx.AsyncClient(timeout=60.0) as client:
-                    response = await client.get(self.base_url, headers=self.headers)
+                    response = await client.get(self.request_url, headers=self.headers)
 
                     if response.status_code == 429:
                         # Rate limited - wait and retry with exponential backoff
@@ -141,8 +146,13 @@ class PeeringDBNetworkCollector(HTTPCollector):
             "User-Agent": "Planet-Intelligence-System/1.0 (Python/collector)",
             "Accept": "application/json",
         }
-        if PEERINGDB_API_KEY:
-            self.base_url = f"{self.base_url}?key={PEERINGDB_API_KEY}"
+    @property
+    def request_url(self) -> str:
+        base = self._resolved_url or self.base_url
+        if not PEERINGDB_API_KEY:
+            return base
+        separator = "&" if "?" in base else "?"
+        return f"{base}{separator}{urlencode({'key': PEERINGDB_API_KEY})}"
 
     async def fetch_with_retry(
         self, max_retries: int = 3, base_delay: float = 2.0
@@ -153,7 +163,7 @@ class PeeringDBNetworkCollector(HTTPCollector):
         for attempt in range(max_retries):
             try:
                 async with httpx.AsyncClient(timeout=60.0) as client:
-                    response = await client.get(self.base_url, headers=self.headers)
+                    response = await client.get(self.request_url, headers=self.headers)
 
                     if response.status_code == 429:
                         delay = base_delay * (2**attempt)
@@ -244,8 +254,13 @@ class PeeringDBFacilityCollector(HTTPCollector):
             "User-Agent": "Planet-Intelligence-System/1.0 (Python/collector)",
             "Accept": "application/json",
         }
-        if PEERINGDB_API_KEY:
-            self.base_url = f"{self.base_url}?key={PEERINGDB_API_KEY}"
+    @property
+    def request_url(self) -> str:
+        base = self._resolved_url or self.base_url
+        if not PEERINGDB_API_KEY:
+            return base
+        separator = "&" if "?" in base else "?"
+        return f"{base}{separator}{urlencode({'key': PEERINGDB_API_KEY})}"
 
     async def fetch_with_retry(
         self, max_retries: int = 3, base_delay: float = 2.0
@@ -256,7 +271,7 @@ class PeeringDBFacilityCollector(HTTPCollector):
         for attempt in range(max_retries):
             try:
                 async with httpx.AsyncClient(timeout=60.0) as client:
-                    response = await client.get(self.base_url, headers=self.headers)
+                    response = await client.get(self.request_url, headers=self.headers)
 
                     if response.status_code == 429:
                         delay = base_delay * (2**attempt)

backend/app/services/collectors/ris_live.py

@@ -33,7 +33,7 @@ class RISLiveCollector(BaseCollector):
 
     def _fetch_via_stream(self) -> list[dict[str, Any]]:
         events: list[dict[str, Any]] = []
-        stream_url = "https://ris-live.ripe.net/v1/stream/?format=json&client=planet-ris-live"
+        stream_url = self._resolved_url or ""
         subscribe = json.dumps(
             {
                 "host": "rrc00",

backend/app/services/collectors/spacetrack.py

@@ -7,6 +7,7 @@ API documentation: https://www.space-track.org/documentation
 import json
 from typing import Dict, Any, List
 import httpx
+from urllib.parse import urlparse
 
 from app.services.collectors.base import BaseCollector
 from app.core.data_sources import get_data_sources_config
@@ -21,12 +22,30 @@ class SpaceTrackTLECollector(BaseCollector):
     data_type = "satellite_tle"
 
     @property
-    def base_url(self) -> str:
+    def query_url(self) -> str:
         config = get_data_sources_config()
         if self._resolved_url:
             return self._resolved_url
         return config.get_yaml_url("spacetrack_tle")
 
+    @property
+    def site_root(self) -> str:
+        config = get_data_sources_config()
+        configured_root = config.get_yaml_value("spacetrack.base_url")
+        if isinstance(configured_root, str) and configured_root:
+            return configured_root.rstrip("/")
+
+        parsed = urlparse(self.query_url)
+        return f"{parsed.scheme}://{parsed.netloc}".rstrip("/")
+
+    @property
+    def login_url(self) -> str:
+        return f"{self.site_root}/ajaxauth/login"
+
+    @property
+    def probe_url(self) -> str:
+        return f"{self.site_root}/basicspacedata/query/class/gp/NORAD_CAT_ID/25544/format/json"
+
     async def fetch(self) -> List[Dict[str, Any]]:
         from app.core.config import settings
 
@@ -47,13 +66,13 @@ class SpaceTrackTLECollector(BaseCollector):
                     "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36",
                     "Accept": "application/json, text/html, */*",
                     "Accept-Language": "en-US,en;q=0.9",
-                    "Referer": "https://www.space-track.org/",
+                    "Referer": f"{self.site_root}/",
                 },
             ) as client:
-                await client.get("https://www.space-track.org/")
+                await client.get(f"{self.site_root}/")
 
                 login_response = await client.post(
-                    "https://www.space-track.org/ajaxauth/login",
+                    self.login_url,
                     data={
                         "identity": username,
                         "password": password,
@@ -69,7 +88,7 @@ class SpaceTrackTLECollector(BaseCollector):
                         timeout=120.0,
                         follow_redirects=True,
                     ) as alt_client:
-                        await alt_client.get("https://www.space-track.org/")
+                        await alt_client.get(f"{self.site_root}/")
 
                         form_data = {
                             "username": username,
@@ -77,7 +96,7 @@ class SpaceTrackTLECollector(BaseCollector):
                             "query": "class/gp/NORAD_CAT_ID/25544/format/json",
                         }
                         alt_login = await alt_client.post(
-                            "https://www.space-track.org/ajaxauth/login",
+                            self.login_url,
                             data={
                                 "identity": username,
                                 "password": password,
@@ -86,9 +105,7 @@ class SpaceTrackTLECollector(BaseCollector):
                         print(f"SPACETRACK: Alt login status: {alt_login.status_code}")
 
                         if alt_login.status_code == 200:
-                            tle_response = await alt_client.get(
-                                "https://www.space-track.org/basicspacedata/query/class/gp/NORAD_CAT_ID/25544/format/json"
-                            )
+                            tle_response = await alt_client.get(self.probe_url)
                             if tle_response.status_code == 200:
                                 data = tle_response.json()
                                 print(f"SPACETRACK: Received {len(data)} records via alt method")
@@ -98,9 +115,7 @@ class SpaceTrackTLECollector(BaseCollector):
                     print(f"SPACETRACK: Login failed, using sample data")
                     return self._get_sample_data()
 
-                tle_response = await client.get(
-                    "https://www.space-track.org/basicspacedata/query/class/gp/NORAD_CAT_ID/25544/format/json"
-                )
+                tle_response = await client.get(self.probe_url)
                 print(f"SPACETRACK: TLE query status: {tle_response.status_code}")
 
                 if tle_response.status_code != 200:
@@ -127,11 +142,11 @@ class SpaceTrackTLECollector(BaseCollector):
                 },
             ) as client:
                 # First, visit the main page to get any cookies
-                await client.get("https://www.space-track.org/")
+                await client.get(f"{self.site_root}/")
 
                 # Login to get session cookie
                 login_response = await client.post(
-                    "https://www.space-track.org/ajaxauth/login",
+                    self.login_url,
                     data={
                         "identity": username,
                         "password": password,
@@ -146,13 +161,7 @@ class SpaceTrackTLECollector(BaseCollector):
                     return self._get_sample_data()
 
                 # Query for TLE data (get first 1000 satellites)
-                tle_response = await client.get(
-                    "https://www.space-track.org/basicspacedata/query"
-                    "/class/gp"
-                    "/orderby/EPOCH%20desc"
-                    "/limit/1000"
-                    "/format/json"
-                )
+                tle_response = await client.get(self.query_url)
                 print(f"SPACETRACK: TLE query status: {tle_response.status_code}")
 
                 if tle_response.status_code != 200:

backend/app/services/collectors/telegeography.py

@@ -11,6 +11,7 @@ from datetime import UTC, datetime
 from bs4 import BeautifulSoup
 import httpx
 
+from app.core.data_sources import get_data_sources_config
 from app.services.collectors.base import BaseCollector
 
 
@@ -24,15 +25,17 @@ class TeleGeographyCableCollector(BaseCollector):
 
     async def fetch(self) -> List[Dict[str, Any]]:
         """Fetch submarine cable data from Wayback Machine"""
+        config = get_data_sources_config()
         # Try multiple data sources
         sources = [
-            # Wayback Machine archive of TeleGeography
-            "https://web.archive.org/web/2024/https://www.submarinecablemap.com/api/v3/cable",
-            # Alternative: Try scraping the page
-            "https://www.submarinecablemap.com",
+            self._resolved_url or "",
+            str(config.get_yaml_value("telegeography.archived_cable_url") or ""),
+            str(config.get_yaml_value("telegeography.live_map_url") or ""),
         ]
 
         for url in sources:
+            if not url:
+                continue
             try:
                 async with httpx.AsyncClient(timeout=60.0, follow_redirects=True) as client:
                     response = await client.get(url)
@@ -161,7 +164,7 @@ class TeleGeographyLandingPointCollector(BaseCollector):
 
     async def fetch(self) -> List[Dict[str, Any]]:
         """Fetch landing point data from GitHub mirror"""
-        url = "https://raw.githubusercontent.com/lintaojlu/submarine_cable_information/main/landing_point.json"
+        url = self._resolved_url or ""
 
         async with httpx.AsyncClient(timeout=60.0) as client:
             response = await client.get(url)
@@ -225,7 +228,7 @@ class TeleGeographyCableSystemCollector(BaseCollector):
 
     async def fetch(self) -> List[Dict[str, Any]]:
         """Fetch cable system data"""
-        url = "https://raw.githubusercontent.com/lintaojlu/submarine_cable_information/main/cable.json"
+        url = self._resolved_url or ""
 
         async with httpx.AsyncClient(timeout=60.0) as client:
             response = await client.get(url)

backend/app/services/collectors/top500.py

@@ -10,6 +10,7 @@ from typing import Dict, Any, List
 from bs4 import BeautifulSoup
 import httpx
 
+from app.core.data_sources import get_data_sources_config
 from app.services.collectors.base import BaseCollector
 
 
@@ -22,7 +23,7 @@ class TOP500Collector(BaseCollector):
 
     async def fetch(self) -> List[Dict[str, Any]]:
         """Fetch TOP500 list data and enrich each row with detail-page metadata."""
-        url = "https://top500.org/lists/top500/list/2025/11/"
+        url = self._resolved_url or ""
 
         async with httpx.AsyncClient(timeout=60.0, follow_redirects=True) as client:
             response = await client.get(url)
@@ -48,11 +49,13 @@ class TOP500Collector(BaseCollector):
             return await asyncio.gather(*(enrich(entry) for entry in entries))
 
     def _extract_system_fields(self, system_cell) -> Dict[str, str]:
+        config = get_data_sources_config()
+        top500_base_url = config.get_yaml_value("top500.base_url") or "https://top500.org"
         link = system_cell.find("a")
         system_name = link.get_text(" ", strip=True) if link else system_cell.get_text(" ", strip=True)
         detail_url = ""
         if link and link.get("href"):
-            detail_url = f"https://top500.org{link.get('href')}"
+            detail_url = f"{str(top500_base_url).rstrip('/')}{link.get('href')}"
 
         manufacturer = ""
         if link and link.next_sibling:

backend/app/services/playground_chat_service.py

@@ -0,0 +1,694 @@
+from __future__ import annotations
+
+import asyncio
+from collections.abc import Sequence
+from datetime import UTC, datetime
+from time import perf_counter
+from uuid import uuid4
+
+from fastapi import HTTPException, status
+from sqlalchemy import func, select, update
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.db.session import async_session_factory
+from app.models.playground_message import PlaygroundMessage
+from app.models.playground_session import PlaygroundSession
+from app.schemas.ai import (
+    PlaygroundMessageEditRequest,
+    PlaygroundMessageActionResponse,
+    PlaygroundMessageCreateRequest,
+    PlaygroundMessageRecord,
+    PlaygroundMessageResendRequest,
+    PlaygroundMessageStopRequest,
+    PlaygroundSessionResponse,
+    PlaygroundSessionState,
+    PlaygroundSessionUpsertRequest,
+    PlaygroundThreadResponse,
+    SituationalAnalysisRequest,
+)
+from app.services.ai_client import AIProviderClient
+from app.services.playground_session_store import _to_response as session_to_response
+from app.services.playground_session_store import upsert_playground_session
+
+STREAM_CHUNK_SIZE = 24
+STREAM_INTERVAL_SECONDS = 0.08
+THINKING_PREVIEW_SECONDS = 2.6
+
+
+class _ActiveRun:
+    def __init__(self, task: asyncio.Task[None]) -> None:
+        self.task = task
+        self.stop_requested = asyncio.Event()
+
+
+_ACTIVE_RUNS: dict[str, _ActiveRun] = {}
+
+
+async def _get_session_by_key(
+    db: AsyncSession,
+    *,
+    user_id: int,
+    session_key: str,
+) -> PlaygroundSession | None:
+    result = await db.execute(
+        select(PlaygroundSession).where(
+            PlaygroundSession.user_id == user_id,
+            PlaygroundSession.session_key == session_key,
+        )
+    )
+    return result.scalar_one_or_none()
+
+
+async def _require_session(
+    db: AsyncSession,
+    *,
+    user_id: int,
+    session_key: str,
+) -> PlaygroundSession:
+    session = await _get_session_by_key(db, user_id=user_id, session_key=session_key)
+    if session is None:
+        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Playground session not found")
+    return session
+
+
+async def _require_visible_message(
+    db: AsyncSession,
+    *,
+    user_id: int,
+    public_id: str,
+    role: str | None = None,
+) -> PlaygroundMessage:
+    conditions = [
+        PlaygroundMessage.user_id == user_id,
+        PlaygroundMessage.public_id == public_id,
+        PlaygroundMessage.is_visible.is_(True),
+    ]
+    if role is not None:
+        conditions.append(PlaygroundMessage.role == role)
+
+    result = await db.execute(select(PlaygroundMessage).where(*conditions))
+    message = result.scalar_one_or_none()
+    if message is None:
+        detail = "User message not found" if role == "user" else "Playground message not found"
+        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail=detail)
+    return message
+
+
+def _message_to_record(message: PlaygroundMessage, parent_public_id: str | None = None) -> PlaygroundMessageRecord:
+    return PlaygroundMessageRecord(
+        id=message.public_id,
+        role=message.role,
+        kind=message.kind,
+        status=message.status,
+        title=message.title,
+        content=message.content or "",
+        thinking_content=message.thinking_content or "",
+        meta=list(message.meta or []),
+        markdown=message.role != "system",
+        provider=message.provider,
+        model=message.model,
+        request_id=message.request_id,
+        raw_response=dict(message.raw_response or {}),
+        content_blocks=list(message.content_blocks or []),
+        text_blocks=list(message.text_blocks or []),
+        thinking_blocks=list(message.thinking_blocks or []),
+        parent_message_id=parent_public_id,
+        created_at=message.created_at.isoformat(),
+        updated_at=message.updated_at.isoformat(),
+    )
+
+
+async def _ensure_session(
+    db: AsyncSession,
+    *,
+    user_id: int,
+    session_key: str,
+    title: str,
+    state: PlaygroundSessionState | None = None,
+) -> PlaygroundSession:
+    result = await db.execute(
+        select(PlaygroundSession).where(
+            PlaygroundSession.user_id == user_id,
+            PlaygroundSession.session_key == session_key,
+        )
+    )
+    session = result.scalar_one_or_none()
+    if session is not None:
+        if title:
+            session.title = title[:200]
+        if state is not None:
+            session.state = state.model_dump(mode="json")
+        await db.flush()
+        await db.refresh(session)
+        return session
+
+    payload = PlaygroundSessionUpsertRequest(
+        session_key=session_key,
+        title=title[:200],
+        state=state or PlaygroundSessionState(title=title[:200]),
+    )
+    await upsert_playground_session(db, user_id=user_id, payload=payload)
+    result = await db.execute(
+        select(PlaygroundSession).where(
+            PlaygroundSession.user_id == user_id,
+            PlaygroundSession.session_key == session_key,
+        )
+    )
+    return result.scalar_one()
+
+
+async def _list_visible_messages(
+    db: AsyncSession,
+    *,
+    session_id: int,
+) -> list[PlaygroundMessage]:
+    result = await db.execute(
+        select(PlaygroundMessage)
+        .where(
+            PlaygroundMessage.session_id == session_id,
+            PlaygroundMessage.is_visible.is_(True),
+        )
+        .order_by(PlaygroundMessage.sort_order.asc(), PlaygroundMessage.id.asc())
+    )
+    return list(result.scalars().all())
+
+
+async def _build_thread_response(
+    db: AsyncSession,
+    *,
+    session: PlaygroundSession,
+) -> PlaygroundThreadResponse:
+    messages = await _list_visible_messages(db, session_id=session.id)
+    id_map = {item.id: item.public_id for item in messages}
+    return PlaygroundThreadResponse(
+        session=session_to_response(session),
+        messages=[_message_to_record(item, id_map.get(item.parent_message_id)) for item in messages],
+    )
+
+
+async def get_thread(
+    db: AsyncSession,
+    *,
+    user_id: int,
+    session_key: str,
+) -> PlaygroundThreadResponse | None:
+    session = await _get_session_by_key(db, user_id=user_id, session_key=session_key)
+    if session is None:
+        return None
+    return await _build_thread_response(db, session=session)
+
+
+async def _build_action_response(
+    db: AsyncSession,
+    *,
+    session: PlaygroundSession,
+    active_message_id: str | None = None,
+) -> PlaygroundMessageActionResponse:
+    thread = await _build_thread_response(db, session=session)
+    return PlaygroundMessageActionResponse(
+        session=thread.session,
+        messages=thread.messages,
+        active_message_id=active_message_id,
+    )
+
+
+def _collect_constraints(raw_constraints: str) -> list[str]:
+    return [item.strip() for item in raw_constraints.split("\n") if item.strip()]
+
+
+async def _next_sort_order(db: AsyncSession, session_id: int) -> int:
+    result = await db.execute(
+        select(func.max(PlaygroundMessage.sort_order)).where(PlaygroundMessage.session_id == session_id)
+    )
+    current = result.scalar_one_or_none()
+    return int(current or 0)
+
+
+async def _set_session_state(
+    db: AsyncSession,
+    *,
+    session: PlaygroundSession,
+    payload: PlaygroundMessageCreateRequest,
+) -> PlaygroundSession:
+    session.state = PlaygroundSessionState(
+        messages=[],
+        selectedPresetKey=payload.selected_preset_key,
+        title=payload.title,
+        objective=payload.objective,
+        constraints=payload.constraints,
+        inputValue="",
+        analysis=None,
+        latestAnalysisMessageId=None,
+        analysisMeta={},
+        helpExpanded=payload.help_expanded,
+    ).model_dump(mode="json")
+    session.title = payload.title[:200]
+    await db.flush()
+    await db.refresh(session)
+    return session
+
+
+def _spawn_assistant_run(
+    *,
+    user_id: int,
+    session_id: int,
+    session_key: str,
+    user_message_id: int,
+    assistant_message_id: int,
+    assistant_public_id: str,
+    payload: PlaygroundMessageCreateRequest,
+    provider_client: AIProviderClient,
+) -> None:
+    task = asyncio.create_task(
+        _run_assistant_message(
+            user_id=user_id,
+            session_id=session_id,
+            session_key=session_key,
+            user_message_id=user_message_id,
+            assistant_message_id=assistant_message_id,
+            payload=payload,
+            provider_client=provider_client,
+        )
+    )
+    _ACTIVE_RUNS[assistant_public_id] = _ActiveRun(task)
+
+
+async def create_turn(
+    db: AsyncSession,
+    *,
+    user_id: int,
+    payload: PlaygroundMessageCreateRequest,
+    provider_client: AIProviderClient,
+) -> PlaygroundMessageActionResponse:
+    session = await _ensure_session(
+        db,
+        user_id=user_id,
+        session_key=payload.session_key,
+        title=payload.title,
+        state=PlaygroundSessionState(
+            selectedPresetKey=payload.selected_preset_key,
+            title=payload.title,
+            objective=payload.objective,
+            constraints=payload.constraints,
+            inputValue="",
+            helpExpanded=payload.help_expanded,
+        ),
+    )
+    session = await _set_session_state(db, session=session, payload=payload)
+    base_order = await _next_sort_order(db, session.id)
+
+    user_message = PlaygroundMessage(
+        public_id=uuid4().hex,
+        session_id=session.id,
+        user_id=user_id,
+        role="user",
+        kind="message",
+        status="done",
+        title=payload.selected_preset_key,
+        content=payload.input,
+        meta=[payload.title],
+        sort_order=base_order + 10,
+    )
+    assistant_message = PlaygroundMessage(
+        public_id=uuid4().hex,
+        session_id=session.id,
+        user_id=user_id,
+        parent_message_id=None,
+        role="assistant",
+        kind="thinking",
+        status="pending",
+        title="AI 回应",
+        content="",
+        thinking_content="",
+        meta=[],
+        sort_order=base_order + 20,
+    )
+    db.add(user_message)
+    await db.flush()
+    assistant_message.parent_message_id = user_message.id
+    db.add(assistant_message)
+    await db.flush()
+    await db.refresh(user_message)
+    await db.refresh(assistant_message)
+    await db.commit()
+    await db.refresh(session)
+    await db.refresh(user_message)
+    await db.refresh(assistant_message)
+
+    _spawn_assistant_run(
+        user_id=user_id,
+        session_id=session.id,
+        session_key=payload.session_key,
+        user_message_id=user_message.id,
+        assistant_message_id=assistant_message.id,
+        assistant_public_id=assistant_message.public_id,
+        payload=payload,
+        provider_client=provider_client,
+    )
+
+    return await _build_action_response(
+        db,
+        session=session,
+        active_message_id=assistant_message.public_id,
+    )
+
+
+async def _create_assistant_retry_turn(
+    db: AsyncSession,
+    *,
+    user_id: int,
+    session: PlaygroundSession,
+    user_message: PlaygroundMessage,
+    payload: PlaygroundMessageCreateRequest,
+    provider_client: AIProviderClient,
+) -> PlaygroundMessageActionResponse:
+    base_order = await _next_sort_order(db, session.id)
+    assistant_message = PlaygroundMessage(
+        public_id=uuid4().hex,
+        session_id=session.id,
+        user_id=user_id,
+        parent_message_id=user_message.id,
+        role="assistant",
+        kind="thinking",
+        status="pending",
+        title="AI 回应",
+        content="",
+        thinking_content="",
+        meta=[],
+        sort_order=base_order + 10,
+    )
+    db.add(assistant_message)
+    await db.flush()
+    await db.commit()
+    await db.refresh(session)
+    await db.refresh(assistant_message)
+
+    _spawn_assistant_run(
+        user_id=user_id,
+        session_id=session.id,
+        session_key=payload.session_key,
+        user_message_id=user_message.id,
+        assistant_message_id=assistant_message.id,
+        assistant_public_id=assistant_message.public_id,
+        payload=payload,
+        provider_client=provider_client,
+    )
+
+    return await _build_action_response(
+        db,
+        session=session,
+        active_message_id=assistant_message.public_id,
+    )
+
+
+async def stop_message(
+    db: AsyncSession,
+    *,
+    user_id: int,
+    payload: PlaygroundMessageStopRequest,
+) -> PlaygroundMessageActionResponse:
+    session = await _require_session(db, user_id=user_id, session_key=payload.session_key)
+    message = await _require_visible_message(db, user_id=user_id, public_id=payload.message_id)
+
+    if message.status not in {"pending", "thinking", "answering"}:
+        return await _build_action_response(db, session=session)
+
+    active_run = _ACTIVE_RUNS.get(message.public_id)
+    if active_run is not None:
+        active_run.stop_requested.set()
+        active_run.task.cancel()
+
+    message.status = "stopped"
+    if "已手动停止生成" not in (message.meta or []):
+        message.meta = [*(message.meta or []), "已手动停止生成"]
+    await db.flush()
+    await db.commit()
+    await db.refresh(message)
+
+    return await _build_action_response(db, session=session)
+
+
+async def resend_turn(
+    db: AsyncSession,
+    *,
+    user_id: int,
+    payload: PlaygroundMessageResendRequest,
+    provider_client: AIProviderClient,
+) -> PlaygroundMessageActionResponse:
+    session = await _require_session(db, user_id=user_id, session_key=payload.session_key)
+    user_message = await _require_visible_message(
+        db,
+        user_id=user_id,
+        public_id=payload.user_message_id,
+        role="user",
+    )
+
+    later_messages = await db.execute(
+        select(PlaygroundMessage).where(
+            PlaygroundMessage.session_id == session.id,
+            PlaygroundMessage.sort_order > user_message.sort_order,
+            PlaygroundMessage.is_visible.is_(True),
+        )
+    )
+    for item in later_messages.scalars().all():
+        item.is_visible = False
+        if item.status in {"pending", "thinking", "answering"}:
+            active_run = _ACTIVE_RUNS.get(item.public_id)
+            if active_run is not None:
+                active_run.stop_requested.set()
+                active_run.task.cancel()
+    await db.flush()
+    await db.commit()
+
+    session_state = PlaygroundSessionState.model_validate(session.state or {})
+    create_payload = PlaygroundMessageCreateRequest(
+        session_key=payload.session_key,
+        title=session_state.title or session.title,
+        objective=session_state.objective or "继续当前对话",
+        constraints=session_state.constraints or "",
+        input=user_message.content,
+        selected_preset_key=session_state.selectedPresetKey or "bgp-brief",
+        help_expanded=session_state.helpExpanded,
+    )
+    return await _create_assistant_retry_turn(
+        db,
+        user_id=user_id,
+        session=session,
+        user_message=user_message,
+        payload=create_payload,
+        provider_client=provider_client,
+    )
+
+
+async def edit_user_message(
+    db: AsyncSession,
+    *,
+    user_id: int,
+    payload: PlaygroundMessageEditRequest,
+) -> PlaygroundMessageActionResponse:
+    session = await _require_session(db, user_id=user_id, session_key=payload.session_key)
+    user_message = await _require_visible_message(
+        db,
+        user_id=user_id,
+        public_id=payload.user_message_id,
+        role="user",
+    )
+
+    user_message.content = payload.content.strip()
+    await db.flush()
+    await db.commit()
+    await db.refresh(user_message)
+
+    return await _build_action_response(db, session=session)
+
+
+async def _append_meta_if_missing(db: AsyncSession, message_id: int, meta_line: str) -> None:
+    result = await db.execute(select(PlaygroundMessage).where(PlaygroundMessage.id == message_id))
+    message = result.scalar_one_or_none()
+    if message is None:
+        return
+    if meta_line not in (message.meta or []):
+        message.meta = [*(message.meta or []), meta_line]
+        await db.flush()
+
+
+async def _should_stop(message_public_id: str) -> bool:
+    active_run = _ACTIVE_RUNS.get(message_public_id)
+    return active_run.stop_requested.is_set() if active_run is not None else False
+
+
+async def _mark_message_state(
+    db: AsyncSession,
+    *,
+    message_id: int,
+    **updates,
+) -> PlaygroundMessage:
+    result = await db.execute(select(PlaygroundMessage).where(PlaygroundMessage.id == message_id))
+    message = result.scalar_one()
+    for key, value in updates.items():
+        setattr(message, key, value)
+    await db.flush()
+    await db.refresh(message)
+    return message
+
+
+def _build_conversation_history(messages: Sequence[PlaygroundMessage], current_user_message_id: int) -> list[dict]:
+    history: list[dict] = []
+    for item in messages:
+        if item.id >= current_user_message_id:
+            break
+        if item.role == "system":
+            continue
+        history.append(
+            {
+                "role": item.role,
+                "kind": item.kind or "message",
+                "title": item.title,
+                "content": item.content or "",
+            }
+        )
+    return history[-8:]
+
+
+async def _run_assistant_message(
+    *,
+    user_id: int,
+    session_id: int,
+    session_key: str,
+    user_message_id: int,
+    assistant_message_id: int,
+    payload: PlaygroundMessageCreateRequest,
+    provider_client: AIProviderClient,
+) -> None:
+    request_id = str(uuid4())
+    started_at = perf_counter()
+    assistant_public_id: str | None = None
+    try:
+        async with async_session_factory() as db:
+            session = await db.get(PlaygroundSession, session_id)
+            user_message = await db.get(PlaygroundMessage, user_message_id)
+            assistant_message = await db.get(PlaygroundMessage, assistant_message_id)
+            if session is None or user_message is None or assistant_message is None:
+                return
+            assistant_public_id = assistant_message.public_id
+
+            visible_messages = await _list_visible_messages(db, session_id=session_id)
+            conversation_history = _build_conversation_history(visible_messages, user_message_id)
+
+            request_payload = SituationalAnalysisRequest(
+                title=payload.title,
+                objective=payload.objective,
+                observations=[item.strip() for item in payload.input.split("\n") if item.strip()],
+                constraints=_collect_constraints(payload.constraints),
+                context={
+                    "source": "playground",
+                    "preset": payload.selected_preset_key,
+                    "conversation_history": conversation_history,
+                    "history_size": len(conversation_history),
+                },
+                thinking={"type": "enabled"},
+            )
+
+        analysis = await provider_client.analyze(request_payload, request_id=request_id)
+
+        async with async_session_factory() as db:
+            assistant_message = await _mark_message_state(
+                db,
+                message_id=assistant_message_id,
+                status="thinking" if analysis.thinking_blocks else "answering",
+                title=f"{analysis.provider} / {analysis.model}",
+                provider=analysis.provider,
+                model=analysis.model,
+                request_id=request_id,
+                raw_response=analysis.raw_response,
+                content_blocks=[item.model_dump(mode="json") for item in analysis.content_blocks],
+                text_blocks=analysis.text_blocks,
+                thinking_blocks=analysis.thinking_blocks,
+                thinking_content="\n\n".join(analysis.thinking_blocks).strip(),
+            )
+            session = await db.get(PlaygroundSession, session_id)
+            if session is not None:
+                session_state = PlaygroundSessionState.model_validate(session.state or {})
+                session.state = session_state.model_copy(
+                    update={
+                        "latestAnalysisMessageId": assistant_message.public_id,
+                        "analysis": analysis.model_dump(mode="json"),
+                    }
+                ).model_dump(mode="json")
+                await db.flush()
+            await db.commit()
+
+        if assistant_public_id and analysis.thinking_blocks:
+            await asyncio.sleep(THINKING_PREVIEW_SECONDS)
+            if await _should_stop(assistant_public_id):
+                return
+
+        content = analysis.content or ""
+        cursor = 0
+        while cursor < len(content):
+            if assistant_public_id and await _should_stop(assistant_public_id):
+                return
+            cursor = min(len(content), cursor + STREAM_CHUNK_SIZE)
+            async with async_session_factory() as db:
+                await _mark_message_state(
+                    db,
+                    message_id=assistant_message_id,
+                    status="answering",
+                    content=content[:cursor],
+                )
+                await db.commit()
+            await asyncio.sleep(STREAM_INTERVAL_SECONDS)
+
+        duration_ms = round((perf_counter() - started_at) * 1000)
+        async with async_session_factory() as db:
+            assistant_message = await _mark_message_state(
+                db,
+                message_id=assistant_message_id,
+                status="done",
+                content=content,
+                meta=[
+                    f"Request ID: {request_id}",
+                    f"耗时: {duration_ms} ms",
+                    f"完成时间: {datetime.now(UTC).astimezone().isoformat()}",
+                ],
+            )
+            session = await db.get(PlaygroundSession, session_id)
+            if session is not None:
+                session_state = PlaygroundSessionState.model_validate(session.state or {})
+                session.state = session_state.model_copy(
+                    update={
+                        "latestAnalysisMessageId": assistant_message.public_id,
+                        "analysis": analysis.model_dump(mode="json"),
+                        "analysisMeta": {
+                            "requestId": request_id,
+                            "durationMs": duration_ms,
+                            "completedAt": datetime.now().isoformat(),
+                        },
+                    }
+                ).model_dump(mode="json")
+                await db.flush()
+            await db.commit()
+    except asyncio.CancelledError:
+        async with async_session_factory() as db:
+            result = await db.execute(select(PlaygroundMessage).where(PlaygroundMessage.id == assistant_message_id))
+            message = result.scalar_one_or_none()
+            if message is not None and message.status in {"pending", "thinking", "answering"}:
+                message.status = "stopped"
+                if "已手动停止生成" not in (message.meta or []):
+                    message.meta = [*(message.meta or []), "已手动停止生成"]
+                await db.flush()
+                await db.commit()
+        raise
+    except Exception as exc:
+        async with async_session_factory() as db:
+            result = await db.execute(select(PlaygroundMessage).where(PlaygroundMessage.id == assistant_message_id))
+            message = result.scalar_one_or_none()
+            if message is not None:
+                message.status = "error"
+                message.content = message.content or "分析失败，请检查 AI Provider 配置或稍后再试。"
+                message.meta = [*(message.meta or []), f"错误: {type(exc).__name__}"]
+                await db.flush()
+                await db.commit()
+    finally:
+        if assistant_public_id:
+            _ACTIVE_RUNS.pop(assistant_public_id, None)

backend/app/services/playground_session_store.py

@@ -0,0 +1,72 @@
+from __future__ import annotations
+
+from sqlalchemy import select
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.models.playground_session import PlaygroundSession
+from app.schemas.ai import (
+    PlaygroundSessionResponse,
+    PlaygroundSessionState,
+    PlaygroundSessionUpsertRequest,
+)
+
+
+def _to_response(record: PlaygroundSession) -> PlaygroundSessionResponse:
+    return PlaygroundSessionResponse(
+        id=str(record.id),
+        session_key=record.session_key,
+        title=record.title,
+        state=PlaygroundSessionState.model_validate(record.state or {}),
+        created_at=record.created_at.isoformat(),
+        updated_at=record.updated_at.isoformat(),
+    )
+
+
+async def get_playground_session(
+    db: AsyncSession,
+    *,
+    user_id: int,
+    session_key: str = "default",
+) -> PlaygroundSessionResponse | None:
+    result = await db.execute(
+        select(PlaygroundSession).where(
+            PlaygroundSession.user_id == user_id,
+            PlaygroundSession.session_key == session_key,
+        )
+    )
+    record = result.scalar_one_or_none()
+    if record is None:
+        return None
+    return _to_response(record)
+
+
+async def upsert_playground_session(
+    db: AsyncSession,
+    *,
+    user_id: int,
+    payload: PlaygroundSessionUpsertRequest,
+) -> PlaygroundSessionResponse:
+    result = await db.execute(
+        select(PlaygroundSession).where(
+            PlaygroundSession.user_id == user_id,
+            PlaygroundSession.session_key == payload.session_key,
+        )
+    )
+    record = result.scalar_one_or_none()
+    title = (payload.title or payload.state.title or "Playground 会话").strip()[:200] or "Playground 会话"
+
+    if record is None:
+        record = PlaygroundSession(
+            user_id=user_id,
+            session_key=payload.session_key,
+            title=title,
+            state=payload.state.model_dump(mode="json"),
+        )
+        db.add(record)
+    else:
+        record.title = title
+        record.state = payload.state.model_dump(mode="json")
+
+    await db.flush()
+    await db.refresh(record)
+    return _to_response(record)

backend/app/services/scheduler.py

@@ -19,6 +19,30 @@ logger = logging.getLogger(__name__)
 
 scheduler = AsyncIOScheduler()
 RUNNING_TASK_GUARD_TIMEOUT_MINUTES = 90
+RUNNING_COLLECTOR_TASKS: dict[str, asyncio.Task[Any]] = {}
+
+
+def _collector_task_name(collector_name: str) -> str:
+    return f"collector:{collector_name}"
+
+
+def get_running_collector_task(collector_name: str) -> asyncio.Task[Any] | None:
+    task = RUNNING_COLLECTOR_TASKS.get(collector_name)
+    if task is not None and not task.done():
+        return task
+
+    if task is not None and task.done():
+        RUNNING_COLLECTOR_TASKS.pop(collector_name, None)
+
+    target_name = _collector_task_name(collector_name)
+    for candidate in asyncio.all_tasks():
+        if candidate.done():
+            continue
+        if candidate.get_name() == target_name:
+            RUNNING_COLLECTOR_TASKS[collector_name] = candidate
+            return candidate
+
+    return None
 
 
 async def _update_next_run_at(datasource: DataSource, session) -> None:
@@ -133,6 +157,12 @@ async def run_collector_task(collector_name: str):
             datasource.last_status = task_result.get("status")
             await _update_next_run_at(datasource, db)
             logger.info("Collector %s completed: %s", collector_name, task_result)
+        except asyncio.CancelledError:
+            datasource.last_run_at = datetime.now(UTC)
+            datasource.last_status = "cancelled"
+            await db.commit()
+            logger.warning("Collector %s cancelled by operator", collector_name)
+            raise
         except Exception as exc:
             datasource.last_run_at = datetime.now(UTC)
             datasource.last_status = "failed"
@@ -244,10 +274,37 @@ def run_collector_now(collector_name: str) -> bool:
         logger.error("Collector not found: %s", collector_name)
         return False
 
+    existing_task = get_running_collector_task(collector_name)
+    if existing_task is not None and not existing_task.done():
+        logger.warning("Collector %s is already running in-memory; skipping duplicate trigger", collector_name)
+        return False
+
     try:
-        asyncio.create_task(run_collector_task(collector_name))
+        task = asyncio.create_task(run_collector_task(collector_name), name=_collector_task_name(collector_name))
+        RUNNING_COLLECTOR_TASKS[collector_name] = task
+
+        def _cleanup_task(done_task: asyncio.Task[Any]) -> None:
+            current = RUNNING_COLLECTOR_TASKS.get(collector_name)
+            if current is done_task:
+                RUNNING_COLLECTOR_TASKS.pop(collector_name, None)
+
+        task.add_done_callback(_cleanup_task)
         logger.info("Triggered collector: %s", collector_name)
         return True
     except Exception as exc:
         logger.error("Failed to trigger collector %s: %s", collector_name, exc)
-        return False
+        return False
+
+
+async def cancel_running_collector_now(collector_name: str) -> bool:
+    task = get_running_collector_task(collector_name)
+    if task is None or task.done():
+        RUNNING_COLLECTOR_TASKS.pop(collector_name, None)
+        return False
+
+    task.cancel()
+    try:
+        await task
+    except asyncio.CancelledError:
+        return True
+    return task.cancelled()

backend/app/services/situational_alert_ai_brief.py

@@ -0,0 +1,174 @@
+from __future__ import annotations
+
+from collections import Counter
+from typing import Any
+
+from sqlalchemy import func, select
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.models.alert import Alert, AlertSeverity, AlertStatus
+from app.models.bgp_anomaly import BGPAnomaly
+from app.models.bgp_incident import BGPIncident
+from app.schemas.ai import SituationalAnalysisRequest
+from app.services.bgp_ai_brief_store import get_latest_bgp_brief_record
+
+
+def _format_pairs(pairs: list[tuple[str, int]], empty_text: str = "无") -> str:
+    if not pairs:
+        return empty_text
+    return "，".join(f"{key} {value}" for key, value in pairs if key)
+
+
+async def build_situational_alert_brief_request(
+    db: AsyncSession,
+) -> tuple[SituationalAnalysisRequest, list[str], dict[str, Any]]:
+    total_alerts_result = await db.execute(select(func.count(Alert.id)))
+    active_alerts_result = await db.execute(
+        select(func.count(Alert.id)).where(Alert.status == AlertStatus.ACTIVE)
+    )
+    alert_severity_result = await db.execute(
+        select(Alert.severity, func.count(Alert.id))
+        .where(Alert.status == AlertStatus.ACTIVE)
+        .group_by(Alert.severity)
+    )
+    alert_source_result = await db.execute(
+        select(Alert.datasource_name, func.count(Alert.id))
+        .where(Alert.status == AlertStatus.ACTIVE)
+        .group_by(Alert.datasource_name)
+        .order_by(func.count(Alert.id).desc())
+        .limit(6)
+    )
+    recent_alerts_result = await db.execute(
+        select(Alert)
+        .order_by(Alert.created_at.desc(), Alert.id.desc())
+        .limit(6)
+    )
+
+    total_incidents_result = await db.execute(select(func.count(BGPIncident.id)))
+    active_incidents_result = await db.execute(
+        select(func.count(BGPIncident.id)).where(BGPIncident.status == "active")
+    )
+    bgp_severity_result = await db.execute(
+        select(BGPIncident.severity, func.count(BGPIncident.id))
+        .where(BGPIncident.status == "active")
+        .group_by(BGPIncident.severity)
+    )
+    bgp_region_counter: Counter[str] = Counter()
+    recent_incidents_result = await db.execute(
+        select(BGPIncident)
+        .order_by(BGPIncident.created_at.desc(), BGPIncident.id.desc())
+        .limit(5)
+    )
+
+    total_anomalies_result = await db.execute(select(func.count(BGPAnomaly.id)))
+    active_anomalies_result = await db.execute(
+        select(func.count(BGPAnomaly.id)).where(BGPAnomaly.status == "active")
+    )
+    anomaly_type_result = await db.execute(
+        select(BGPAnomaly.anomaly_type, func.count(BGPAnomaly.id))
+        .where(BGPAnomaly.status == "active")
+        .group_by(BGPAnomaly.anomaly_type)
+        .order_by(func.count(BGPAnomaly.id).desc())
+        .limit(6)
+    )
+
+    recent_incidents = recent_incidents_result.scalars().all()
+    for incident in recent_incidents:
+        for region in incident.affected_regions or []:
+            if not isinstance(region, dict):
+                continue
+            label = ", ".join(part for part in [region.get("city"), region.get("country")] if part) or "未知区域"
+            bgp_region_counter[label] += 1
+
+    latest_bgp_brief = get_latest_bgp_brief_record()
+    active_alert_severities = [
+        (item[0].value if isinstance(item[0], AlertSeverity) else str(item[0]), item[1])
+        for item in alert_severity_result.fetchall()
+        if item[0]
+    ]
+    active_bgp_severities = [
+        (str(item[0]), item[1])
+        for item in bgp_severity_result.fetchall()
+        if item[0]
+    ]
+    active_anomaly_types = [(str(item[0]), item[1]) for item in anomaly_type_result.fetchall() if item[0]]
+    active_alert_sources = [
+        (str(item[0] or "未命名数据源"), item[1])
+        for item in alert_source_result.fetchall()
+    ]
+
+    facts = [
+        (
+            f"系统告警侧：总告警 {total_alerts_result.scalar() or 0} 条，active {active_alerts_result.scalar() or 0} 条；"
+            f"活跃告警严重度分布为 {_format_pairs(active_alert_severities)}。"
+        ),
+        (
+            f"BGP态势侧：累计 incidents {total_incidents_result.scalar() or 0} 条，active incidents {active_incidents_result.scalar() or 0} 条；"
+            f"活跃 incidents 严重度分布为 {_format_pairs(active_bgp_severities)}。"
+        ),
+        (
+            f"BGP异常侧：累计 anomalies {total_anomalies_result.scalar() or 0} 条，active anomalies {active_anomalies_result.scalar() or 0} 条；"
+            f"活跃 anomaly 类型分布为 {_format_pairs(active_anomaly_types)}。"
+        ),
+    ]
+
+    if active_alert_sources:
+        facts.append(f"当前系统告警主要集中在：{_format_pairs(active_alert_sources)}。")
+    if bgp_region_counter:
+        facts.append(f"BGP近期高风险区域线索：{_format_pairs(bgp_region_counter.most_common(5))}。")
+
+    recent_alerts = recent_alerts_result.scalars().all()
+    if recent_alerts:
+        facts.append(
+            "最近系统告警摘录："
+            + "；".join(
+                [
+                    f"{alert.datasource_name or '未命名数据源'} / {alert.severity.value if alert.severity else '-'} / {alert.status.value if alert.status else '-'} / {alert.message or '-'}"
+                    for alert in recent_alerts
+                ]
+            )
+        )
+
+    if recent_incidents:
+        facts.append(
+            "最近BGP事件摘录："
+            + "；".join(
+                [
+                    f"{incident.incident_type} / {incident.severity} / {incident.status} / {incident.summary}"
+                    for incident in recent_incidents
+                ]
+            )
+        )
+
+    if latest_bgp_brief:
+        facts.append(
+            f"最近一份 BGP AI 简报生成于 {latest_bgp_brief.generated_at}，模型 {latest_bgp_brief.model}，可作为当前态势的补充说明。"
+        )
+
+    context = {
+        "source": "situational-alerts",
+        "active_system_alerts": active_alerts_result.scalar() or 0,
+        "active_system_alert_severities": dict(active_alert_severities),
+        "top_system_alert_sources": dict(active_alert_sources),
+        "active_bgp_incidents": active_incidents_result.scalar() or 0,
+        "active_bgp_incident_severities": dict(active_bgp_severities),
+        "active_bgp_anomalies": active_anomalies_result.scalar() or 0,
+        "active_bgp_anomaly_types": dict(active_anomaly_types),
+        "bgp_hot_regions": dict(bgp_region_counter.most_common(5)),
+        "latest_bgp_brief_id": latest_bgp_brief.id if latest_bgp_brief else None,
+        "latest_bgp_brief_generated_at": latest_bgp_brief.generated_at if latest_bgp_brief else None,
+    }
+
+    request = SituationalAnalysisRequest(
+        title="态势告警 AI 简报",
+        objective="综合系统告警、BGP incidents、BGP anomalies 与近期 BGP AI 简报，生成一份面向值班人员的态势告警简报，指出当前最需要关注的风险域、跨模块联动迹象和优先动作。",
+        observations=facts,
+        constraints=[
+            "明确区分事实、推断与建议。",
+            "优先指出仍在 active 状态的系统告警与 BGP 风险是否存在联动。",
+            "不要把单一数据源的局部异常夸大成全局态势。",
+            "如果证据不足，请明确写出仍缺哪些模块或区域信息。",
+        ],
+        context=context,
+    )
+    return request, facts, context

backend/app/services/tv_streams.py

@@ -0,0 +1,466 @@
+from __future__ import annotations
+
+from datetime import UTC, datetime
+from typing import Any
+from urllib.parse import urlparse
+
+from sqlalchemy import select
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.core.time import to_iso8601_utc
+from app.models.collected_data import CollectedData
+from app.models.system_setting import SystemSetting
+
+DEFAULT_TV_SOURCE_ID = "cgtn-en"
+TV_SETTINGS_CATEGORY = "tv"
+TV_LIVE_SOURCE_COLLECTOR = "news_live_streams"
+TV_LIVE_SOURCE_DATA_TYPE = "news_live_stream"
+
+DEFAULT_TV_SETTINGS = {
+        "default_source_id": DEFAULT_TV_SOURCE_ID,
+    "auto_fallback": True,
+    "sources": [
+        {
+            "id": "cctv4",
+            "name": "CCTV-4 中文国际",
+            "provider": "CCTV",
+            "region": "China",
+            "language": "zh-CN",
+            "source_type": "hls",
+            "embed_url": "https://tv.cctv.com/live/cctv4/",
+            "stream_url": "https://ldocctvwbcdtxy.liveplay.myqcloud.com/ldocctvwbcd/cdrmldcctv4_1_td.m3u8",
+            "homepage_url": "https://tv.cctv.com/live/cctv4/",
+            "poster_url": "",
+            "is_enabled": True,
+            "is_fallback": True,
+            "sort_order": 10,
+            "collector_source": None,
+            "notes": "默认兜底新闻直播源。优先尝试 CCTV-4 官方 HLS 播放流，若直播放失败则回退到央视官网直播页。",
+        },
+        {
+            "id": "reuters-tv",
+            "name": "Reuters TV",
+            "provider": "Reuters",
+            "region": "Global",
+            "language": "en",
+            "source_type": "hls",
+            "embed_url": "",
+            "stream_url": "https://reuters-reutersnow-1-eu.rakuten.wurl.tv/playlist.m3u8",
+            "homepage_url": "https://www.reuters.com/video/live/",
+            "poster_url": "",
+            "is_enabled": True,
+            "is_fallback": False,
+            "sort_order": 20,
+            "collector_source": None,
+            "notes": "参考 worldmonitor 的默认新闻频道清单，优先作为全球英文新闻直播放源。",
+        },
+        {
+            "id": "cgtn-en",
+            "name": "CGTN English",
+            "provider": "CGTN",
+            "region": "Global",
+            "language": "en",
+            "source_type": "youtube",
+            "embed_url": "https://www.youtube.com/watch?v=BOy2xDU1LC8",
+            "stream_url": "https://news.cgtn.com/resource/live/english/cgtn-news.m3u8",
+            "youtube_video_id": "BOy2xDU1LC8",
+            "homepage_url": "https://news.cgtn.com/",
+            "poster_url": "",
+            "is_enabled": True,
+            "is_fallback": False,
+            "sort_order": 30,
+            "collector_source": None,
+            "notes": "优先使用官方 YouTube 直播源，保留 HLS 直播放流作为候选信息。",
+        },
+        {
+            "id": "cgtn-es",
+            "name": "CGTN Espanol",
+            "provider": "CGTN",
+            "region": "Latin America",
+            "language": "es",
+            "source_type": "hls",
+            "embed_url": "",
+            "stream_url": "https://news.cgtn.com/resource/live/espanol/cgtn-e.m3u8",
+            "homepage_url": "https://news.cgtn.com/",
+            "poster_url": "",
+            "is_enabled": True,
+            "is_fallback": False,
+            "sort_order": 40,
+            "collector_source": None,
+            "notes": "西语国际新闻频道，覆盖拉美方向态势。",
+        },
+        {
+            "id": "dw-espanol",
+            "name": "DW Espanol",
+            "provider": "Deutsche Welle",
+            "region": "Europe",
+            "language": "es",
+            "source_type": "hls",
+            "embed_url": "",
+            "stream_url": "https://dwamdstream104.akamaized.net/hls/live/2015530/dwstream104/stream04/streamPlaylist.m3u8",
+            "homepage_url": "https://www.dw.com/es/",
+            "poster_url": "",
+            "is_enabled": True,
+            "is_fallback": False,
+            "sort_order": 50,
+            "collector_source": None,
+            "notes": "来自 worldmonitor 可选频道清单的直播放源。",
+        },
+        {
+            "id": "dw-arabic",
+            "name": "DW Arabic",
+            "provider": "Deutsche Welle",
+            "region": "Middle East",
+            "language": "ar",
+            "source_type": "hls",
+            "embed_url": "",
+            "stream_url": "https://dwamdstream103.akamaized.net/hls/live/2015526/dwstream103/index.m3u8",
+            "homepage_url": "https://www.dw.com/ar/",
+            "poster_url": "",
+            "is_enabled": True,
+            "is_fallback": False,
+            "sort_order": 60,
+            "collector_source": None,
+            "notes": "阿拉伯语新闻流，适合作为中东方向新闻补充源。",
+        },
+        {
+            "id": "aljazeera-mubasher",
+            "name": "Al Jazeera Mubasher",
+            "provider": "Al Jazeera",
+            "region": "Middle East",
+            "language": "ar",
+            "source_type": "hls",
+            "embed_url": "",
+            "stream_url": "https://live-hls-web-ajm.getaj.net/AJM/index.m3u8",
+            "homepage_url": "https://www.aljazeera.net/live",
+            "poster_url": "",
+            "is_enabled": True,
+            "is_fallback": False,
+            "sort_order": 70,
+            "collector_source": None,
+            "notes": "中东实时新闻流，来自 worldmonitor HLS 频道目录。",
+        },
+        {
+            "id": "arirang-news",
+            "name": "Arirang News",
+            "provider": "Arirang",
+            "region": "Korea",
+            "language": "en",
+            "source_type": "hls",
+            "embed_url": "",
+            "stream_url": "https://amdlive-ch01-ctnd-com.akamaized.net/arirang_1ch/smil:arirang_1ch.smil/playlist.m3u8",
+            "homepage_url": "https://www.arirang.com/",
+            "poster_url": "",
+            "is_enabled": True,
+            "is_fallback": False,
+            "sort_order": 80,
+            "collector_source": None,
+            "notes": "东北亚英语新闻源，适合补充韩半岛与东亚视角。",
+        },
+        {
+            "id": "abp-news",
+            "name": "ABP News",
+            "provider": "ABP",
+            "region": "India",
+            "language": "hi",
+            "source_type": "hls",
+            "embed_url": "",
+            "stream_url": "https://abplivetv.pc.cdn.bitgravity.com/httppush/abp_livetv/abp_abpnews/master.m3u8",
+            "homepage_url": "https://news.abplive.com/live-tv",
+            "poster_url": "",
+            "is_enabled": True,
+            "is_fallback": False,
+            "sort_order": 90,
+            "collector_source": None,
+            "notes": "印度新闻直播放源，补充南亚区域视角。",
+        },
+        {
+            "id": "sabc-news",
+            "name": "SABC News",
+            "provider": "SABC",
+            "region": "Africa",
+            "language": "en",
+            "source_type": "hls",
+            "embed_url": "",
+            "stream_url": "https://sabconetanw.cdn.mangomolo.com/news/smil:news.stream.smil/playlist.m3u8",
+            "homepage_url": "https://www.sabcnews.com/sabcnews/",
+            "poster_url": "",
+            "is_enabled": True,
+            "is_fallback": False,
+            "sort_order": 100,
+            "collector_source": None,
+            "notes": "非洲英语新闻源，补充非洲区域新闻覆盖。",
+        },
+    ],
+}
+
+
+def _clean_text(value: Any) -> str:
+    if value is None:
+        return ""
+    return str(value).strip()
+
+
+def _clean_url(value: Any) -> str:
+    text = _clean_text(value)
+    if not text:
+        return ""
+
+    parsed = urlparse(text)
+    if parsed.scheme and parsed.scheme not in {"http", "https"}:
+        return ""
+    if parsed.scheme and not parsed.netloc:
+        return ""
+    return text
+
+
+def _clean_bool(value: Any, *, default: bool) -> bool:
+    if isinstance(value, bool):
+        return value
+    if value in (None, ""):
+        return default
+    if isinstance(value, str):
+        lowered = value.strip().lower()
+        if lowered in {"1", "true", "yes", "on"}:
+            return True
+        if lowered in {"0", "false", "no", "off"}:
+            return False
+    return bool(value)
+
+
+def _clean_int(value: Any, *, default: int) -> int:
+    try:
+        return int(value)
+    except (TypeError, ValueError):
+        return default
+
+
+def normalize_tv_source(source: dict[str, Any] | None, *, index: int = 0) -> dict[str, Any]:
+    payload = dict(source or {})
+    source_id = _clean_text(payload.get("id")) or f"tv-source-{index + 1}"
+    source_type = _clean_text(payload.get("source_type")).lower()
+    youtube_video_id = _clean_text(payload.get("youtube_video_id"))
+    youtube_channel = _clean_text(payload.get("youtube_channel"))
+    if source_type not in {"iframe", "hls", "video", "external", "youtube"}:
+        if youtube_video_id or youtube_channel:
+            source_type = "youtube"
+        else:
+            source_type = "iframe" if _clean_text(payload.get("embed_url")) else "external"
+
+    if source_type == "youtube" and not youtube_video_id and not youtube_channel:
+        source_type = "iframe" if _clean_text(payload.get("embed_url")) else "external"
+
+    return {
+        "id": source_id,
+        "name": _clean_text(payload.get("name")) or f"新闻直播源 {index + 1}",
+        "provider": _clean_text(payload.get("provider")) or "Unknown",
+        "region": _clean_text(payload.get("region")) or "Global",
+        "language": _clean_text(payload.get("language")) or "und",
+        "source_type": source_type,
+        "embed_url": _clean_url(payload.get("embed_url")),
+        "stream_url": _clean_url(payload.get("stream_url")),
+        "homepage_url": _clean_url(payload.get("homepage_url")),
+        "poster_url": _clean_url(payload.get("poster_url")),
+        "youtube_video_id": youtube_video_id,
+        "youtube_channel": youtube_channel,
+        "is_enabled": _clean_bool(payload.get("is_enabled"), default=True),
+        "is_fallback": _clean_bool(payload.get("is_fallback"), default=False),
+        "sort_order": _clean_int(payload.get("sort_order"), default=(index + 1) * 10),
+        "collector_source": payload.get("collector_source"),
+        "notes": _clean_text(payload.get("notes")),
+        "updated_at": _clean_text(payload.get("updated_at")),
+    }
+
+
+def normalize_tv_settings(payload: dict[str, Any] | None) -> dict[str, Any]:
+    merged = {
+        "default_source_id": DEFAULT_TV_SETTINGS["default_source_id"],
+        "auto_fallback": DEFAULT_TV_SETTINGS["auto_fallback"],
+        "sources": [],
+    }
+
+    raw_sources = []
+    if isinstance(payload, dict):
+        merged["default_source_id"] = (
+            _clean_text(payload.get("default_source_id")) or merged["default_source_id"]
+        )
+        merged["auto_fallback"] = _clean_bool(
+            payload.get("auto_fallback"),
+            default=DEFAULT_TV_SETTINGS["auto_fallback"],
+        )
+        if isinstance(payload.get("sources"), list):
+            raw_sources = payload["sources"]
+
+    if not raw_sources:
+        raw_sources = DEFAULT_TV_SETTINGS["sources"]
+
+    normalized_sources = [
+        normalize_tv_source(source, index=index)
+        for index, source in enumerate(raw_sources)
+    ]
+
+    if not any(source["id"] == DEFAULT_TV_SOURCE_ID for source in normalized_sources):
+        normalized_sources.append(
+            normalize_tv_source(DEFAULT_TV_SETTINGS["sources"][0], index=len(normalized_sources))
+        )
+
+    default_source_exists = any(
+        source["id"] == merged["default_source_id"] and source["is_enabled"]
+        for source in normalized_sources
+    )
+    if not default_source_exists:
+        fallback_source = next(
+            (source for source in normalized_sources if source["is_fallback"] and source["is_enabled"]),
+            None,
+        )
+        first_enabled_source = next(
+            (source for source in normalized_sources if source["is_enabled"]),
+            None,
+        )
+        merged["default_source_id"] = (
+            fallback_source["id"]
+            if fallback_source
+            else first_enabled_source["id"]
+            if first_enabled_source
+            else DEFAULT_TV_SOURCE_ID
+        )
+
+    merged["sources"] = sorted(
+        normalized_sources,
+        key=lambda item: (item["sort_order"], item["name"], item["id"]),
+    )
+    return merged
+
+
+async def get_tv_settings_payload(db: AsyncSession) -> dict[str, Any]:
+    result = await db.execute(
+        select(SystemSetting).where(SystemSetting.category == TV_SETTINGS_CATEGORY)
+    )
+    record = result.scalar_one_or_none()
+    payload = record.payload if record else None
+    return normalize_tv_settings(payload)
+
+
+def _build_collected_tv_source(record: CollectedData, index: int) -> dict[str, Any]:
+    metadata = dict(record.extra_data or {})
+    return normalize_tv_source(
+        {
+            "id": metadata.get("id") or record.source_id or record.entity_key,
+            "name": record.name or record.title or metadata.get("name") or f"采集直播源 {index + 1}",
+            "provider": metadata.get("provider") or metadata.get("publisher") or "Collector",
+            "region": metadata.get("region") or metadata.get("country") or "Global",
+            "language": metadata.get("language") or "und",
+            "source_type": metadata.get("source_type") or "iframe",
+            "embed_url": metadata.get("embed_url") or metadata.get("url") or "",
+            "stream_url": metadata.get("stream_url") or "",
+            "homepage_url": metadata.get("homepage_url") or metadata.get("source_url") or "",
+            "poster_url": metadata.get("poster_url") or "",
+            "youtube_video_id": metadata.get("youtube_video_id") or metadata.get("video_id") or "",
+            "youtube_channel": metadata.get("youtube_channel") or metadata.get("channel_handle") or "",
+            "is_enabled": metadata.get("is_enabled", True),
+            "is_fallback": False,
+            "sort_order": metadata.get("sort_order", 200 + index),
+            "collector_source": record.source,
+            "notes": record.description or metadata.get("notes") or "",
+            "updated_at": to_iso8601_utc(record.updated_at or record.reference_date or datetime.now(UTC)),
+        },
+        index=index,
+    )
+
+
+async def get_collected_tv_sources(db: AsyncSession) -> list[dict[str, Any]]:
+    result = await db.execute(
+        select(CollectedData)
+        .where(CollectedData.source == TV_LIVE_SOURCE_COLLECTOR)
+        .where(CollectedData.data_type == TV_LIVE_SOURCE_DATA_TYPE)
+        .where(CollectedData.is_current.is_(True))
+        .where(CollectedData.is_valid == 1)
+        .order_by(CollectedData.reference_date.desc().nullslast(), CollectedData.id.desc())
+    )
+    rows = result.scalars().all()
+    return [_build_collected_tv_source(record, index) for index, record in enumerate(rows)]
+
+
+def build_public_tv_payload(
+    settings_payload: dict[str, Any],
+    collected_sources: list[dict[str, Any]],
+) -> dict[str, Any]:
+    configured_sources = [
+        source for source in settings_payload["sources"] if source["is_enabled"]
+    ]
+
+    merged_by_id = {source["id"]: source for source in configured_sources}
+    for source in collected_sources:
+        if source["id"] in merged_by_id or not source["is_enabled"]:
+            continue
+        merged_by_id[source["id"]] = source
+
+    available_sources = sorted(
+        merged_by_id.values(),
+        key=lambda item: (item["sort_order"], item["name"], item["id"]),
+    )
+
+    default_source = next(
+        (
+            source
+            for source in available_sources
+            if source["id"] == settings_payload["default_source_id"]
+        ),
+        None,
+    )
+    fallback_source = next(
+        (source for source in available_sources if source["is_fallback"]),
+        None,
+    )
+
+    resolved_source = default_source or fallback_source or (available_sources[0] if available_sources else None)
+    latest_updated_at = max(
+        (source.get("updated_at") or "" for source in available_sources),
+        default="",
+    )
+
+    return {
+        "default_source_id": settings_payload["default_source_id"],
+        "auto_fallback": settings_payload["auto_fallback"],
+        "selected_source": resolved_source,
+        "fallback_source": fallback_source,
+        "sources": available_sources,
+        "source_count": len(available_sources),
+        "latest_updated_at": latest_updated_at or to_iso8601_utc(datetime.now(UTC)),
+        "generated_at": to_iso8601_utc(datetime.now(UTC)),
+    }
+
+
+async def get_public_tv_payload(db: AsyncSession) -> dict[str, Any]:
+    settings_payload = await get_tv_settings_payload(db)
+    collected_sources = await get_collected_tv_sources(db)
+    return build_public_tv_payload(settings_payload, collected_sources)
+
+
+def _extract_allowed_tv_hosts(sources: list[dict[str, Any]]) -> set[str]:
+    hosts: set[str] = set()
+    for source in sources:
+        for field in ("stream_url", "embed_url", "homepage_url", "youtube_channel"):
+            value = _clean_url(source.get(field))
+            if not value:
+                continue
+            parsed = urlparse(value)
+            if parsed.hostname:
+                hosts.add(parsed.hostname.lower())
+    return hosts
+
+
+def is_allowed_tv_proxy_url(url: str, sources: list[dict[str, Any]]) -> bool:
+    cleaned = _clean_url(url)
+    if not cleaned:
+        return False
+
+    parsed = urlparse(cleaned)
+    hostname = (parsed.hostname or "").lower()
+    if not hostname:
+        return False
+
+    allowed_hosts = _extract_allowed_tv_hosts(sources)
+    if hostname in allowed_hosts:
+        return True
+    return any(hostname.endswith(f".{allowed_host}") for allowed_host in allowed_hosts)

backend/tests/test_api.py

@@ -10,7 +10,12 @@ from app.core.config import settings
 from app.core.security import create_access_token
 from app.db.session import get_db
 from app.models.user import User
-from app.schemas.ai import AIProviderStatusResponse, SituationalAnalysisResponse
+from app.schemas.ai import (
+    AIProviderStatusResponse,
+    PlaygroundSessionResponse,
+    PlaygroundSessionState,
+    SituationalAnalysisResponse,
+)
 
 
 @pytest.fixture
@@ -90,6 +95,15 @@ async def test_alerts_without_auth():
     assert response.status_code == 401
 
 
+@pytest.mark.asyncio
+async def test_datasource_task_status_without_auth():
+    """Test datasource task-status endpoint requires authentication"""
+    transport = ASGITransport(app=app)
+    async with AsyncClient(transport=transport, base_url="http://test") as client:
+        response = await client.get("/api/v1/datasources/1/task-status")
+    assert response.status_code == 401
+
+
 @pytest.mark.asyncio
 async def test_alerts_endpoint_with_auth(auth_headers):
     """Test alerts endpoint with authentication"""
@@ -165,7 +179,8 @@ async def test_ai_provider_status_with_auth(auth_headers):
     class _FakeAIProviderClient:
         async def get_status(self, request_id=None):
             return AIProviderStatusResponse(
-                provider="openai_compatible",
+                provider="minimax",
+                api="anthropic-messages",
                 enabled=True,
                 configured=True,
                 model="test-model",
@@ -193,6 +208,7 @@ async def test_ai_provider_status_with_auth(auth_headers):
         assert response.status_code == 200
         data = response.json()
         assert "provider" in data
+        assert "api" in data
         assert "configured" in data
     finally:
         app.dependency_overrides.clear()
@@ -207,6 +223,9 @@ async def test_ai_situational_analysis_returns_503_when_disabled(auth_headers):
                 provider="openai_compatible",
                 model="test-model",
                 content="1) 态势摘要: 测试返回",
+                content_blocks=[],
+                text_blocks=["1) 态势摘要: 测试返回"],
+                thinking_blocks=[],
                 raw_response={"id": "mock-response"},
             )
 
@@ -241,5 +260,297 @@ async def test_ai_situational_analysis_returns_503_when_disabled(auth_headers):
         data = response.json()
         assert data["provider"] == "openai_compatible"
         assert data["content"]
+        assert "content_blocks" in data
+        assert "text_blocks" in data
+        assert "thinking_blocks" in data
+
+
+@pytest.mark.asyncio
+async def test_get_playground_session_with_auth(auth_headers):
+    """Test playground session restore endpoint."""
+
+    def override_get_current_user():
+        return User(
+            id=1,
+            username="testuser",
+            email="test@example.com",
+            password_hash="hashed",
+            role="admin",
+            is_active=True,
+        )
+
+    async def override_get_db():
+        yield AsyncMock()
+
+    app.dependency_overrides = {
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+        get_db: override_get_db,
+    }
+    transport = ASGITransport(app=app)
+    try:
+        with patch(
+            "app.api.v1.ai.get_playground_session",
+            new=AsyncMock(
+                return_value=PlaygroundSessionResponse(
+                    id="1",
+                    session_key="default",
+                    title="Playground 会话",
+                    state=PlaygroundSessionState(
+                        messages=[{"id": "msg-1", "role": "user", "content": "hello"}],
+                        title="测试标题",
+                        objective="测试目标",
+                    ),
+                    created_at="2026-04-10T00:00:00+00:00",
+                    updated_at="2026-04-10T00:00:00+00:00",
+                )
+            ),
+        ):
+            async with AsyncClient(transport=transport, base_url="http://test") as client:
+                response = await client.get("/api/v1/ai/playground/session", headers=auth_headers)
+        assert response.status_code == 200
+        data = response.json()
+        assert data["session_key"] == "default"
+        assert data["state"]["messages"][0]["content"] == "hello"
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_save_playground_session_with_auth(auth_headers):
+    """Test playground session save endpoint."""
+
+    def override_get_current_user():
+        return User(
+            id=1,
+            username="testuser",
+            email="test@example.com",
+            password_hash="hashed",
+            role="admin",
+            is_active=True,
+        )
+
+    async def override_get_db():
+        yield AsyncMock()
+
+    app.dependency_overrides = {
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+        get_db: override_get_db,
+    }
+    transport = ASGITransport(app=app)
+    try:
+        with patch(
+            "app.api.v1.ai.upsert_playground_session",
+            new=AsyncMock(
+                return_value=PlaygroundSessionResponse(
+                    id="1",
+                    session_key="default",
+                    title="测试标题",
+                    state=PlaygroundSessionState(
+                        messages=[{"id": "msg-1", "role": "user", "content": "hello"}],
+                        title="测试标题",
+                        objective="测试目标",
+                    ),
+                    created_at="2026-04-10T00:00:00+00:00",
+                    updated_at="2026-04-10T00:00:00+00:00",
+                )
+            ),
+        ):
+            async with AsyncClient(transport=transport, base_url="http://test") as client:
+                response = await client.put(
+                    "/api/v1/ai/playground/session",
+                    headers=auth_headers,
+                    json={
+                        "session_key": "default",
+                        "title": "测试标题",
+                        "state": {
+                            "messages": [{"id": "msg-1", "role": "user", "content": "hello"}],
+                            "selectedPresetKey": "bgp-brief",
+                            "title": "测试标题",
+                            "objective": "测试目标",
+                            "constraints": "",
+                            "inputValue": "",
+                            "analysis": None,
+                            "latestAnalysisMessageId": None,
+                            "analysisMeta": {},
+                            "helpExpanded": True,
+                        },
+                    },
+                )
+        assert response.status_code == 200
+        data = response.json()
+        assert data["title"] == "测试标题"
+        assert data["state"]["objective"] == "测试目标"
+    finally:
+        app.dependency_overrides.clear()
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_ai_bgp_brief_endpoint_persists_fact_snapshot(auth_headers):
+    class _FakeAIProviderClient:
+        async def analyze(self, _payload, request_id=None):
+            return SituationalAnalysisResponse(
+                provider="minimax",
+                model="MiniMax-M2.5",
+                content="# BGP AI 简报\n\n事实摘要：测试",
+                content_blocks=[],
+                text_blocks=["# BGP AI 简报\n\n事实摘要：测试"],
+                thinking_blocks=[],
+                raw_response={"id": "mock-bgp-brief"},
+            )
+
+    def override_get_current_user():
+        return User(
+            id=1,
+            username="testuser",
+            email="test@example.com",
+            password_hash="hashed",
+            role="admin",
+            is_active=True,
+        )
+
+    async def override_get_db():
+        yield AsyncMock()
+
+    async def _fake_build_bgp_brief_request(_db, **_kwargs):
+        request_payload = __import__("app.schemas.ai", fromlist=["SituationalAnalysisRequest"]).SituationalAnalysisRequest(
+            title="BGP 态势 AI 简报",
+            objective="生成值班简报",
+            observations=["事实A", "事实B"],
+            constraints=["不要编造"],
+            context={"incident_total": 2, "active_collectors": 3},
+        )
+        return request_payload, ["事实A", "事实B"], {"incident_total": 2, "active_collectors": 3}
+
+    app.dependency_overrides = {
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+        __import__("app.services.ai_client", fromlist=["get_ai_provider_client"]).get_ai_provider_client: lambda: _FakeAIProviderClient(),
+        get_db: override_get_db,
+    }
+    transport = ASGITransport(app=app)
+    try:
+        with patch("app.api.v1.ai.build_bgp_brief_request", side_effect=_fake_build_bgp_brief_request):
+            async with AsyncClient(transport=transport, base_url="http://test") as client:
+                response = await client.post("/api/v1/ai/bgp/brief", headers=auth_headers, json={})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["facts"] == ["事实A", "事实B"]
+        assert data["context"]["incident_total"] == 2
+        assert data["content_markdown"]
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_ai_alert_brief_endpoint_with_auth(auth_headers):
+    class _FakeAIProviderClient:
+        async def analyze(self, _payload, request_id=None):
+            return SituationalAnalysisResponse(
+                provider="minimax",
+                model="MiniMax-M2.7",
+                content="事实摘要：告警测试。风险研判：告警测试。建议动作：告警测试。",
+                content_blocks=[],
+                text_blocks=["事实摘要：告警测试。风险研判：告警测试。建议动作：告警测试。"],
+                thinking_blocks=[],
+                raw_response={"id": "mock-alert-brief"},
+            )
+
+    def override_get_current_user():
+        return User(
+            id=1,
+            username="testuser",
+            email="test@example.com",
+            password_hash="hashed",
+            role="admin",
+            is_active=True,
+        )
+
+    async def override_get_db():
+        yield AsyncMock()
+
+    async def _fake_build_alert_brief_request(_db, **_kwargs):
+        request_payload = __import__("app.schemas.ai", fromlist=["SituationalAnalysisRequest"]).SituationalAnalysisRequest(
+            title="告警态势 AI 简报",
+            objective="输出告警简报",
+            observations=["告警事实A", "告警事实B"],
+            constraints=["不要编造"],
+            context={"active_alerts": 3, "top_datasources": {"bgp": 2}},
+        )
+        return request_payload, ["告警事实A", "告警事实B"], {"active_alerts": 3, "top_datasources": {"bgp": 2}}
+
+    app.dependency_overrides = {
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+        __import__("app.services.ai_client", fromlist=["get_ai_provider_client"]).get_ai_provider_client: lambda: _FakeAIProviderClient(),
+        get_db: override_get_db,
+    }
+    transport = ASGITransport(app=app)
+    try:
+        with patch("app.api.v1.ai.build_alert_brief_request", side_effect=_fake_build_alert_brief_request):
+            async with AsyncClient(transport=transport, base_url="http://test") as client:
+                response = await client.post("/api/v1/ai/alerts/brief", headers=auth_headers, json={})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["title"] == "告警态势 AI 简报"
+        assert data["facts"] == ["告警事实A", "告警事实B"]
+        assert data["context"]["active_alerts"] == 3
+        assert data["content"]
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_ai_situational_alert_brief_endpoint_with_auth(auth_headers):
+    class _FakeAIProviderClient:
+        async def analyze(self, _payload, request_id=None):
+            return SituationalAnalysisResponse(
+                provider="minimax",
+                model="MiniMax-M2.7",
+                content="事实摘要：态势测试。风险研判：态势测试。建议动作：态势测试。",
+                content_blocks=[],
+                text_blocks=["事实摘要：态势测试。风险研判：态势测试。建议动作：态势测试。"],
+                thinking_blocks=[],
+                raw_response={"id": "mock-situational-brief"},
+            )
+
+    def override_get_current_user():
+        return User(
+            id=1,
+            username="testuser",
+            email="test@example.com",
+            password_hash="hashed",
+            role="admin",
+            is_active=True,
+        )
+
+    async def override_get_db():
+        yield AsyncMock()
+
+    async def _fake_build_situational_alert_brief_request(_db):
+        request_payload = __import__("app.schemas.ai", fromlist=["SituationalAnalysisRequest"]).SituationalAnalysisRequest(
+            title="态势告警 AI 简报",
+            objective="输出态势告警简报",
+            observations=["态势事实A", "态势事实B"],
+            constraints=["不要编造"],
+            context={"active_system_alerts": 2, "active_bgp_incidents": 1},
+        )
+        return request_payload, ["态势事实A", "态势事实B"], {"active_system_alerts": 2, "active_bgp_incidents": 1}
+
+    app.dependency_overrides = {
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+        __import__("app.services.ai_client", fromlist=["get_ai_provider_client"]).get_ai_provider_client: lambda: _FakeAIProviderClient(),
+        get_db: override_get_db,
+    }
+    transport = ASGITransport(app=app)
+    try:
+        with patch("app.api.v1.ai.build_situational_alert_brief_request", side_effect=_fake_build_situational_alert_brief_request):
+            async with AsyncClient(transport=transport, base_url="http://test") as client:
+                response = await client.post("/api/v1/ai/situational-alerts/brief", headers=auth_headers, json={})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["title"] == "态势告警 AI 简报"
+        assert data["facts"] == ["态势事实A", "态势事实B"]
+        assert data["context"]["active_system_alerts"] == 2
+        assert data["content"]
     finally:
         app.dependency_overrides.clear()

docker-compose.yml

@@ -5,6 +5,8 @@ services:
     build:
       context: .
       dockerfile: aiprovider/Dockerfile
+    env_file:
+      - ./aiprovider/.env
     container_name: planet_aiprovider
     ports:
       - "8010:8010"

docs/CHANGELOG.md

@@ -7,6 +7,416 @@ This project follows the repository versioning rule:
 - `feature` -> `+0.1.0`
 - `bugfix` -> `+0.0.1`
 
+## 0.27.0
+
+Released: 2026-04-14
+
+### Highlights
+
+- 全面重构 Earth HUD 布局：品牌面板、图层控制面板、信息详情卡片各自独立，支持拖拽与折叠，信息卡片改为跟随点击位置悬浮显示。
+- 新增图层控制面板（Layer Panel），海缆、卫星、地形、BGP 等图层集中管理，支持关键字搜索过滤。
+- Earth 大气层渲染升级，引入 Fresnel 着色器双层辉光效果和深度遮挡球体。
+
+### ✨ Features
+
+- 新增独立 Layer Panel（`js/controls.js`, `css/layer-panel.css`），图层开关、搜索过滤、折叠收起，取代原工具栏弹出菜单
+- 信息详情面板（info-panel）改为点击时定位到鼠标附近（`js/info-card.js`），悬停改为轻量 tooltip，降低视觉干扰
+- Earth 材质重构（`js/earth.js`, `js/constants.js`）：新增 `EARTH_MATERIAL_CONFIG`，Fresnel 内外大气层辉光、深度遮挡球体，纹理加载独立为 `loadEarthTexture()`
+
+### 🔧 Improvements
+
+- Earth Stats 面板改为 2 列 KPI 网格布局，支持拖拽和关闭（`css/earth-stats.css`）
+- 数据加载改为分步串行（登陆点 → 海缆 → 卫星 → BGP → 纹理），每步之间 yield 帧，改善视觉渐现体验（`js/main.js`）
+- 图层按钮状态更新逻辑统一至 `updateLayerButtonState()`，消除重复实现
+- 海缆状态识别新增 `active` 枚举值（兼容旧 `In Service`）
+
+---
+
+## 0.26.1
+
+Released: 2026-04-12
+
+### Highlights
+
+- Cleaned up the first TV follow-up and replaced the dashboard sidebar's brittle one-off scroll behavior with a reusable `Scrollbar` component, so the new live module code is easier to maintain and the console navigation can scroll without layout jitter.
+
+### Improved
+
+- Improved [frontend/src/components/Scrollbar/Scrollbar.tsx](/home/ray/dev/linkong/planet/frontend/src/components/Scrollbar/Scrollbar.tsx), [frontend/src/components/AppLayout/AppLayout.tsx](/home/ray/dev/linkong/planet/frontend/src/components/AppLayout/AppLayout.tsx), and [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css) by extracting the sidebar scrollbar into a dedicated component with explicit `x / y / both` axis support, hidden native scrollbars, compact account/version rows, and a sidebar-only vertical setup instead of the earlier patchwork CSS glued directly onto the layout.
+
+### Fixed
+
+- Fixed [frontend/public/earth/js/tv.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/tv.js) by refactoring repeated iframe/video reset paths into shared helpers, so TV source switching, empty-state fallback, and playback retry handling no longer duplicate cleanup logic across multiple branches.
+
+## 0.26.2
+
+Released: 2026-04-12
+
+### Highlights
+
+- Stabilized the new reusable scrollbar work by restoring reliable sidebar visibility and extending the same floating scrollbar language to the Data Sources tables without letting scrollbars squeeze layout width or regress the console navigation.
+
+### Added
+
+- Added [frontend/src/components/Scrollbar/ScrollbarOverlay.tsx](/home/ray/dev/linkong/planet/frontend/src/components/Scrollbar/ScrollbarOverlay.tsx) as an overlay variant that binds to existing scroll containers such as Ant Table bodies, so heavy data grids can adopt the new scrollbar visuals without replacing their built-in scrolling mechanics.
+
+### Improved
+
+- Improved [frontend/src/components/Scrollbar/Scrollbar.tsx](/home/ray/dev/linkong/planet/frontend/src/components/Scrollbar/Scrollbar.tsx), [frontend/src/components/AppLayout/AppLayout.tsx](/home/ray/dev/linkong/planet/frontend/src/components/AppLayout/AppLayout.tsx), and [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css) by reverting the sidebar to a reliable vertical-first scrollbar path, then reintroducing automatic dual-axis support with independent floating tracks that no longer hide the thumb when only the sidebar needs vertical scrolling.
+
+### Fixed
+
+- Fixed [frontend/src/pages/DataSources/DataSources.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/DataSources/DataSources.tsx) and [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css) so the built-in and custom datasource tables now use the new overlay scrollbar instead of native table scrollbars, keeping horizontal and vertical scrolling available without changing Ant Table’s internal layout behavior.
+
+## 0.26.0
+
+Released: 2026-04-12
+
+### Highlights
+
+- Added an operator-facing TV live module to Earth, including backend-configurable live sources, a draggable/resizable live-news HUD window, default global news channels, and a dedicated settings workflow so the Earth page can open real news playback instead of only static telemetry.
+
+### Added
+
+- Added [backend/app/api/v1/tv.py](/home/ray/dev/linkong/planet/backend/app/api/v1/tv.py), [backend/app/services/tv_streams.py](/home/ray/dev/linkong/planet/backend/app/services/tv_streams.py), and [backend/app/services/collectors/news_live_streams.py](/home/ray/dev/linkong/planet/backend/app/services/collectors/news_live_streams.py) to provide TV source configuration, public stream payloads, a guarded HLS proxy path, and a collector entry point for future world-news live-source ingestion.
+- Added the Earth TV HUD workspace through [frontend/public/earth/index.html](/home/ray/dev/linkong/planet/frontend/public/earth/index.html), [frontend/public/earth/js/tv.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/tv.js), and [frontend/public/earth/css/tv-panel.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/tv-panel.css), including toolbar access, draggable/closable behavior, resize support, direct video/HLS playback, iframe fallback, and per-channel external-open handling.
+- Added [docs/earth-tv-live-module-plan.md](/home/ray/dev/linkong/planet/docs/earth-tv-live-module-plan.md) and [docs/news-live-streams-collector-format.md](/home/ray/dev/linkong/planet/docs/news-live-streams-collector-format.md) to document the TV module rollout plan and the expected collector payload format for future curated live-channel ingestion.
+
+### Improved
+
+- Improved [frontend/src/pages/Settings/Settings.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Settings/Settings.tsx), [backend/app/api/v1/settings.py](/home/ray/dev/linkong/planet/backend/app/api/v1/settings.py), and [backend/app/core/datasource_defaults.py](/home/ray/dev/linkong/planet/backend/app/core/datasource_defaults.py) by adding TV source administration to system settings and registering the `news_live_streams` datasource as a first-class configurable collector.
+- Improved [backend/app/services/tv_streams.py](/home/ray/dev/linkong/planet/backend/app/services/tv_streams.py) by seeding a curated first-pass news channel catalog that now defaults to `CGTN English` YouTube playback while keeping `CCTV-4` as a built-in fallback and exposing additional Reuters, CGTN, DW, Al Jazeera, Arirang, ABP, and SABC entries for operator testing.
+
+### Fixed
+
+- Fixed [frontend/public/earth/js/tv.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/tv.js) and [frontend/public/earth/index.html](/home/ray/dev/linkong/planet/frontend/public/earth/index.html) so HLS/video playback now actively attempts autoplay in the TV panel instead of only loading metadata and leaving the player visually idle.
+- Fixed [frontend/public/earth/css/tv-panel.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/tv-panel.css) so the TV source selector and action controls better match the Earth HUD dark theme instead of falling back to a bright native dropdown presentation.
+
+## 0.25.3
+
+Released: 2026-04-11
+
+### Highlights
+
+- Refined the Earth HUD visual system into a calmer operator-facing style, turned the top-left Earth brand into a real reusable component with language-driven rendering, and cleaned up duplicated brand assets so the page now has a single source of truth for HUD branding.
+
+### Added
+
+- Added [frontend/public/earth/js/brand.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/brand.js) as a reusable Earth `brand` component that renders the top-left logo/title/subtitle block from a shared config instead of hardcoding the structure in HTML.
+- Added [frontend/public/earth/assets/brand/earth-logo.svg](/home/ray/dev/linkong/planet/frontend/public/earth/assets/brand/earth-logo.svg), [frontend/public/earth/assets/brand/title-zh.svg](/home/ray/dev/linkong/planet/frontend/public/earth/assets/brand/title-zh.svg), and [frontend/public/earth/assets/brand/title-en.svg](/home/ray/dev/linkong/planet/frontend/public/earth/assets/brand/title-en.svg) as the canonical Earth brand asset set.
+
+### Improved
+
+- Improved [frontend/public/earth/css/base.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/base.css), [frontend/public/earth/css/hud.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/hud.css), [frontend/public/earth/css/coordinates-display.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/coordinates-display.css), [frontend/public/earth/css/legend.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/legend.css), [frontend/public/earth/css/earth-stats.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/earth-stats.css), and [frontend/public/earth/css/toolbar.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/toolbar.css) by rebalancing the Earth HUD into a more restrained deep-blue control-room look instead of the earlier over-layered glass-and-neon mix.
+- Improved [frontend/public/earth/index.html](/home/ray/dev/linkong/planet/frontend/public/earth/index.html), [frontend/public/earth/js/main.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/main.js), and [frontend/public/earth/js/constants.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/constants.js) by moving the Earth brand mount to a dedicated root and letting `HUD_CONFIG.brandLanguage` choose between `zh` and `en` without embedding the language decision in the DOM.
+
+### Fixed
+
+- Fixed [frontend/public/earth/css/info-panel.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/info-panel.css) so the Earth brand area now uses consistent `earth-brand` component selectors and English-specific typography hooks, avoiding the earlier one-off `brand-banner` naming drift and duplicated title styles.
+
+## 0.25.2
+
+Released: 2026-04-10
+
+### Highlights
+
+- Refined the Earth HUD operator polish so settings now behave like a true bounded menu, HUD panels render at the correct scale from the first frame, and dragged panels animate cleanly into and out of maximized layout targets without drifting to screen edges.
+
+### Improved
+
+- Improved [frontend/public/earth/index.html](/home/ray/dev/linkong/planet/frontend/public/earth/index.html) by precomputing the initial `--hud-scale` before Earth CSS loads, so HUD panels no longer flash at full size before shrinking to the target scale.
+- Improved [frontend/public/earth/css/hud.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/hud.css) by cleaning up duplicated settings-modal close-button styles, aligning the settings title with the shared HUD title system, and constraining the settings sheet to a stable centered width instead of viewport-relative modal sizing.
+
+### Fixed
+
+- Fixed [frontend/public/earth/js/controls.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/controls.js) so dragged HUD panels now fly from their dragged positions into maximized layout targets, closed panels stay out of the transition, and restoring layout clears drag overrides back to the initial positions.
+- Fixed [frontend/public/earth/js/controls.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/controls.js) and [frontend/public/earth/css/hud.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/hud.css) so layout transitions no longer visibly stick to screen edges before landing; the FLIP motion now composes with the existing corner transforms instead of fighting them.
+
+## 0.25.1
+
+Released: 2026-04-10
+
+### Highlights
+
+- Cleaned up the first persistent Playground rollout, fixed sidebar submenu persistence to match the intended operator behavior, and added reusable code-hygiene rules to prevent this class of drift from accumulating again.
+
+### Improved
+
+- Improved [backend/app/services/playground_chat_service.py](/home/ray/dev/linkong/planet/backend/app/services/playground_chat_service.py) and [frontend/src/pages/Playground/Playground.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Playground/Playground.tsx) by extracting repeated lookup, response, and request-action paths into shared helpers, reducing duplicated Playground flow code without changing behavior.
+- Improved [rules.md](/home/ray/dev/linkong/planet/rules.md) by adding a new `Code Hygiene - MANDATORY` section covering single-source-of-truth state, transitional cleanup, repeated-logic extraction, layout debugging order, and post-feature cleanup expectations.
+
+### Fixed
+
+- Fixed [frontend/src/components/AppLayout/AppLayout.tsx](/home/ray/dev/linkong/planet/frontend/src/components/AppLayout/AppLayout.tsx) so first-level menu expansion now behaves correctly across in-app navigation: `采集与数据` remains the default expanded group after refresh, while manually expanded groups stay open when navigating to their own child routes and reset only on page reload.
+
+## 0.25.0
+
+Released: 2026-04-10
+
+### Highlights
+
+- Turned `AI Playground` into a persistent backend-backed chat workspace, added dedicated alert workspaces as a foundation for future situational analysis, and aligned the operator UI around a more structured AI + alerts workflow instead of one-off playground calls.
+
+### Added
+
+- Added [backend/app/models/playground_session.py](/home/ray/dev/linkong/planet/backend/app/models/playground_session.py), [backend/app/models/playground_message.py](/home/ray/dev/linkong/planet/backend/app/models/playground_message.py), and [backend/app/services/playground_chat_service.py](/home/ray/dev/linkong/planet/backend/app/services/playground_chat_service.py) so Playground conversations, execution state, edits, retries, and stop/resume semantics are persisted in the backend database rather than living only in browser state.
+- Added [backend/app/services/alert_ai_brief.py](/home/ray/dev/linkong/planet/backend/app/services/alert_ai_brief.py), [backend/app/services/situational_alert_ai_brief.py](/home/ray/dev/linkong/planet/backend/app/services/situational_alert_ai_brief.py), and the dedicated alert pages [frontend/src/pages/Alerts/SystemAlerts.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Alerts/SystemAlerts.tsx), [frontend/src/pages/Alerts/BGPAlerts.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Alerts/BGPAlerts.tsx), and [frontend/src/pages/Alerts/SituationalAlerts.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Alerts/SituationalAlerts.tsx) to establish the alert-analysis foundation for later situational awareness expansion.
+
+### Improved
+
+- Improved [frontend/src/pages/Playground/Playground.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Playground/Playground.tsx) and [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css) by rebuilding Playground into a true chatbox workflow with persistent history, edit-and-resend behavior, grounded message actions, responsive composer behavior, bottom-stick scrolling, and tighter mobile layout handling.
+- Improved [frontend/src/components/AppLayout/AppLayout.tsx](/home/ray/dev/linkong/planet/frontend/src/components/AppLayout/AppLayout.tsx), [frontend/src/App.tsx](/home/ray/dev/linkong/planet/frontend/src/App.tsx), and [frontend/src/pages/Alerts/Alerts.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Alerts/Alerts.tsx) by reorganizing navigation around `采集与数据`, `专题观测`, and split alert entries so the app can scale to more observability and situational modules without turning the top-level UI into a single overloaded page.
+- Improved [README.md](/home/ray/dev/linkong/planet/README.md) and [docs/situational-awareness-foundation-plan.md](/home/ray/dev/linkong/planet/docs/situational-awareness-foundation-plan.md) by documenting the current AI/alerts base, planned situational-awareness direction, and the new persistent Playground foundation.
+
+### Fixed
+
+- Fixed [backend/app/services/playground_chat_service.py](/home/ray/dev/linkong/planet/backend/app/services/playground_chat_service.py) so background Playground runs explicitly commit state transitions, allowing frontend polling to observe real pending/thinking/answering/done states instead of seeing stale empty threads.
+- Fixed [frontend/src/services/situational-awareness/index.ts](/home/ray/dev/linkong/planet/frontend/src/services/situational-awareness/index.ts) by removing the temporary mock gateway path, so Playground and alert-related AI flows now reflect the real backend/provider chain instead of local fake responses.
+
+## 0.24.8
+
+Released: 2026-04-10
+
+### Highlights
+
+- Refined the BGP AI brief operator workflow so the tab now stays compact and metadata-focused, while the full Markdown brief opens in a bounded modal that respects the repo’s single-screen layout rules.
+
+### Improved
+
+- Improved [frontend/src/pages/BGP/BGP.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/BGP/BGP.tsx) by turning the brief action row into a clearer `历史简报下拉框 + 查看 + 生成` flow, keeping inline metadata visible in the tab while moving full Markdown reading into a dedicated modal workspace.
+
+### Fixed
+
+- Fixed [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css) and [frontend/src/pages/BGP/BGP.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/BGP/BGP.tsx) so the BGP brief modal no longer spills below the viewport; long brief content now scrolls inside the modal body instead of extending past the visible screen.
+
+## 0.24.7
+
+Released: 2026-04-10
+
+### Highlights
+
+- Formalized repository release hygiene and frontend layout guardrails so repeated versioning chores and recurring layout regressions now have explicit repo-level rules instead of living only in conversation context.
+
+### Added
+
+- Added [release-workflow/SKILL.md](/home/ray/dev/linkong/planet/.codex/skills/release-workflow/SKILL.md), defining the repository release workflow for version bumps, changelog/version-history updates, minimal validation, and commit/push sequencing.
+
+### Improved
+
+- Improved [rules.md](/home/ray/dev/linkong/planet/rules.md) by adding mandatory release-workflow requirements and a new frontend layout constraint section covering single-screen workspaces, overflow ownership, tab-pane behavior, compact-mode expectations, and readable-card fallbacks.
+- Improved [frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/frontend-layout-guidelines.md) by summarizing the recurring Earth, Playground, BGP, and admin-layout regressions into concrete constraints for future frontend work, including “prefer scrollbars over unreadable compression” and “do not treat every tab as a table pane.”
+
+## 0.24.6
+
+Released: 2026-04-10
+
+### Highlights
+
+- Tightened several backend hot paths outside the original BGP page fixes, stabilized BGP collector coverage after the recent query refactors, and rebuilt the BGP AI brief tab so saved Markdown briefs render and scroll like a proper operator workspace instead of collapsing inside the shared table layout.
+
+### Improved
+
+- Improved [backend/app/api/v1/datasources.py](/home/ray/dev/linkong/planet/backend/app/api/v1/datasources.py) by replacing per-datasource task, count, and endpoint lookups with batched prefetch helpers, reducing the worst `1 + N` behavior on the datasource list and `trigger-all` flow.
+- Improved [backend/app/api/v1/visualization.py](/home/ray/dev/linkong/planet/backend/app/api/v1/visualization.py) by switching the main Earth-facing `CollectedData` endpoints to `is_current` records, batching multi-source loads for aggregate endpoints, and removing stale Python-side dedupe paths from the hot route.
+- Improved [backend/app/services/bgp_incidents.py](/home/ray/dev/linkong/planet/backend/app/services/bgp_incidents.py) and [backend/app/services/bgp_enrichment.py](/home/ray/dev/linkong/planet/backend/app/services/bgp_enrichment.py) by avoiding historical full-table infrastructure scans, narrowing observation baseline payloads to required columns, and pushing more ASN filtering into the database.
+- Improved [backend/app/api/v1/alerts.py](/home/ray/dev/linkong/planet/backend/app/api/v1/alerts.py), [backend/app/api/v1/dashboard.py](/home/ray/dev/linkong/planet/backend/app/api/v1/dashboard.py), and [backend/app/api/v1/settings.py](/home/ray/dev/linkong/planet/backend/app/api/v1/settings.py) by collapsing several repeated count and settings queries into fewer aggregate or batched reads.
+- Improved [frontend/src/pages/BGP/BGP.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/BGP/BGP.tsx), [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css), and [frontend/src/components/MarkdownRenderer/MarkdownRenderer.tsx](/home/ray/dev/linkong/planet/frontend/src/components/MarkdownRenderer/MarkdownRenderer.tsx) by rebuilding the `AI 简报` tab layout, fixing saved brief scrolling behavior, and extending the renderer to handle tables, separators, and stored metadata comments more gracefully.
+- Improved [docs/ai-playground-development-plan.md](/home/ray/dev/linkong/planet/docs/ai-playground-development-plan.md) by explicitly recording that the current BGP brief is only the first-stage summary flow and that regional prefix-geography analysis remains a planned Phase B follow-up.
+
+### Fixed
+
+- Fixed [backend/app/services/bgp_collectors.py](/home/ray/dev/linkong/planet/backend/app/services/bgp_collectors.py) and [backend/app/services/bgp_enrichment.py](/home/ray/dev/linkong/planet/backend/app/services/bgp_enrichment.py) so JSON field extraction no longer depends on the less portable `.astext` path that could break BGP collector endpoints in local environments.
+- Fixed the BGP AI brief tab in [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css) and [frontend/src/pages/BGP/BGP.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/BGP/BGP.tsx) so long saved briefs are no longer compressed into a tiny clipped viewport by the shared tab/table overflow rules.
+
+## 0.24.4
+
+Released: 2026-04-09
+
+### Highlights
+
+- Refined the `planet.sh` AI Provider rebuild UX so image rebuilds now feel like first-class scripted tasks, with clearer stage boundaries and cleaner fallback messaging instead of leaking raw Compose output into the terminal.
+
+### Improved
+
+- Improved [planet.sh](/home/ray/dev/linkong/planet/planet.sh) by keeping AI Provider image build logs in a temporary file, surfacing stage-specific detail copy for `docker compose v2` and `docker-compose v1`, and showing an explicit success line when the image rebuild finishes before container health checks begin.
+
+### Fixed
+
+- Fixed [planet.sh](/home/ray/dev/linkong/planet/planet.sh) so the AI Provider image rebuild stage no longer dumps raw Compose build output into the main spinner flow during normal successful runs.
+- Fixed [planet.sh](/home/ray/dev/linkong/planet/planet.sh) so the `构建 AI Provider 镜像` phase now ends with an explicit completion signal instead of visually blending into the subsequent container health-check phase.
+
+## 0.24.3
+
+Released: 2026-04-09
+
+### Highlights
+
+- Extended `AI Playground` from a minimal prompt form into a more repeatable diagnostics workspace, and hardened `planet.sh` so AI Provider restarts can rebuild changed images and expose clearer Compose fallback behavior.
+
+### Improved
+
+- Improved [frontend/src/pages/Playground/Playground.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Playground/Playground.tsx) by adding preset scenarios, response metadata, `thinking` block visibility, raw JSON inspection, and copy actions so the page works more like a proper AI diagnostics console.
+- Improved [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css) by styling Playground presets, result metadata, raw response sections, and responsive action groups without breaking the single-screen workspace layout.
+- Improved [planet.sh](/home/ray/dev/linkong/planet/planet.sh) so AI Provider restarts detect code/config changes, rebuild the image when needed, and surface which Compose path is being used during image and container operations.
+
+### Fixed
+
+- Fixed the AI Provider restart path in [planet.sh](/home/ray/dev/linkong/planet/planet.sh) so code changes inside `aiprovider/` no longer stay hidden behind an old container image after `restart -a`.
+- Fixed Compose error reporting in [planet.sh](/home/ray/dev/linkong/planet/planet.sh) by making the script explicitly show `docker compose v2` first, then `docker-compose v1`, and only fail after both execution paths are exhausted.
+
+## 0.24.2
+
+Released: 2026-04-09
+
+### Highlights
+
+- Fixed the public-entry and AI diagnostics regressions introduced during the recent frontend routing and playground work, while also reducing the main frontend bundle by switching to route-level lazy loading and more targeted vendor chunking.
+
+### Improved
+
+- Improved [frontend/src/App.tsx](/home/ray/dev/linkong/planet/frontend/src/App.tsx) by lazy-loading admin pages and large workspaces through `React.lazy()` plus `Suspense`, so the initial frontend entry no longer pulls every route into the first bundle.
+- Improved [frontend/vite.config.ts](/home/ray/dev/linkong/planet/frontend/vite.config.ts) by adding targeted manual chunking for React, icon, network, and Earth-related vendor dependencies instead of leaving everything in one monolithic application bundle.
+- Improved [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css) by adding a shared route-loading state and making the Playground help panel size to its content instead of stretching to fill the sidebar.
+
+### Fixed
+
+- Fixed [frontend/src/App.tsx](/home/ray/dev/linkong/planet/frontend/src/App.tsx) so anonymous visits to `/` once again reach the public Earth entry through the existing `/ -> /earth` redirect instead of being intercepted by the login screen.
+- Fixed [frontend/src/pages/Playground/Playground.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Playground/Playground.tsx) so cached provider status is shown immediately but still refreshed from the backend, avoiding stale diagnostics after `.env` or provider changes within the same browser tab.
+- Fixed [frontend/src/pages/Playground/Playground.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Playground/Playground.tsx) and [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css) so the `测试说明` card no longer over-expands and now fits its content more naturally in the sidebar.
+
+## 0.24.1
+
+Released: 2026-04-09
+
+### Highlights
+
+- Refined the `/earth` HUD into a cleaner class-first structure with responsive scaling, clearer CSS layer boundaries, and lower coupling between HTML, CSS, and runtime UI updates.
+- Hardened local startup conventions around Bun so frontend tooling, docs, and `planet.sh` behave more predictably in fresh Ubuntu and mixed WSL environments.
+
+### Added
+
+- Added [frontend/public/earth/css/hud.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/hud.css), extracting shared HUD panel surfaces, shared HUD typography rows, status messaging, tooltip overlays, and layout-expanded panel transitions out of the old monolithic base layer.
+- Added [frontend/public/earth/css/toolbar.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/toolbar.css), isolating Earth toolbar, popover, zoom dock, liquid-glass button, and toolbar-tooltip behavior into a dedicated toolbar layer.
+- Added explicit Bun package-manager metadata to [frontend/package.json](/home/ray/dev/linkong/planet/frontend/package.json) so the frontend package manager choice is declared instead of inferred from the lockfile alone.
+
+### Improved
+
+- Improved [frontend/public/earth/css/base.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/base.css) by reducing it to app-shell concerns only: global tokens, Earth app container, loading panel, and shared animation primitives.
+- Improved [frontend/public/earth/index.html](/home/ray/dev/linkong/planet/frontend/public/earth/index.html) by wiring the new CSS layer order, standardizing HUD utility classes on `hud-panel-*`, and replacing generic toolbar/tooltip hooks with more explicit Earth-specific classes.
+- Improved [frontend/public/earth/css/info-panel.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/info-panel.css), [frontend/public/earth/css/coordinates-display.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/coordinates-display.css), [frontend/public/earth/css/legend.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/legend.css), and [frontend/public/earth/css/earth-stats.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/earth-stats.css) by leaving only panel-specific responsibilities in each file after the shared HUD and toolbar primitives moved out.
+- Improved [frontend/public/earth/js/main.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/main.js), [frontend/public/earth/js/controls.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/controls.js), [frontend/public/earth/js/ui.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/ui.js), and [frontend/public/earth/js/legend.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/legend.js) by aligning runtime DOM queries and generated markup with the new class-first HUD and toolbar structure.
+- Improved [frontend/public/earth/js/constants.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/constants.js) and [frontend/public/earth/js/main.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/main.js) by keeping HUD scaling configurable through extracted constants instead of scattering scaling assumptions through the main Earth entrypoint.
+- Improved [planet.sh](/home/ray/dev/linkong/planet/planet.sh) by prepending `~/.local/bin` and `~/.bun/bin` automatically, preferring direct `~/.bun/bin/bun` detection, and auto-installing missing `uv`/`bun` instead of requiring the user's interactive shell config to expose them first.
+- Improved [README.md](/home/ray/dev/linkong/planet/README.md), [project_context.md](/home/ray/dev/linkong/planet/project_context.md), [rules.md](/home/ray/dev/linkong/planet/rules.md), and [scripts/bootstrap-dev.sh](/home/ray/dev/linkong/planet/scripts/bootstrap-dev.sh) by making the frontend Bun-only workflow explicit in both onboarding docs and command-line guidance.
+
+### Fixed
+
+- Fixed the Earth HUD cleanup path where splitting styles previously left the page with a missing `base.css` entry and mismatched utility class names during the refactor.
+- Fixed several Earth UI update paths so toolbar tooltip text, legend re-rendering, status message classes, and mouse coordinate styling continue to work after removing legacy fallback selectors.
+- Fixed the local developer bootstrap path where `planet.sh start` could fail in a clean Ubuntu or agent shell session simply because Bun or uv were installed outside the current shell's inherited `PATH`.
+
+## 0.24.0
+
+Released: 2026-04-09
+
+### Highlights
+
+- Added a dedicated `AI Playground` admin entry so operators can validate provider connectivity and run controlled situational-analysis prompts from the main frontend without introducing a separate UI service.
+- Established a first explicit frontend layout rulebook centered on single-screen workspaces, internal module scrolling, and BGP-style page composition for future admin pages.
+
+### Added
+
+- Added [frontend/src/pages/Playground/Playground.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Playground/Playground.tsx), introducing the first dedicated AI testing workspace with provider status visibility, prompt/result tabs, and collapsible operator guidance.
+- Added [docs/frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/frontend-layout-guidelines.md), documenting the repository standard for one-screen admin workspaces and module-local overflow handling.
+- Added [docs/ai-playground-development-plan.md](/home/ray/dev/linkong/planet/docs/ai-playground-development-plan.md), capturing the completed AI gateway/UI work and the next delivery phases for BGP briefs, evidence-first inputs, and future agent runtime expansion.
+
+### Improved
+
+- Improved [frontend/src/components/AppLayout/AppLayout.tsx](/home/ray/dev/linkong/planet/frontend/src/components/AppLayout/AppLayout.tsx) and [frontend/src/App.tsx](/home/ray/dev/linkong/planet/frontend/src/App.tsx) by wiring `AI Playground` into the main admin navigation and route tree instead of pointing operators to a nonexistent `aiprovider` chat page.
+- Improved [planet.sh](/home/ray/dev/linkong/planet/planet.sh) by replacing the incorrect `localhost:8010/chat` closeout link with the frontend `AI Playground` entry.
+- Improved [docker-compose.yml](/home/ray/dev/linkong/planet/docker-compose.yml) by attaching `./aiprovider/.env` to the `aiprovider` service so provider identity, model, and credentials actually reach the running container.
+- Improved [README.md](/home/ray/dev/linkong/planet/README.md) by linking the new frontend layout guidance and AI Playground development plan.
+- Improved [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css) by refining the Playground workspace into a notebook-friendly left-sidebar plus right-tabbed layout, reusing thin scrollbars, and making provider/help/result regions degrade more gracefully under constrained height.
+
+### Fixed
+
+- Fixed the local AI status flow so provider configuration no longer appeared permanently `disabled / not configured` merely because `aiprovider/.env` was not mounted into the container.
+- Fixed repeated `Provider 状态` refetching when switching away from and back to `/playground` by caching the last known provider status within the browser session until the operator explicitly refreshes it.
+- Fixed several Playground layout regressions where auxiliary panels could push the result area out of view or clip provider details without exposing internal scrolling.
+
+## 0.23.4
+
+Released: 2026-04-08
+
+### Highlights
+
+- Fixed the BGP overview workspace so summary cards and tabular data now share viewport space more predictably, keeping the table header visible while preserving internal horizontal and vertical scrolling.
+
+### Improved
+
+- Improved [BGP.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/BGP/BGP.tsx) by switching BGP tables from frontend pagination to in-table scrolling, setting explicit horizontal scroll baselines per tab, and reshaping the summary cards into a desktop two-row layout with a compact single-row horizontal strip on tighter screens.
+- Improved [index.css](/home/ray/dev/linkong/planet/frontend/src/index.css) by hardening the BGP table layout chain from card body to Ant Design table body, so width overflow stays inside the table region and compact summary cards no longer consume unnecessary vertical space above the workspace.
+- Improved release consistency by aligning [VERSION](/home/ray/dev/linkong/planet/VERSION), [pyproject.toml](/home/ray/dev/linkong/planet/pyproject.toml), [frontend/package.json](/home/ray/dev/linkong/planet/frontend/package.json), and [uv.lock](/home/ray/dev/linkong/planet/uv.lock) on the same `0.23.4` bugfix version.
+
+### Fixed
+
+- Fixed the BGP collector/events tables so shrinking the browser window no longer clips the table card frame without exposing a usable horizontal scrollbar.
+- Fixed the BGP table region so the Ant Design header row is no longer obscured by an overgrown table body after the page switched to full-height internal scrolling.
+- Fixed the BGP summary area so medium and large screens no longer collapse the six KPI cards into overly narrow single-row tiles.
+
+## 0.23.3
+
+Released: 2026-04-08
+
+### Highlights
+
+- Refined `planet.sh` startup and restart presentation so service bring-up, health checks, and frontend dev-server readiness are easier to follow in real time.
+
+### Improved
+
+- Improved [planet.sh](/home/ray/dev/linkong/planet/planet.sh) by standardizing the script on `zsh`, tightening frontend startup stage rendering, and making spinner-driven subtask feedback show up step by step instead of bunching at the end.
+- Improved [planet.sh](/home/ray/dev/linkong/planet/planet.sh) by clarifying health-check subtask copy, filtering noisy frontend startup lines, and aligning terminal output spacing across spinner and status rows.
+
+### Fixed
+
+- Fixed several `zsh` compatibility issues in [planet.sh](/home/ray/dev/linkong/planet/planet.sh), including reserved variable name collisions during `restart` and container health checks.
+- Fixed [planet.sh](/home/ray/dev/linkong/planet/planet.sh) so frontend readiness animation no longer appears frozen while waiting for Vite startup and health probes.
+
+## 0.23.2
+
+Released: 2026-04-08
+
+### Highlights
+
+- Hardened datasource retrigger handling so operators can safely force reruns without losing control of task state visibility or rollback behavior.
+
+### Improved
+
+- Improved [frontend/src/pages/DataSources/DataSources.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/DataSources/DataSources.tsx) by mapping phase-local task percentages into a continuous overall progress bar, pre-checking running task status before retriggering, and consolidating the trigger/force-confirm flow.
+- Improved [planet.sh](/home/ray/dev/linkong/planet/planet.sh) by polishing startup CLI output, reducing duplicated spinner cleanup, and standardizing log/help output around the newer terminal presentation.
+
+### Fixed
+
+- Fixed forced datasource reruns in [backend/app/services/collectors/base.py](/home/ray/dev/linkong/planet/backend/app/services/collectors/base.py) by rolling back invalid SQLAlchemy session state before marking cancelled or failed task cleanup, preventing `PendingRollbackError` during operator-triggered cancellation.
+- Fixed datasource trigger UX in [frontend/src/pages/DataSources/DataSources.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/DataSources/DataSources.tsx) so batch and single-source progress no longer jump straight to `100%` on frontend-side status coercion while the backend is still reporting real progress.
+
+## 0.23.1
+
+Released: 2026-04-07
+
+### Highlights
+
+- Fixed the BGP overview page so high-DPI and lower-height screens can keep the observation workspace visible within a single viewport.
+
+### Improved
+
+- Improved [BGP.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/BGP/BGP.tsx) by switching the page to a height-aware shell, compact summary layout, and tabbed data workspace so observation tables retain the majority of the screen.
+- Improved [index.css](/home/ray/dev/linkong/planet/frontend/src/index.css) by adding responsive BGP-specific compact spacing, denser table paddings, and an internal scroll region for collector coverage and event tables.
+
+### Fixed
+
+- Fixed the BGP observation page on small or high-scale displays where stacked stats and full-height blocks consumed too much vertical space, leaving only a couple of visible table rows.
+
 ## 0.23.0
 
 Released: 2026-04-07
@@ -31,6 +441,8 @@ Released: 2026-04-07
 - Improved backend-to-provider tracing by propagating `X-Request-ID` through the AI call chain and returning the same header from both backend and `aiprovider`.
 - Improved resilience by adding lightweight retry handling to both `backend -> aiprovider` and `aiprovider -> model provider` HTTP calls.
 - Improved operator workflow by folding `aiprovider` startup, health checks, restart support, and log viewing into [planet.sh](/home/ray/dev/linkong/planet/planet.sh).
+- Improved [planet.sh](/home/ray/dev/linkong/planet/planet.sh) by adding bounded retry and container-health self-recovery for dependency installs, database startup, `aiprovider` startup, and interactive PostgreSQL boot paths.
+- Improved [README.md](/home/ray/dev/linkong/planet/README.md) by documenting the new `planet.sh` retry and health-check tuning environment variables with concrete override examples.
 - Improved container consistency by switching [backend/Dockerfile](/home/ray/dev/linkong/planet/backend/Dockerfile) and [aiprovider/Dockerfile](/home/ray/dev/linkong/planet/aiprovider/Dockerfile) to `uv sync` / `uv run`.
 
 ### Changed

docs/agent-architecture-plan.md

@@ -0,0 +1,647 @@
+# Agent Architecture Plan
+
+## Overview
+
+This document defines the agent architecture for Planet.
+
+The architecture is intentionally broader than datasource health checking.
+
+It is designed to support both:
+
+- datasource health governance
+- future situational-awareness workflows
+
+The core idea is to avoid building a one-off "repair broken API links" agent.
+
+Instead, Planet should grow a reusable agent runtime that can:
+
+- collect evidence
+- evaluate signals
+- reason over incomplete information
+- generate proposals
+- produce assessments
+- execute limited actions under policy
+
+
+## Design Goal
+
+Build an agent foundation that can evolve in this order:
+
+1. datasource health checks
+2. datasource repair proposals
+3. signal correlation
+4. situational assessments
+5. controlled runtime actions
+
+This means the architecture should treat datasource health as one use case of the larger agent system, not as the whole system.
+
+
+## Core Principles
+
+1. Separate evidence from reasoning
+
+- raw signals should be gathered first
+- deterministic checks should run before LLM reasoning
+
+2. Agents do not own the defaults
+
+- repository defaults remain human-owned
+- agents operate on runtime state, proposals, and overrides
+
+3. Reasoning and action are different responsibilities
+
+- many agents should be read-only or propose-only
+- only tightly controlled flows may apply changes
+
+4. Shared runtime, specialized roles
+
+- multiple agent roles should share the same object model and orchestration patterns
+- health and situational-awareness agents should not invent incompatible payloads
+
+5. Auditability is mandatory
+
+- every proposal, assessment, and applied action should be attributable
+
+
+## System Layers
+
+Planet agent architecture should be split into four layers.
+
+### 1. Signal Layer
+
+Purpose:
+
+- gather raw evidence from internal and external systems
+
+Example sources:
+
+- collector outputs
+- datasource health checks
+- logs
+- snapshots
+- alerts
+- web search results
+- scraped pages
+- external APIs
+- operator inputs
+
+Responsibilities:
+
+- fetch
+- normalize
+- timestamp
+- tag with source and trust level
+
+This layer should not make high-level judgments.
+
+
+### 2. Evaluation Layer
+
+Purpose:
+
+- perform deterministic analysis
+
+Examples:
+
+- reachability checks
+- schema validation
+- threshold checks
+- time-window comparisons
+- anomaly counters
+- completeness checks
+
+Responsibilities:
+
+- classify signals into machine-readable findings
+- attach deterministic evidence
+
+This layer should avoid LLM dependency whenever possible.
+
+
+### 3. Reasoning Layer
+
+Purpose:
+
+- use LLMs when semantic interpretation or incomplete-information reasoning is needed
+
+Examples:
+
+- endpoint migration inference
+- multi-source event correlation
+- causality hypotheses
+- ambiguity reduction
+- assessment narrative generation
+- action recommendation generation
+
+Responsibilities:
+
+- synthesize evidence
+- produce hypotheses
+- rank confidence
+- explain reasoning boundaries
+
+This is the main place where `aiprovider` and web search are used.
+
+
+### 4. Action Layer
+
+Purpose:
+
+- convert proposals or assessments into controlled system actions
+
+Examples:
+
+- create runtime override
+- create proposal
+- publish alert
+- update operator task queue
+- generate summary artifact
+- trigger follow-up verification
+
+Responsibilities:
+
+- enforce policy
+- enforce approval requirements
+- verify post-action outcomes
+- record audit trails
+
+
+## Architecture Sketch
+
+```mermaid
+flowchart TD
+    A["Collectors / Logs / Snapshots / External APIs"] --> B["Signal Layer"]
+    W["Web Search / Page Fetch / Docs"] --> B
+    B --> C["Evaluation Layer"]
+    C --> D["Findings"]
+    D --> E["Reasoning Layer (LLM + Tools)"]
+    E --> F["Proposals"]
+    E --> G["Assessments"]
+    F --> H["Action Layer"]
+    H --> I["Runtime Overrides / Alerts / Tasks"]
+    H --> J["Verification Loop"]
+    J --> B
+
+    K["Policy Engine"] --> H
+    L["Audit / History Store"] --> H
+    L --> E
+    L --> C
+```
+
+
+## Agent Roles
+
+The first version should define these logical roles.
+
+### 1. Health Agent
+
+Primary use case:
+
+- datasource health governance
+
+Inputs:
+
+- datasource metadata
+- current endpoint
+- latest health records
+- latest failures
+- deterministic findings
+
+Outputs:
+
+- health interpretation
+- repair proposal
+- confidence
+- evidence references
+
+Typical action level:
+
+- propose-only
+
+
+### 2. Correlation Agent
+
+Primary use case:
+
+- identify whether multiple signals describe the same event or related events
+
+Inputs:
+
+- findings from multiple collectors
+- time windows
+- region / ASN / prefix / cable relationships
+- prior incidents
+
+Outputs:
+
+- grouped event candidates
+- correlation rationale
+- confidence per relationship
+
+Typical action level:
+
+- read-only
+
+
+### 3. Assessment Agent
+
+Primary use case:
+
+- produce situational-awareness outputs
+
+Inputs:
+
+- grouped events
+- findings
+- current context
+- historical context
+- operator constraints
+
+Outputs:
+
+- structured assessment
+- risk summary
+- evidence-backed recommendations
+- missing-information list
+
+Typical action level:
+
+- read-only or propose-only
+
+
+### 4. Recovery Agent
+
+Primary use case:
+
+- carry low-risk proposals into controlled runtime actions
+
+Inputs:
+
+- approved proposal
+- policy constraints
+- trusted-domain rules
+- verification checks
+
+Outputs:
+
+- applied override
+- failed application
+- rollback request
+
+Typical action level:
+
+- apply-limited
+
+
+## Shared Object Model
+
+All agents should work on a shared object model.
+
+That prevents the health subsystem and situational-awareness subsystem from drifting into incompatible payloads.
+
+### Signal
+
+Represents a raw observed fact.
+
+Examples:
+
+- a datasource returned HTTP 404
+- a collector returned empty results
+- BGP updates spiked in one region
+- a known endpoint now redirects elsewhere
+
+Suggested shape:
+
+```json
+{
+  "id": "sig_123",
+  "type": "datasource.http_failure",
+  "source": "ris_live_bgp",
+  "occurred_at": "2026-04-08T10:00:00Z",
+  "severity": "medium",
+  "payload": {},
+  "trust": 0.95
+}
+```
+
+
+### Finding
+
+Represents a deterministic or semi-deterministic interpretation of one or more signals.
+
+Examples:
+
+- `schema_changed`
+- `endpoint_unreachable`
+- `data_volume_abnormally_low`
+- `event_cluster_detected`
+
+Suggested shape:
+
+```json
+{
+  "id": "find_123",
+  "type": "datasource.schema_changed",
+  "source_ids": ["sig_123"],
+  "confidence": 0.92,
+  "evidence": [],
+  "details": {}
+}
+```
+
+
+### Proposal
+
+Represents a recommended action, not an already-applied action.
+
+Examples:
+
+- switch endpoint to new URL
+- disable bad override
+- escalate issue for manual review
+
+Suggested shape:
+
+```json
+{
+  "id": "prop_123",
+  "kind": "endpoint_override",
+  "target": "telegeography_cables",
+  "confidence": 0.84,
+  "reason": "Official docs now point to a new API path",
+  "payload": {},
+  "evidence_urls": [],
+  "status": "proposed"
+}
+```
+
+
+### Assessment
+
+Represents a structured situational-awareness output for operators or downstream systems.
+
+Examples:
+
+- current network posture summary
+- incident impact assessment
+- risk and response recommendations
+
+Suggested shape:
+
+```json
+{
+  "id": "assess_123",
+  "scope": "regional-network",
+  "risk_level": "high",
+  "summary": "Regional routing instability is increasing.",
+  "key_risks": [],
+  "evidence": [],
+  "recommendations": [],
+  "missing_data": []
+}
+```
+
+
+## State Machine
+
+The shared orchestration flow should look like this:
+
+```mermaid
+stateDiagram-v2
+    [*] --> Collect
+    Collect --> Validate
+    Validate --> Classify
+    Classify --> Reason
+    Reason --> Propose
+    Reason --> Assess
+    Propose --> Review
+    Review --> Apply
+    Apply --> Verify
+    Verify --> Archive
+    Assess --> Archive
+    Archive --> [*]
+```
+
+Definitions:
+
+- `Collect`: gather signals
+- `Validate`: run deterministic checks
+- `Classify`: create findings
+- `Reason`: invoke LLM reasoning when needed
+- `Propose`: create change proposals
+- `Review`: policy or human approval
+- `Apply`: perform limited runtime action
+- `Verify`: confirm action effect
+- `Archive`: store artifacts and decisions
+
+
+## Permission Model
+
+Each agent role should be assigned one of these action levels.
+
+### `read-only`
+
+Allowed:
+
+- read signals
+- search web
+- fetch pages
+- read internal state
+- generate findings and assessments
+
+Not allowed:
+
+- mutate config
+- write overrides
+- change live runtime behavior
+
+
+### `propose-only`
+
+Allowed:
+
+- everything in `read-only`
+- create proposals
+- create review tasks
+
+Not allowed:
+
+- apply live changes
+
+
+### `apply-limited`
+
+Allowed:
+
+- everything in `propose-only`
+- write approved runtime overrides
+- trigger verification checks
+
+Not allowed:
+
+- mutate repository defaults
+- make destructive data changes
+- bypass policy engine
+
+
+## Runtime Components
+
+The first durable architecture should introduce these components.
+
+### 1. Signal Store
+
+Stores normalized evidence and health outputs.
+
+
+### 2. Finding Store
+
+Stores deterministic classifications that can be reused by multiple agents.
+
+
+### 3. Proposal Store
+
+Stores recommended actions with evidence and confidence.
+
+
+### 4. Assessment Store
+
+Stores structured situational-awareness outputs.
+
+
+### 5. Policy Engine
+
+Decides:
+
+- whether agent may run
+- whether proposal requires review
+- whether proposal may auto-apply
+- whether post-apply verification passed
+
+
+### 6. Override Store
+
+Stores runtime-only configuration changes.
+
+This is where endpoint repairs should live.
+
+
+## Relation To `aiprovider`
+
+`aiprovider` should remain the model gateway.
+
+It should not become the full agent runtime.
+
+Recommended split:
+
+- `aiprovider`
+  - provider adaptation
+  - prompt transport
+  - model execution
+  - protocol compatibility
+
+- agent runtime
+  - orchestration
+  - signal handling
+  - tool selection
+  - proposal generation
+  - policy and audit
+
+This keeps provider concerns and agent behavior concerns separate.
+
+
+## Relation To Datasource Health
+
+Datasource health becomes one vertical slice of this architecture.
+
+Mapping:
+
+- signal:
+  - endpoint unreachable
+  - schema mismatch
+  - bad content type
+- finding:
+  - `failed`
+  - `schema_changed`
+  - `moved_endpoint_suspected`
+- proposal:
+  - runtime override suggestion
+- assessment:
+  - datasource health summary for operators
+
+
+## Relation To Situational Awareness
+
+Future situational-awareness capabilities should reuse the same flow:
+
+- raw telemetry becomes signals
+- anomaly detection becomes findings
+- LLM correlation becomes reasoning
+- operator-facing output becomes assessments
+- policy-approved mitigations become actions
+
+This lets the platform evolve from operational health governance into broader cyber/network posture workflows without changing the architecture.
+
+
+## Suggested Delivery Sequence
+
+### Phase A
+
+- finalize shared object model
+- implement health-oriented signal and finding storage
+
+### Phase B
+
+- implement Health Agent
+- generate proposals only
+
+### Phase C
+
+- implement Assessment Agent
+- expose structured assessments via API
+
+### Phase D
+
+- implement Correlation Agent
+- support multi-source incident grouping
+
+### Phase E
+
+- implement Recovery Agent with policy-gated runtime actions
+
+
+## Recommended First Build
+
+The first build should not try to implement every agent role.
+
+Recommended initial slice:
+
+- shared object model
+- health signals
+- health findings
+- Health Agent
+- proposal generation only
+
+This gives immediate value while preserving the longer-term architecture.
+
+
+## Non-Goals For The First Iteration
+
+- repository YAML auto-rewrites
+- unrestricted autonomous action
+- full incident graph reasoning
+- automatic large-scale remediation
+- agent-owned configuration source of truth
+
+
+## Summary
+
+Planet should treat agents as a reusable runtime for evidence, reasoning, proposals, and assessments.
+
+The datasource health use case is the first practical entrypoint, but the architecture should already assume future situational-awareness expansion.
+
+The safest path is:
+
+- deterministic checks first
+- agent reasoning second
+- proposals before actions
+- runtime overrides instead of default mutation

docs/agent-runtime-roadmap.md

@@ -0,0 +1,346 @@
+# Agent Runtime Roadmap
+
+## Overview
+
+This document connects three existing planning threads into one implementation roadmap:
+
+- `aiprovider` as the model gateway
+- datasource health governance as the first practical agent use case
+- situational awareness as the broader long-term target
+
+Related documents:
+
+- [aiprovider](/home/ray/dev/linkong/planet/docs/aiprovider.md)
+- [datasource-health-plan](/home/ray/dev/linkong/planet/docs/datasource-health-plan.md)
+- [agent-architecture-plan](/home/ray/dev/linkong/planet/docs/agent-architecture-plan.md)
+
+
+## Big Picture
+
+Planet should evolve in layers:
+
+1. stable model gateway
+2. deterministic health and evidence collection
+3. agent runtime for reasoning and proposal generation
+4. situational-awareness assessments and controlled actions
+
+This prevents the system from collapsing into a single giant "AI feature" with unclear boundaries.
+
+
+## Architecture Overview
+
+```mermaid
+flowchart TD
+    U["Frontend / Backend APIs / Operators"] --> B["Planet Backend"]
+    B --> H["Datasource Health Services"]
+    B --> R["Agent Runtime"]
+    R --> P["aiprovider"]
+    P --> M["OpenAI / Anthropic / MiniMax / Ollama / Local Models"]
+
+    C["Collectors / Snapshots / Logs / Alerts / BGP Signals"] --> S["Signal Store"]
+    H --> S
+    S --> E["Evaluation Layer"]
+    E --> F["Findings"]
+    F --> R
+
+    W["Web Search / Page Fetch / Docs Fetch"] --> R
+    R --> PR["Proposals"]
+    R --> AS["Assessments"]
+
+    PR --> O["Runtime Overrides / Review Queue / Tasks"]
+    AS --> SA["Situational Awareness APIs / UI"]
+
+    O --> V["Verification Loop"]
+    V --> S
+```
+
+
+## Role Boundaries
+
+### `aiprovider`
+
+Responsibilities:
+
+- provider compatibility
+- protocol adaptation
+- auth and model transport
+- request/response normalization
+
+Not responsible for:
+
+- agent orchestration
+- business workflows
+- datasource repair policy
+- situational-awareness domain logic
+
+
+### Backend
+
+Responsibilities:
+
+- stable business APIs
+- auth and permissions
+- task orchestration
+- health records
+- proposal and override persistence
+- assessment exposure
+
+
+### Agent Runtime
+
+Responsibilities:
+
+- consume findings and context
+- invoke LLMs via `aiprovider`
+- invoke tools such as web search
+- create proposals
+- create assessments
+- route to policy-controlled action paths
+
+
+## Delivery Sequence
+
+## Stage 1: Gateway Foundation
+
+Status:
+
+- already in place
+
+Delivered by current work:
+
+- `aiprovider`
+- multi-provider compatibility
+- backend AI facade
+- MiniMax / Anthropic-compatible support
+- request-id propagation
+
+Primary outcome:
+
+- the system already has a stable way to call models
+
+
+## Stage 2: Datasource Health MVP
+
+Goal:
+
+- establish deterministic health observability
+
+Key work:
+
+- health check task runner
+- health result table
+- datasource health APIs
+- UI visibility
+- collector endpoint override precedence cleanup
+
+Primary outcome:
+
+- Planet knows which collectors are healthy before asking an LLM anything
+
+
+## Stage 3: Health Agent
+
+Goal:
+
+- let the first agent role operate on health failures
+
+Key work:
+
+- convert health failures into signals/findings
+- invoke agent only for failed or suspicious cases
+- produce repair proposals with evidence and confidence
+
+Primary outcome:
+
+- Planet can suggest endpoint repairs without mutating defaults
+
+
+## Stage 4: Runtime Repair Application
+
+Goal:
+
+- safely apply approved datasource repair proposals
+
+Key work:
+
+- override storage
+- policy-gated apply flow
+- verification after apply
+- rollback path
+
+Primary outcome:
+
+- datasource repair becomes operationally useful without polluting repository defaults
+
+
+## Stage 5: Situational Awareness Assessments
+
+Goal:
+
+- reuse the same runtime for broader operator-facing assessment
+
+Key work:
+
+- normalize telemetry and incident evidence into signals/findings
+- build Assessment Agent
+- expose structured assessments through backend APIs and UI
+
+Primary outcome:
+
+- LLM output becomes evidence-backed situational summary, not just ad hoc chat output
+
+
+## Stage 6: Correlation and Controlled Actions
+
+Goal:
+
+- connect multiple sources into higher-level posture and event groupings
+
+Key work:
+
+- event correlation
+- incident grouping
+- recommendation scoring
+- controlled action routing
+
+Primary outcome:
+
+- Planet becomes a true agent-assisted situational-awareness system
+
+
+## Implementation Tracks
+
+These tracks can progress in parallel, but they should stay loosely coupled.
+
+### Track A: Config and Runtime Resolution
+
+Scope:
+
+- datasource defaults
+- overrides
+- runtime precedence
+- audit trails
+
+First milestone:
+
+- health-safe override layer
+
+
+### Track B: Health and Evidence
+
+Scope:
+
+- deterministic checks
+- failure categorization
+- signal and finding persistence
+
+First milestone:
+
+- datasource health record system
+
+
+### Track C: Agent Runtime
+
+Scope:
+
+- shared object model
+- orchestration flow
+- prompt/tool pipeline
+- policy integration
+
+First milestone:
+
+- Health Agent proposal pipeline
+
+
+### Track D: Situational Awareness
+
+Scope:
+
+- assessment schema
+- multi-source context assembly
+- operator-facing outputs
+
+First milestone:
+
+- structured assessment API
+
+
+## Shared Artifacts
+
+To avoid fragmentation, these artifacts should be shared across all future agent work.
+
+### Shared object model
+
+- `Signal`
+- `Finding`
+- `Proposal`
+- `Assessment`
+
+### Shared orchestration flow
+
+- collect
+- validate
+- classify
+- reason
+- propose or assess
+- review or apply
+- verify
+- archive
+
+### Shared policy model
+
+- read-only
+- propose-only
+- apply-limited
+
+
+## Recommended Next Concrete Steps
+
+1. Build Stage 2 first
+
+- datasource health records
+- deterministic checks
+- no automatic repair
+
+2. Then build Stage 3
+
+- Health Agent
+- proposal generation only
+
+3. Then Stage 4
+
+- override apply flow
+- rollback and verification
+
+4. Only after that start Stage 5
+
+- broader situational-awareness assessment workflows
+
+
+## Why This Order
+
+Because situational-awareness quality depends on reliable upstream data.
+
+If datasource health is weak:
+
+- agent reasoning quality will degrade
+- false explanations will increase
+- assessment trust will drop
+
+So datasource health is not a side task.
+
+It is the first operational foundation for the later situational-awareness system.
+
+
+## Summary
+
+Planet should be built as:
+
+- `aiprovider` for model access
+- backend services for orchestration and persistence
+- datasource health as the first evidence-governance layer
+- agent runtime as the reusable reasoning core
+- situational awareness as the long-term application layer
+
+That path keeps the architecture coherent and lets each phase produce useful functionality without forcing a rewrite later.

docs/ai-playground-development-plan.md

@@ -0,0 +1,361 @@
+# AI Playground Development Plan
+
+## 目标
+
+这份计划用于统一 `aiprovider`、`backend AI facade`、`Playground` 页面，以及后续 `BGP / 告警 / 数据源健康` 等 AI 入口的演进方向。
+
+当前原则：
+
+- `aiprovider` 继续作为独立模型网关
+- `backend` 继续作为稳定业务入口
+- `frontend` 负责测试台和业务 UI
+- 先做“可控、可验证、可解释”的 AI 能力，再逐步引入 agent/tool calling
+
+## 当前已完成
+
+### 1. AI 网关基础层
+
+已完成：
+
+- 独立 `aiprovider` 服务
+- `backend -> aiprovider -> model provider` 调用链
+- `provider/status` 与 `situational-awareness/analyze` 稳定接口
+- `X-Request-ID` 透传
+- 轻量超时与重试
+- MiniMax / Anthropic-compatible / OpenAI-compatible / Ollama 适配
+
+相关文件：
+
+- [backend/app/api/v1/ai.py](/home/ray/dev/linkong/planet/backend/app/api/v1/ai.py)
+- [backend/app/services/ai_client.py](/home/ray/dev/linkong/planet/backend/app/services/ai_client.py)
+- [aiprovider/main.py](/home/ray/dev/linkong/planet/aiprovider/main.py)
+- [aiprovider/provider_service.py](/home/ray/dev/linkong/planet/aiprovider/provider_service.py)
+- [docs/aiprovider.md](/home/ray/dev/linkong/planet/docs/aiprovider.md)
+
+### 2. 本地运行与配置打通
+
+已完成：
+
+- `planet.sh` 启动链路纳入 `aiprovider`
+- `planet.sh` 启动完成后输出 Playground 入口
+- `docker-compose.yml` 为 `aiprovider` 加入 `env_file`
+- `backend/.env` 与 `aiprovider/.env` 两侧 service token 对齐
+- `Playground` 状态缓存，避免页面切换时每次都重新请求 provider 状态
+
+相关文件：
+
+- [planet.sh](/home/ray/dev/linkong/planet/planet.sh)
+- [docker-compose.yml](/home/ray/dev/linkong/planet/docker-compose.yml)
+- [backend/.env.example](/home/ray/dev/linkong/planet/backend/.env.example)
+- [aiprovider/.env.example](/home/ray/dev/linkong/planet/aiprovider/.env.example)
+
+### 3. Playground UI 基础版
+
+已完成：
+
+- 新增前端路由 `/playground`
+- 左侧 `Provider 状态 + 测试说明`
+- 右侧 `请求 / 结果` Tabs
+- `Provider 状态` 支持手动刷新
+- `测试说明` 支持折叠
+- 内部区域采用细滚动条
+- 页面布局开始遵循“单屏工作区 + 模块内部滚动”规范
+
+相关文件：
+
+- [frontend/src/pages/Playground/Playground.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Playground/Playground.tsx)
+- [frontend/src/App.tsx](/home/ray/dev/linkong/planet/frontend/src/App.tsx)
+- [frontend/src/components/AppLayout/AppLayout.tsx](/home/ray/dev/linkong/planet/frontend/src/components/AppLayout/AppLayout.tsx)
+- [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css)
+
+### 4. 前端布局规范沉淀
+
+已完成：
+
+- 把“一屏工作区、主模块优先、模块内部滚动”的规范文档化
+- 明确 `BGP` 页面为当前参考实现
+
+相关文件：
+
+- [docs/frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/frontend-layout-guidelines.md)
+- [frontend/src/pages/BGP/BGP.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/BGP/BGP.tsx)
+
+## 当前限制
+
+### 1. Playground 还是 prompt playground，不是 agent playground
+
+当前 `Playground` 的 `观察项 / 目标 / 约束条件` 都是人工输入。
+
+模型现在拿到的是：
+
+- 你手工输入的结构化字段
+- 后端传递的少量静态上下文
+
+模型现在拿不到：
+
+- 实时 BGP 事件
+- 真实告警列表
+- 数据源健康状态
+- 自动检索结果
+- tool calling / skills / 自主取数
+
+### 2. `situational-awareness/analyze` 还是通用提示词接口
+
+当前更适合：
+
+- 测试链路
+- 测试模型输出风格
+- 验证不同 provider 是否正常返回
+
+当前还不适合：
+
+- 直接当真实态势系统主入口
+- 让用户手工维护长期分析模板
+- 代替专用业务研判接口
+
+### 3. 还没有可验证的真实业务输入注入
+
+目前最缺的是：
+
+- 从业务系统自动整理“事实输入”
+- 再把这些事实喂给 AI
+
+而不是继续让用户在 Playground 手工输入真实事件摘要。
+
+## 短期计划
+
+### Phase A: Playground 收敛为稳定测试台
+
+目标：
+
+- 保持 Playground 简洁可用
+- 不再继续堆“高级参数”
+
+工作项：
+
+- 继续微调左侧 `Provider 状态` 与 `测试说明` 的空间策略
+- 保持 `请求 / 结果` 为单一主工作区
+- 不引入盲填式高级字段
+- 统一滚动条、卡片、溢出行为
+
+完成标准：
+
+- 笔记本视口下依然可用
+- 各模块标题可见
+- 主要阅读区始终是右侧 Tabs
+
+### Phase B: BGP AI 简报
+
+目标：
+
+- 不再依赖手工填写“观察项”
+- 让系统自动把真实 BGP 数据注入 AI
+- 让 BGP 页面逐步从“摘要汇总”升级为“证据驱动的区域态势分析”
+
+建议实现：
+
+- 新增专用后端接口，例如：
+  - `POST /api/v1/ai/bgp/brief`
+- 后端自动读取：
+  - incidents summary
+  - anomalies
+  - recent events
+  - collector coverage summary
+- 后端将结构化事实注入 `context / observations`
+- 前端在 BGP 页面增加“生成 AI 简报”
+
+当前阶段说明：
+
+- 第一版 `BGP AI 简报` 允许先落地为“值班摘要生成器”
+- 也就是先把 incidents / anomalies / events / collector coverage 自动注入
+- 允许模型先做事实摘要、风险归纳、建议动作
+
+但这不应被视为 Phase B 的最终形态。
+
+Phase B 后续还需要补齐：
+
+- prefix geography 证据注入
+  - `iptoasn`
+  - `opengeofeed`
+  - `nro_delegated`
+- 基于 `affected_regions` 与 prefix geography 的区域聚合
+- 区分“真实区域热度”与“collector coverage 偏差”
+- 对高风险 prefix / ASN 给出更明确的国家、城市、运营商归属线索
+- 让 AI 输出明确回答：
+  - 哪些区域正在异常升温
+  - 哪些结论只是观测站偏差
+  - 当前还缺哪些区域证据
+
+完成标准：
+
+- 用户不需要手工录入 BGP 观察项
+- AI 输出能明确区分“事实”和“研判”
+- AI 不只是复述总量和最近几条事件，还能利用 prefix geography 与 affected regions 做区域态势判断
+- 输出中能明确指出：
+  - 高风险区域
+  - 区域证据来源
+  - collector coverage 偏差对判断的影响
+
+### Phase C: 告警 / 数据源健康 AI 简报
+
+目标：
+
+- 复用同样模式，扩展到其他模块
+
+建议入口：
+
+- `Alerts` 页面：异常与告警摘要
+- `DataSources` 页面：采集失败与健康状态总结
+
+原则：
+
+- 每个业务页优先做“专用 AI 简报”
+- 不优先做“万能大聊天框”
+
+## 中期计划
+
+### 1. Assessment Layer
+
+目标：
+
+- 不只返回自由文本
+- 返回结构化的 assessment
+
+建议输出字段：
+
+- summary
+- key_risks
+- evidence
+- recommendations
+- confidence
+- missing_data
+
+这样后续才能：
+
+- 持久化
+- 回看
+- 对比不同时间的 AI 结论
+- 在 Earth / Dashboard / BGP 页面稳定展示
+
+### 2. Evidence-first Runtime
+
+目标：
+
+- 所有 AI 分析先取真实数据，再调模型
+
+原则：
+
+- 先 evidence
+- 再 prompt
+- 最后才是自由生成
+
+优先要做的不是更强聊天，而是：
+
+- 更稳定的数据注入
+- 更一致的事实模板
+- 更清晰的结果结构
+
+### 3. 按页面提供专用入口
+
+目标：
+
+- 让 AI 成为业务视图的一部分，而不是孤立 playground
+
+优先顺序建议：
+
+1. `BGP` AI 简报
+2. `Alerts` AI 简报
+3. `DataSources` 健康研判
+4. `Dashboard` 总览总结
+
+## 长期计划
+
+### 1. Tool Calling / Agent Runtime
+
+只有在以下基础稳定后再推进：
+
+- 数据源健康信号稳定
+- BGP / Alerts / Datasource evidence 注入稳定
+- assessment 结构稳定
+
+长期可做能力：
+
+- AI 调用受控工具查询业务数据
+- AI 调用检索/web search 做外部验证
+- AI 生成建议而不是直接修改系统
+- 审核后触发受控动作
+
+### 2. 受控动作与闭环
+
+潜在方向：
+
+- 根据健康异常生成修复建议
+- 根据态势变化生成处理建议
+- 进入 review queue
+- 审批后执行
+- 验证结果并形成闭环
+
+### 3. 多模块统一 AI 体验
+
+长期目标不是一个孤立 Playground，而是：
+
+- 每个业务页都有自己的 AI 入口
+- 共享统一的 backend AI facade
+- 共享统一的 assessment 结构
+- 共享统一的 evidence 注入与审计链路
+
+## 设计决策总结
+
+### 为什么保留 `aiprovider`
+
+因为它已经很好地承担了：
+
+- provider 适配
+- 协议兼容
+- service token 边界
+- 独立重启与部署
+
+因此短期内不建议把它并回 `backend`。
+
+### 为什么 Playground 不做成万能聊天页
+
+因为当前更需要的是：
+
+- 稳定测试链路
+- 可验证业务输入
+- 专用分析入口
+
+而不是一个泛化但没有真实数据支撑的聊天框。
+
+### 为什么优先做专用 AI 简报
+
+因为：
+
+- 数据可以自动注入
+- 用户心智更清晰
+- 输出更容易结构化
+- 更容易校验事实与研判是否一致
+
+## 下一步建议
+
+按优先级建议接下来这样做：
+
+1. 稳住 `Playground` 当前布局，不再大幅重做
+2. 在 `BGP` 页面新增专用 “AI 简报” 入口
+3. 后端新增 `BGP brief` 专用接口，自动注入真实数据
+4. 补齐 `BGP brief` 的区域态势证据层
+5. 把 AI 输出逐步从自由文本升级为结构化 assessment
+
+### BGP Brief 后续子项
+
+为避免把“已有 AI 简报”误判成“区域分析已完成”，这里单独记录 `BGP brief` 的后续 backlog：
+
+1. 把高风险 prefix 命中的 `iptoasn / opengeofeed / nro_delegated` 结果注入 brief context
+2. 按国家/城市聚合 active incidents、anomalies、affected prefixes，生成区域热点事实层
+3. 把 collector coverage 与区域热点并排注入，避免模型把观测偏差误判成区域风险
+4. 对高风险 ASN / prefix 追加归属线索，如国家、城市、可能运营商或注册区域
+5. 在输出结构中单独增加：
+   - 区域态势
+   - 证据来源
+   - 观测偏差说明
+   - 缺失区域证据

docs/aiprovider.md

@@ -31,22 +31,40 @@ The recommended default is:
 - timeout and lightweight retry
 - request tracing via `X-Request-ID`
 
+This now follows an OpenClaw-like seam:
+
+- `AI_PROVIDER` identifies the vendor or logical provider
+- `AI_PROVIDER_API` identifies the wire adapter
+
+That split makes MiniMax, Claude-compatible gateways, and self-hosted OpenAI-compatible services easier to model without overloading one config field.
+
 ## Supported Providers
 
-`aiprovider` currently supports:
+`aiprovider` currently supports these provider identities:
 
 - `openai`
-- `openai_compatible`
 - `anthropic`
+- `minimax`
+- `ollama`
+
+Supported request adapters:
+
+- `openai-completions`
+- `anthropic-messages`
+- `ollama-generate`
+
+Backward-compatible aliases still accepted:
+
+- `openai_compatible`
 - `anthropic_compatible`
 - `claude_compatible`
-- `ollama`
 
 Provider mapping:
 
-- `vLLM`, `LM Studio`, `One API`: `openai_compatible`
-- `MiniMax`, Claude-compatible gateways: `claude_compatible`
-- `Ollama`: `ollama`
+- `vLLM`, `LM Studio`, `One API`: `AI_PROVIDER=openai`, `AI_PROVIDER_API=openai-completions`
+- `MiniMax`: `AI_PROVIDER=minimax`, `AI_PROVIDER_API=anthropic-messages`
+- Claude-compatible gateways: `AI_PROVIDER=anthropic`, `AI_PROVIDER_API=anthropic-messages`
+- `Ollama`: `AI_PROVIDER=ollama`, `AI_PROVIDER_API=ollama-generate`
 
 ## API Surfaces
 
@@ -137,9 +155,13 @@ Both backend and `aiprovider` return the same payload shape:
 
 ```json
 {
-  "provider": "openai_compatible",
-  "model": "gpt-4o-mini",
+  "provider": "minimax",
+  "api": "anthropic-messages",
+  "model": "MiniMax-M2.7",
   "content": "1) 态势摘要 ...",
+  "content_blocks": [],
+  "text_blocks": [],
+  "thinking_blocks": [],
   "raw_response": {}
 }
 ```
@@ -189,17 +211,37 @@ AI_ANALYSIS_SYSTEM_PROMPT=你是态势感知分析助手。请基于输入的上
 ### OpenAI-compatible example
 
 ```env
-AI_PROVIDER=openai_compatible
+AI_PROVIDER=openai
+AI_PROVIDER_API=openai-completions
 AI_BASE_URL=http://127.0.0.1:8001/v1
 AI_API_KEY=local-key
 AI_MODEL=your-local-model
 ```
 
-### Claude-compatible example
+### MiniMax CN example
 
 ```env
-AI_PROVIDER=claude_compatible
-AI_BASE_URL=https://your-claude-compatible-endpoint.example.com
+AI_PROVIDER=minimax
+AI_PROVIDER_API=anthropic-messages
+AI_BASE_URL=https://api.minimaxi.com/anthropic
+AI_API_KEY=sk-cp-xxxxx
+AI_MODEL=MiniMax-M2.7
+AI_MAX_TOKENS=1200
+AI_ANTHROPIC_VERSION=2023-06-01
+```
+
+MiniMax note:
+
+- This follows the same Anthropic Messages request shape as the official MiniMax examples.
+- For MiniMax, `aiprovider` now disables `thinking` by default unless the caller explicitly passes a `thinking` object.
+- This mirrors OpenClaw's caution around MiniMax Anthropic-compatible behavior.
+
+### Anthropic-compatible example
+
+```env
+AI_PROVIDER=anthropic
+AI_PROVIDER_API=anthropic-messages
+AI_BASE_URL=https://your-claude-compatible-endpoint.example.com/anthropic
 AI_API_KEY=your_api_key
 AI_MODEL=your-model
 AI_MAX_TOKENS=1200
@@ -210,6 +252,7 @@ AI_ANTHROPIC_VERSION=2023-06-01
 
 ```env
 AI_PROVIDER=ollama
+AI_PROVIDER_API=ollama-generate
 AI_BASE_URL=http://127.0.0.1:11434
 AI_API_KEY=
 AI_MODEL=qwen2.5:7b

docs/datasource-health-plan.md

@@ -0,0 +1,486 @@
+# Datasource Health Plan
+
+## Overview
+
+This document defines a phased plan for datasource health governance.
+
+The goal is to make collectors observable, diagnosable, and recoverable when upstream APIs change, while avoiding unsafe automatic mutation of repository defaults.
+
+The key principle is:
+
+- do not let runtime automation rewrite repository default config
+
+Instead, split responsibilities across:
+
+- default config
+- runtime overrides
+- health check records
+- agent-generated repair proposals
+
+
+## Problem Statement
+
+Collectors currently depend on third-party APIs, data downloads, mirrored JSON files, archive links, and web pages.
+
+These upstream dependencies can fail in several ways:
+
+- endpoint becomes unreachable
+- endpoint still responds but schema changes
+- content-type changes
+- website shuts down or moves
+- mirror link disappears
+- HTML structure changes and scraping fails
+- endpoint requires a new path or new host
+
+We want a system that can:
+
+- detect datasource health degradation early
+- identify likely cause
+- search for updated endpoints when reasonable
+- apply safe runtime fixes without polluting default repo config
+- preserve auditability and rollback
+
+
+## Design Principles
+
+1. Default config is stable
+
+- `backend/app/core/data_sources.yaml` remains the repository baseline.
+- It should be changed intentionally through normal development flow, not by autonomous runtime agents.
+
+2. Runtime fixes are isolated
+
+- Emergency or adaptive fixes should live in a runtime override layer.
+- Overrides should be reversible and auditable.
+
+3. Deterministic checks come first
+
+- Use normal programmatic health checks before using LLMs.
+- Only call an agent when deterministic checks indicate a meaningful failure.
+
+4. Agents suggest before they mutate
+
+- Agents should produce proposals with evidence and confidence.
+- Application of a proposal should be controlled by policy.
+
+5. Every repair is attributable
+
+- Store what changed, why, who or what suggested it, and when it was applied.
+
+
+## Configuration Layers
+
+Recommended runtime precedence:
+
+1. datasource endpoint override
+2. datasource DB endpoint override
+3. repository default YAML
+4. collector internal fallback logic
+
+Definitions:
+
+- repository default YAML:
+  - `backend/app/core/data_sources.yaml`
+  - versioned baseline
+- datasource DB endpoint override:
+  - existing `DataSourceConfig.endpoint`
+  - current runtime override entrypoint
+- datasource endpoint override:
+  - a dedicated new override table
+  - used for health-repair and proposal application
+- collector internal fallback logic:
+  - final defensive fallback
+  - should be minimized over time
+
+
+## Recommended Architecture
+
+### 1. Deterministic Health Checks
+
+Each collector gets a health profile with checks such as:
+
+- endpoint resolves
+- HTTP request succeeds
+- status code is acceptable
+- content-type is expected
+- body parses successfully
+- minimum structural fields exist
+- sample item count is plausible
+- latency is within threshold
+
+Output states:
+
+- `healthy`
+- `degraded`
+- `failed`
+- `schema_changed`
+- `rate_limited`
+- `auth_required`
+
+
+### 2. Agent-Assisted Repair Discovery
+
+Only triggered when deterministic health checks fail or return suspicious structure.
+
+Agent responsibilities:
+
+- search for current official endpoint or replacement path
+- inspect likely upstream documentation or landing pages
+- compare candidate endpoint output to collector expectations
+- produce a repair proposal with confidence and evidence
+
+Agent should not directly modify repository defaults.
+
+
+### 3. Safe Runtime Repair Application
+
+Repair proposals can be:
+
+- reviewed manually
+- auto-applied only under strict low-risk policy
+
+Auto-apply should be limited to cases like:
+
+- same trusted domain
+- highly similar response structure
+- repeated successful verification
+- confidence above threshold
+
+
+## Phased Delivery Plan
+
+## Phase 1: Deterministic Health MVP
+
+Goal:
+
+- build health observability without automated repair
+
+Scope:
+
+- datasource health check task runner
+- datasource health result persistence
+- endpoint reachability + parse checks
+- dashboard or API visibility into health status
+
+Deliverables:
+
+- health check service
+- health check record table
+- status endpoint
+- scheduled or manual check trigger
+
+No agent usage yet.
+
+
+## Phase 2: Agent Repair Proposals
+
+Goal:
+
+- let agent investigate failing sources and propose updated endpoints
+
+Scope:
+
+- invoke agent only when datasource health is `failed` or `schema_changed`
+- web search + page inspection
+- candidate endpoint extraction
+- proposal persistence
+
+Deliverables:
+
+- repair proposal schema
+- proposal generation pipeline
+- confidence and evidence model
+- operator review view or API
+
+Still no automatic config mutation.
+
+
+## Phase 3: Runtime Overrides
+
+Goal:
+
+- allow approved proposals to take effect safely at runtime
+
+Scope:
+
+- add dedicated override storage
+- runtime resolution prefers override over default config
+- proposal application writes override only
+
+Deliverables:
+
+- endpoint override table
+- override-aware resolution logic
+- apply/reject endpoints
+- rollback endpoint
+
+Repository default YAML remains untouched.
+
+
+## Phase 4: Limited Auto-Apply
+
+Goal:
+
+- safely automate a narrow slice of low-risk repairs
+
+Scope:
+
+- policy engine for auto-apply
+- same-domain or trusted-domain checks
+- structure validation
+- staged verification after apply
+
+Deliverables:
+
+- auto-apply rules
+- audit logs
+- automatic post-apply health verification
+- auto-disable or rollback on regression
+
+
+## Data Model Draft
+
+### datasource_health_checks
+
+Purpose:
+
+- store each health evaluation result
+
+Suggested fields:
+
+- `id`
+- `datasource_id`
+- `collector_name`
+- `endpoint_checked`
+- `status`
+- `http_status`
+- `content_type`
+- `latency_ms`
+- `sample_count`
+- `error_message`
+- `details`
+- `checked_at`
+
+`details` can store structured diagnostic data such as:
+
+- parsed fields
+- schema mismatch summary
+- retry count
+- exception class
+
+
+### datasource_repair_proposals
+
+Purpose:
+
+- store agent-generated repair suggestions
+
+Suggested fields:
+
+- `id`
+- `datasource_id`
+- `collector_name`
+- `old_endpoint`
+- `candidate_endpoint`
+- `reason`
+- `confidence`
+- `evidence_urls`
+- `evidence_summary`
+- `status`
+- `created_by`
+- `created_at`
+- `reviewed_at`
+
+Suggested `status` values:
+
+- `proposed`
+- `approved`
+- `rejected`
+- `applied`
+- `expired`
+
+
+### datasource_endpoint_overrides
+
+Purpose:
+
+- runtime endpoint override layer
+
+Suggested fields:
+
+- `id`
+- `datasource_id`
+- `collector_name`
+- `endpoint`
+- `reason`
+- `source`
+- `proposal_id`
+- `enabled`
+- `created_at`
+- `updated_at`
+
+Suggested `source` values:
+
+- `manual`
+- `health-agent`
+- `migration`
+
+
+## API Draft
+
+### Health
+
+- `GET /api/v1/datasources/health`
+- `GET /api/v1/datasources/{id}/health`
+- `POST /api/v1/datasources/{id}/health-check`
+- `POST /api/v1/datasources/health-check-all`
+
+### Repair proposals
+
+- `GET /api/v1/datasources/{id}/repair-proposals`
+- `POST /api/v1/datasources/{id}/repair-proposals/generate`
+- `POST /api/v1/datasources/{id}/repair-proposals/{proposal_id}/approve`
+- `POST /api/v1/datasources/{id}/repair-proposals/{proposal_id}/reject`
+- `POST /api/v1/datasources/{id}/repair-proposals/{proposal_id}/apply`
+
+### Overrides
+
+- `GET /api/v1/datasources/{id}/overrides`
+- `POST /api/v1/datasources/{id}/overrides`
+- `PUT /api/v1/datasources/{id}/overrides/{override_id}`
+- `DELETE /api/v1/datasources/{id}/overrides/{override_id}`
+
+
+## Agent Contract Draft
+
+When deterministic health fails, the agent should receive:
+
+- datasource name
+- collector name
+- current endpoint
+- current failure mode
+- expected response shape summary
+- known trusted domains
+
+Expected output:
+
+```json
+{
+  "status": "proposal",
+  "candidate_endpoint": "https://example.com/api/v2/data",
+  "confidence": 0.86,
+  "reason": "Official docs now point to v2 endpoint",
+  "evidence_urls": [
+    "https://example.com/docs/api",
+    "https://example.com/changelog"
+  ],
+  "notes": "Response shape appears compatible after light field remapping"
+}
+```
+
+The agent should never output "rewrite the default yaml" as its primary action.
+
+
+## Risk Analysis
+
+### Risk: wrong endpoint chosen by agent
+
+Mitigation:
+
+- use trusted-domain allowlists
+- require evidence URLs
+- require confidence threshold
+- add manual review for medium-risk sources
+
+
+### Risk: endpoint responds but schema silently changed
+
+Mitigation:
+
+- deterministic schema checks
+- parse and sample validation
+- content-type checks
+- collector-specific required fields
+
+
+### Risk: automatic runtime override causes hidden drift
+
+Mitigation:
+
+- store all overrides explicitly
+- mark source of override
+- keep default YAML unchanged
+- expose active overrides in API/UI
+
+
+### Risk: persistent bad override breaks data collection
+
+Mitigation:
+
+- allow rollback
+- keep parent/default endpoint visible
+- re-run verification after apply
+- auto-disable override on repeated failure
+
+
+## Operational Policy Recommendations
+
+1. Do not auto-apply for high-value or high-fragility sources initially.
+
+2. Use manual approval for:
+
+- scraped HTML sources
+- unofficial mirrors
+- sources with auth or rate-limit complexity
+- sources with legal or trust ambiguity
+
+3. Allow auto-apply only for:
+
+- same-domain version bumps
+- obvious official migration paths
+- repeated passing verification
+
+4. Expose health + proposal + override state together in one operator view.
+
+
+## Suggested Implementation Order
+
+1. Phase 1
+- health result table
+- deterministic checks
+- API and UI visibility
+
+2. Phase 2
+- proposal table
+- agent prompt/output contract
+- proposal generation job
+
+3. Phase 3
+- runtime override table
+- resolver precedence update
+- apply/reject endpoints
+
+4. Phase 4
+- auto-apply rules
+- rollback policy
+- operator automation
+
+
+## Out Of Scope For The First Iteration
+
+- direct automatic mutation of repository default YAML
+- automatic git commits by repair agents
+- unrestricted autonomous endpoint replacement
+- fully generalized schema remapping engine
+
+
+## Recommended First Milestone
+
+The first milestone should be:
+
+- deterministic datasource health checks
+- persisted results
+- manual visibility
+- no automatic repair
+
+This gives immediate operational value with low risk, and prepares clean inputs for the later agent phase.

docs/datasource-health-stage2-tasks.md

@@ -0,0 +1,478 @@
+# Datasource Health Stage 2 Tasks
+
+## Goal
+
+Stage 2 focuses on the first practical operational layer:
+
+- deterministic datasource health checks
+- persisted health results
+- health visibility through API and UI
+- no agent-assisted repair yet
+
+This stage should make Planet capable of answering:
+
+- which collectors are healthy
+- which collectors are degraded
+- which collectors are failing
+- why they are failing at a basic deterministic level
+
+
+## Scope
+
+Included:
+
+- datasource health data model
+- deterministic health check service
+- manual and scheduled health check triggers
+- health result APIs
+- frontend visibility
+
+Excluded:
+
+- LLM reasoning
+- web-search-based repair proposals
+- automatic endpoint rewriting
+- runtime override application
+
+
+## Delivery Target
+
+At the end of Stage 2, an operator should be able to:
+
+1. see health status for each collector
+2. trigger a health check manually
+3. inspect the latest failure reason
+4. inspect the last checked endpoint
+5. understand whether the problem is:
+   - unreachable
+   - auth-related
+   - rate-limit-related
+   - schema-related
+   - empty-data-related
+
+
+## Work Breakdown
+
+## A. Data Model
+
+### A1. Add datasource health record table
+
+Create a new model, for example:
+
+- `backend/app/models/datasource_health_check.py`
+
+Suggested fields:
+
+- `id`
+- `datasource_id`
+- `collector_name`
+- `endpoint_checked`
+- `status`
+- `http_status`
+- `content_type`
+- `latency_ms`
+- `sample_count`
+- `error_message`
+- `details`
+- `checked_at`
+
+Suggested status enum values:
+
+- `healthy`
+- `degraded`
+- `failed`
+- `schema_changed`
+- `rate_limited`
+- `auth_required`
+- `empty_result`
+
+
+### A2. Add datasource health summary fields
+
+Option A:
+
+- keep summary only in the health check table
+
+Option B:
+
+- also add summary fields on `data_sources`
+
+Recommended first step:
+
+- do not mutate `data_sources` schema yet
+- derive summary from the latest health record
+
+
+### A3. Migration task
+
+Add migration for the health table.
+
+Deliverables:
+
+- migration file
+- model registration
+
+
+## B. Health Check Engine
+
+### B1. Define health check service
+
+Add a new service module, for example:
+
+- `backend/app/services/datasource_health.py`
+
+Responsibilities:
+
+- resolve effective endpoint
+- execute deterministic check
+- classify result
+- persist health record
+
+
+### B2. Define shared result schema
+
+Create a typed result object, for example:
+
+- `HealthCheckResult`
+
+Suggested fields:
+
+- `status`
+- `endpoint_checked`
+- `http_status`
+- `content_type`
+- `latency_ms`
+- `sample_count`
+- `error_message`
+- `details`
+
+
+### B3. Implement base deterministic checks
+
+Every datasource should go through a minimal baseline check:
+
+1. resolve endpoint
+2. perform request
+3. measure latency
+4. inspect status code
+5. inspect content type
+6. inspect body shape
+
+Classification rules:
+
+- network error -> `failed`
+- HTTP 401/403 -> `auth_required`
+- HTTP 429 -> `rate_limited`
+- HTTP 404/410 -> `failed`
+- parse failure -> `schema_changed`
+- zero or suspiciously empty results -> `empty_result` or `degraded`
+- valid parse -> `healthy`
+
+
+### B4. Add collector-aware adapters
+
+Some collectors do not use the same fetch semantics.
+
+Add adapter profiles such as:
+
+- `http_json`
+- `http_csv`
+- `html_scrape`
+- `stream_probe`
+- `auth_session_http`
+
+Initial mapping suggestion:
+
+- `huggingface`, `peeringdb`, `cloudflare` -> `http_json`
+- `fao` -> `http_csv`
+- `top500`, `epoch_ai`, `telegeography live_map` -> `html_scrape`
+- `ris_live` -> `stream_probe`
+- `spacetrack` -> `auth_session_http`
+
+
+### B5. Add sample validation hooks
+
+For each adapter, add a lightweight validation rule.
+
+Examples:
+
+- JSON array length > 0
+- CSV rows > 1
+- HTML page contains expected table or script patterns
+- stream source yields at least one valid event within timeout
+
+
+## C. Persistence and Query Layer
+
+### C1. Save every check run
+
+Each health check should insert a record.
+
+Do not overwrite history in Stage 2.
+
+
+### C2. Add latest-health query helpers
+
+Add helper functions to fetch:
+
+- latest health record by datasource
+- latest failed health record
+- recent health history
+
+
+### C3. Optional retention policy
+
+For Stage 2, retention can be deferred.
+
+If desired, keep only:
+
+- last N records per datasource
+
+
+## D. API Layer
+
+### D1. Add health list endpoint
+
+Suggested endpoint:
+
+- `GET /api/v1/datasources/health`
+
+Returns:
+
+- datasource id
+- collector name
+- current endpoint
+- latest health status
+- last checked time
+- short reason
+
+
+### D2. Add per-datasource health detail endpoint
+
+Suggested endpoint:
+
+- `GET /api/v1/datasources/{id}/health`
+
+Returns:
+
+- latest record
+- recent history
+- detailed classification fields
+
+
+### D3. Add manual health trigger endpoint
+
+Suggested endpoint:
+
+- `POST /api/v1/datasources/{id}/health-check`
+
+Behavior:
+
+- run a health check now
+- persist the result
+- return the new record
+
+
+### D4. Add bulk health trigger endpoint
+
+Suggested endpoint:
+
+- `POST /api/v1/datasources/health-check-all`
+
+Behavior:
+
+- enqueue or run health checks for all active datasources
+
+
+## E. Scheduling
+
+### E1. Add health scheduler task
+
+Decide scheduling strategy.
+
+Recommended first version:
+
+- run collector jobs and health checks separately
+- health checks run on a lower frequency
+
+Suggested frequency:
+
+- every 6h or 12h for most datasources
+- optionally on-demand only in the very first cut
+
+
+### E2. Prevent health check collision with collection
+
+Rules:
+
+- health checks should not disrupt active collection
+- they should use light requests
+- if a collector is currently running, health check may:
+  - skip
+  - or use a lightweight endpoint probe only
+
+
+## F. Frontend
+
+### F1. Add health columns to datasource list
+
+Update:
+
+- `frontend/src/pages/DataSources/DataSources.tsx`
+
+Suggested new columns:
+
+- health status
+- last checked
+- reason summary
+
+
+### F2. Add manual health check action
+
+Per datasource:
+
+- button or dropdown action:
+  - `健康检查`
+
+
+### F3. Add health detail drawer or modal
+
+Show:
+
+- endpoint checked
+- status
+- HTTP status
+- content type
+- sample count
+- error message
+- last few results
+
+
+### F4. Add basic visual language
+
+Suggested colors:
+
+- green -> healthy
+- yellow -> degraded
+- orange -> rate-limited / auth-required
+- red -> failed / schema-changed
+
+
+## G. Observability
+
+### G1. Structured logging
+
+Every health check should log:
+
+- datasource id
+- collector name
+- endpoint
+- status
+- latency
+- failure class
+
+
+### G2. Optional metrics
+
+If metrics are added later, useful counters include:
+
+- health checks total
+- health checks failed
+- schema changes detected
+- rate limited checks
+
+
+## H. Tests
+
+### H1. Unit tests
+
+Add tests for:
+
+- status classification
+- content type classification
+- adapter behavior
+- latest-health query helpers
+
+
+### H2. API tests
+
+Add tests for:
+
+- health endpoints require auth
+- manual trigger endpoint works
+- list endpoint returns latest status
+
+
+### H3. Failure-path tests
+
+Add coverage for:
+
+- HTTP 404
+- HTTP 429
+- invalid JSON
+- empty response
+- parse mismatch
+
+
+## Suggested File Plan
+
+Possible implementation files:
+
+- `backend/app/models/datasource_health_check.py`
+- `backend/app/services/datasource_health.py`
+- `backend/app/schemas/datasource_health.py`
+- `backend/app/api/v1/datasource_health.py`
+- migration file under the project migration system
+
+Likely touched existing files:
+
+- `backend/app/api/main.py`
+- `frontend/src/pages/DataSources/DataSources.tsx`
+- `backend/tests/test_api.py`
+
+
+## Suggested Execution Order
+
+1. Add model and migration
+2. Add service and result schema
+3. Add deterministic adapters
+4. Add manual trigger API
+5. Add list/detail API
+6. Add frontend visibility
+7. Add scheduled checks
+8. Expand tests
+
+
+## Minimal First Milestone
+
+If we want the fastest useful slice, do this first:
+
+1. health table
+2. deterministic check service
+3. manual per-datasource health check API
+4. latest health list API
+5. frontend status badge column
+
+That is enough to start operating the system and will provide the input layer for Stage 3.
+
+
+## Dependency On Later Stages
+
+Stage 2 outputs become direct inputs for Stage 3.
+
+Specifically:
+
+- failed or schema-changed health records become agent triggers
+- health history becomes repair context
+- endpoint_checked becomes proposal baseline
+
+
+## Success Criteria
+
+Stage 2 is done when:
+
+- every active datasource can be health-checked deterministically
+- the latest health state is visible in API and UI
+- operators can manually trigger checks
+- failures are categorized into stable machine-readable statuses
+- no LLM is required for core health visibility

docs/docker-compose-buildx-upgrade.md

@@ -0,0 +1,105 @@
+# Docker + Compose + Buildx 升级教程
+
+流程：删除旧版 -> 安装新版 -> 验证
+
+---
+
+# 1. 删除旧版本
+
+## 删除 apt 安装的旧包
+
+```bash
+sudo apt remove -y docker.io docker-compose docker-compose-v2 docker-doc podman-docker containerd runc
+```
+
+---
+
+## 删除系统中的 `docker-compose`（V1）
+
+```bash
+sudo rm -f "$(which docker-compose 2>/dev/null)"
+```
+
+---
+
+## 查找并删除手动安装的 Buildx 插件
+
+```bash
+docker info | sed -n '/Plugins:/,/^ Server:/p' | grep -A2 buildx
+```
+
+从输出中获取 `Path`，然后执行：
+
+```bash
+rm -f <Path中对应的docker-buildx文件>
+```
+
+---
+
+## 清理无用依赖
+
+```bash
+sudo apt autoremove -y
+```
+
+---
+
+# 2. 安装 Docker 官方版本
+
+包含 Docker Engine、Docker Compose 插件、Docker Buildx 插件。
+
+## 安装依赖
+
+```bash
+sudo apt update
+sudo apt install -y ca-certificates curl gnupg
+```
+
+---
+
+## 添加 Docker GPG key
+
+```bash
+sudo install -m 0755 -d /etc/apt/keyrings
+curl -fsSL https://download.docker.com/linux/ubuntu/gpg | \
+  sudo gpg --dearmor -o /etc/apt/keyrings/docker.gpg
+sudo chmod a+r /etc/apt/keyrings/docker.gpg
+```
+
+---
+
+## 添加官方仓库
+
+```bash
+echo "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.gpg] https://download.docker.com/linux/ubuntu $(. /etc/os-release && echo "$VERSION_CODENAME") stable" | \
+sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
+```
+
+---
+
+## 安装 Docker + Compose + Buildx
+
+```bash
+sudo apt update
+sudo apt install -y docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin
+```
+
+---
+
+# 3. 验证安装
+
+```bash
+docker --version
+docker compose version
+docker buildx version
+```
+
+---
+
+# 4. 常用命令
+
+```bash
+docker compose up -d
+docker compose down
+docker buildx build .
+```

docs/earth-tv-live-module-plan.md

@@ -0,0 +1,117 @@
+# Earth 电视直播模块计划
+
+## 目标
+
+为 `Earth` 页面增加一个可配置、可扩展、可拖拽的电视直播模块：
+
+- 后台可配置新闻直播源
+- 默认兜底源为央视 `CCTV-4`
+- 未来可通过采集器接入世界各地新闻直播源
+- Earth 工具栏 `显示控制` 子菜单新增电视按钮
+- 点击后打开一个与其他 HUD 一致的可拖拽/可关闭窗口
+- 窗口内部可播放或承载新闻直播页面
+
+## 设计原则
+
+- 第一阶段先交付“后台可配 + Earth 可用 + 默认可回退”的版本
+- 公开读取接口与后台管理接口分离
+- 手工配置源与采集器源共用统一的前端消费结构
+- Earth 里的电视窗口必须复用现有 HUD 拖拽、关闭、布局最大化逻辑
+- 小屏下优先保证窗口完整显示，超出部分在窗口内部滚动
+
+## 分阶段实现
+
+### Phase 1：后端配置与公开读取
+
+- 在系统设置中新增 `tv` 分类
+- 定义直播源配置结构：
+  - `default_source_id`
+  - `auto_fallback`
+  - `sources[]`
+- 每个直播源至少包含：
+  - `id`
+  - `name`
+  - `provider`
+  - `region`
+  - `language`
+  - `source_type`
+  - `embed_url`
+  - `stream_url`
+  - `homepage_url`
+  - `is_enabled`
+  - `is_fallback`
+  - `sort_order`
+  - `collector_source`
+  - `notes`
+- 默认兜底源使用央视官网 `CCTV-4` 直播页
+- 新增公开读取接口，供 Earth 页面无登录态读取直播源配置
+
+### Phase 2：采集器扩展位
+
+- 新增 `news_live_streams` collector 占位
+- 规范采集器入库数据结构，使其能与后台手工配置源合并
+- TV 公开接口支持合并：
+  - 后台手工配置源
+  - 采集器入库源
+- 保持手工配置源优先级更高，避免采集器覆盖人工兜底配置
+
+### Phase 3：后台配置界面
+
+- 在系统配置页新增 `电视直播` tab
+- 支持：
+  - 查看当前默认源
+  - 开关自动回退
+  - 新增直播源
+  - 编辑直播源
+  - 删除直播源
+  - 启用/禁用直播源
+  - 将某个直播源设为默认源
+- 明确区分：
+  - 手工配置源
+  - 采集器来源
+
+### Phase 4：Earth HUD 集成
+
+- 在 `显示控制` 子菜单加入电视按钮
+- 新增 TV HUD 面板：
+  - 可拖拽
+  - 可关闭
+  - 支持显示/隐藏状态同步
+  - 参与布局最大化与恢复布局
+- 面板内容至少包含：
+  - 当前频道标题
+  - 源切换下拉菜单
+  - 刷新按钮
+  - 打开官网按钮
+  - 播放区域
+
+### Phase 5：播放策略
+
+- 第一版优先支持 `iframe`/嵌入页类直播源
+- 为未来扩展保留：
+  - `hls`
+  - `video`
+  - `external`
+- 如果默认源不可用：
+  - 优先回退到标记为 `is_fallback=true` 的源
+  - 若无明确回退源，则回退到第一个可用源
+- 面板内要有清晰的加载、错误、回退提示
+
+### Phase 6：打磨与清理
+
+- 统一 HUD 风格
+- 小屏下限制窗口尺寸并启用内部滚动
+- 避免窗口超出屏幕
+- 补最小验证
+- 清理临时代码、重复样式和无用资源
+
+## 首版交付定义
+
+当以下条件满足时，认为首版可用：
+
+- 后台可以配置新闻直播源
+- Earth 可以读取并显示默认直播源
+- 工具栏可打开电视窗口
+- 电视窗口可拖拽、可关闭
+- 央视 `CCTV-4` 作为默认兜底源可被使用
+- 代码结构已为后续采集器接入预留统一接口

docs/frontend-layout-guidelines.md

@@ -0,0 +1,309 @@
+# Frontend Layout Guidelines
+
+本项目后台页面默认遵循“单屏工作区”布局规范。目标不是让页面永远不溢出，而是确保在常见桌面视口下：
+
+- 页面主结构能在一屏内看清
+- 用户能同时看到页头、摘要区和主工作区
+- 超出的内容在模块内部滚动，而不是把整页纵向撑爆
+
+当前推荐参考实现：
+
+- [frontend/src/pages/BGP/BGP.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/BGP/BGP.tsx)
+- [frontend/src/index.css](/home/ray/dev/linkong/planet/frontend/src/index.css)
+
+## 核心原则
+
+### 1. 页面优先保证一屏工作区
+
+管理页默认采用：
+
+- 页头：标题、说明、主要操作
+- 主工作区：统计卡、表格、图表、列表、标签页
+
+推荐结构：
+
+```tsx
+<AppLayout>
+  <div className="page-shell">
+    <div className="page-shell__header">...</div>
+    <div className="page-shell__body">...</div>
+  </div>
+</AppLayout>
+```
+
+页面总高度应被限制在 `AppLayout` 内容区内，而不是继续让整个页面自然向下增长。
+
+### 2. 滚动优先发生在模块内部
+
+如果表格、日志、长列表、图表明细超出空间：
+
+- 让卡片内部滚动
+- 让表格内部滚动
+- 让标签页内容区内部滚动
+
+不要默认依赖整个页面滚动去“解决”空间问题。
+
+### 3. 主工作区必须拿到主要空间
+
+页面里最重要的模块必须是视觉和空间上的主角。通常应保证：
+
+- 页头始终可见
+- 摘要区高度被控制
+- 主表格 / 主图表 / 主分析区占据 50% 以上可视高度
+
+如果一个页面有多个大模块，优先顺序是：
+
+1. 先压缩说明区和摘要区
+2. 再把次级模块收进标签页或切换视图
+3. 最后才考虑继续增加整页滚动
+
+### 4. 小屏幕和高缩放必须进入紧凑模式
+
+在窗口高度较低、宽度较窄、或系统缩放较高时，应主动切换紧凑布局，例如：
+
+- 缩小卡片 padding
+- 缩小表头和单元格间距
+- 将摘要区改为更紧凑的单行/横向滚动布局
+- 将次级模块移入标签页、抽屉、折叠区
+
+紧凑模式的目标是保持可用，不是单纯把文字和控件一股脑缩小。
+
+### 5. overflow 责任必须明确
+
+页面中的大块内容必须明确：
+
+- 谁负责占满剩余高度
+- 谁负责裁剪
+- 谁负责滚动
+
+常见要求：
+
+- 父容器链路需要 `min-height: 0`
+- 工作区容器通常需要 `display: flex`
+- 真正的滚动节点要显式 `overflow: auto`
+
+### 6. 卡片不能被压到不可读
+
+历史上我们反复踩到的问题不是“没有滚动条”，而是：
+
+- 卡片被 `flex` 压缩得只剩一小条可视区域
+- 文字能渲染，但读不完整
+- 内容其实存在，却被 `overflow: hidden` 裁掉
+
+因此后续约束是：
+
+- 先保证卡片有可读的最小高度
+- 如果继续压缩会影响阅读，就切换成内部滚动
+- 不要为了“保持一屏”而把正文、表格、描述区压成无法阅读的条状区域
+
+### 7. Tabs 不是天然安全的布局容器
+
+历史上 Tabs 相关回归非常多，典型问题包括：
+
+- 隐藏 tab pane 因为自定义 `display: flex` 而重新露出来
+- 所有 tab 被强行套用同一套高度/overflow 规则
+- 表格 tab 能工作，但 markdown / help / diagnostics tab 被压坏
+
+因此约束是：
+
+- `Tabs` 里的每类内容都要单独定义自己的布局策略
+- 表格 tab 可以是“固定高度 + 内部滚动”
+- 文档/Markdown tab 更适合“tab pane 自身滚动 + 内容正常文档流”
+- 如果覆盖组件库样式，必须同时检查 hidden 状态是否仍然成立
+
+### 8. 摘要区优先进入紧凑模式，而不是挤压正文
+
+历史经验表明，最容易被误处理的是顶部摘要卡：
+
+- 它们经常为了“都放下”被强行压窄
+- 然后正文、表格、AI 结果区一起失去主空间
+
+后续统一约束：
+
+- 小屏或高缩放时，摘要卡优先：
+  - 降低 padding
+  - 改成横向滚动
+  - 改成更紧凑的网格
+- 不要优先牺牲主工作区的可视面积
+
+### 9. 长文档类内容优先保证阅读体验
+
+像下面这些内容，不能直接套用“表格工作区”的逻辑：
+
+- AI 简报
+- 运行日志
+- 原始 JSON
+- 帮助说明
+- 多段描述性文本
+
+这些区域应该优先满足：
+
+- 标题和元信息稳定可见
+- 正文有明确的最小可读高度
+- 正文滚动策略单独定义
+- 支持 Markdown 表格、分隔线、引用、代码块等结构
+
+### 10. 高度关键路径要少包一层
+
+历史上不少滚动问题不是组件本身错，而是多包了一层之后：
+
+- 高度链路断掉
+- `min-height: 0` 没传下去
+- `overflow` 责任被吃掉
+
+因此：
+
+- 对高度关键区域，优先使用最直接的 DOM 结构
+- 使用 `Space`、额外包装 `div`、第三方布局容器时，要确认它们不会改变滚动和高度语义
+- 如果一个区域已经出现“内容明明有，但只剩一条缝”，优先怀疑中间包装层
+
+## 历史坑位总结
+
+从 Earth、Playground、BGP、DataSources 这些页面的 bugfix 可以归纳出几类高频坑：
+
+### 1. 用 `overflow: hidden` 掩盖布局问题
+
+表面上看页面“整齐了”，实际上会导致：
+
+- 内容被裁掉
+- tab 内容只剩一条缝
+- 面板明明渲染成功，但用户看不见
+
+正确做法：
+
+- 让真正的内容节点滚动
+- 不要让上层容器无差别裁剪所有子内容
+
+### 2. 把所有 tab 当成同一种内容
+
+表格、Markdown、帮助卡、日志流的空间需求完全不同。
+
+正确做法：
+
+- 表格：固定工作区 + 内部滚动
+- 文档：普通流式内容 + pane 级滚动
+- 侧边说明：内容驱动高度，不强行拉满
+
+### 3. 只做视觉缩小，不做空间重分配
+
+这会导致：
+
+- 卡片文字被截断
+- 表格只剩 1 到 2 行
+- 按钮和筛选区挤成一团
+
+正确做法：
+
+- 紧凑模式优先重排
+- 横向滚动摘要区
+- 折叠/收纳次级模块
+
+### 4. 父容器高度链不完整
+
+这是最常见的内部滚动失效原因。
+
+检查顺序：
+
+1. 外层是否真的有确定高度
+2. flex 父容器是否带了 `min-height: 0`
+3. 真正滚动节点是否明确 `overflow: auto`
+4. 中间包装层是否偷偷改了布局语义
+
+### 5. UI 状态和显示状态不同步
+
+Earth 相关改动里反复出现：
+
+- 图层隐藏了，但 hover/lock 还在
+- tooltip 还在显示旧对象
+- legend 没跟着切换
+
+这类约束同样适用于后台页面：
+
+- 被隐藏、卸载、切换出视图的内容，不应继续保留活跃交互状态
+
+## 推荐实现模式
+
+### 页面骨架
+
+优先复用项目里已有的通用结构：
+
+- `.dashboard-content-inner`
+- `.page-shell`
+- `.page-shell__header`
+- `.page-shell__body`
+- `.table-scroll-region`
+
+不要每个页面都重新发明一套完全不同的高度和滚动语义。
+
+### 表格工作区
+
+推荐模式：
+
+```tsx
+<Card>
+  <div className="table-scroll-region" ref={tableRegionRef}>
+    <Table
+      pagination={false}
+      scroll={{ x: 1200, y: tableHeight }}
+    />
+  </div>
+</Card>
+```
+
+要求：
+
+- 表格尽量在卡片内部滚动
+- `scroll.y` 应来自实际可用高度估算，而不是完全静态的魔法数字
+- 父容器链路要保证 header、body、content 的 overflow 都在表格内部闭合
+
+### 多模块页面
+
+如果一个页面同时有：
+
+- 摘要卡
+- 表格
+- 异常明细
+- 最近事件
+
+不建议简单纵向堆叠全部模块。优先使用：
+
+- 顶部摘要 + 底部单一主工作区
+- 标签页切换多个次级数据视图
+- 左右分栏，并保证每栏内部独立滚动
+
+## 不推荐的做法
+
+以下模式默认视为不符合本项目页面规范：
+
+- 依赖整页纵向滚动来显示主要工作区
+- 一个页面纵向堆 3 到 4 个大卡片，每个都想完整展示
+- 表格没有内部滚动，导致缩放后只能看到 1 到 2 行数据
+- 父容器缺少 `min-height: 0`，导致内部滚动失效
+- 只做视觉缩小，不处理真正的空间分配
+
+## 页面验收检查清单
+
+提交前至少检查：
+
+- 页头、摘要区、主工作区能否同时出现
+- 主工作区是否拿到了页面中最多的高度
+- 表格或明细溢出时，滚动条是否出现在模块内部
+- 卡片是否被压缩到文字显示不完整；如果会，是否已经切换为内部滚动
+- 浏览器缩放到 `125%` / `150%` 时是否仍可用
+- 低高度窗口下是否还保有合理的可见内容行数
+- Tabs、Card、Table 在 overflow 时是否仍可操作
+- 非表格 tab（Markdown、帮助说明、日志）是否有独立且合理的滚动策略
+
+## 落地顺序
+
+后续新增或重构后台页时，优先按这个顺序设计：
+
+1. 先定义主工作区
+2. 再确定哪些模块必须常驻可见
+3. 最后再做样式和视觉层次
+
+简单说：
+
+- 先保证空间分配正确
+- 再处理滚动边界
+- 最后再做美化

docs/news-live-streams-collector-format.md

@@ -0,0 +1,97 @@
+# News Live Streams Collector Format
+
+`news_live_streams` 采集器面向“频道目录 JSON”输入，而不是直接抓网页。
+
+这样做的目标是：
+
+- 让后台能够稳定接入世界各地新闻直播源
+- 让 `Earth` 页面电视模块始终消费统一结构
+- 便于后续接入类似 `worldmonitor` 那种 YouTube / HLS / iframe 混合频道目录
+
+## 推荐 JSON 结构
+
+```json
+{
+  "sources": [
+    {
+      "id": "bbc-world-news",
+      "name": "BBC World News",
+      "provider": "BBC",
+      "region": "UK",
+      "language": "en",
+      "source_type": "youtube",
+      "youtube_video_id": "dQw4w9WgXcQ",
+      "youtube_channel": "https://www.youtube.com/@BBCNews",
+      "embed_url": "",
+      "stream_url": "",
+      "homepage_url": "https://www.youtube.com/@BBCNews/live",
+      "poster_url": "",
+      "sort_order": 220,
+      "is_enabled": true,
+      "notes": "Primary English global news channel"
+    },
+    {
+      "id": "france24-en",
+      "name": "France 24 English",
+      "provider": "France 24",
+      "region": "France",
+      "language": "en",
+      "source_type": "hls",
+      "stream_url": "https://example.com/live.m3u8",
+      "homepage_url": "https://www.france24.com/en/live",
+      "sort_order": 230,
+      "is_enabled": true
+    },
+    {
+      "id": "cctv4-page",
+      "name": "CCTV-4 中文国际",
+      "provider": "CCTV",
+      "region": "China",
+      "language": "zh-CN",
+      "source_type": "iframe",
+      "embed_url": "https://tv.cctv.com/live/cctv4/",
+      "homepage_url": "https://tv.cctv.com/live/cctv4/",
+      "sort_order": 10,
+      "is_enabled": true
+    }
+  ]
+}
+```
+
+## 字段约定
+
+- `id`: 唯一标识，建议稳定不变
+- `name`: 频道显示名
+- `provider`: 提供方
+- `region`: 国家或地区
+- `language`: 语言代码
+- `source_type`: `iframe` / `hls` / `video` / `external` / `youtube`
+- `embed_url`: 适合 iframe 内嵌的页面
+- `stream_url`: 直接视频流地址
+- `homepage_url`: 官网或频道页
+- `youtube_video_id`: YouTube 直播视频 ID
+- `youtube_channel`: YouTube 频道 handle 或频道 URL
+- `poster_url`: 封面图，可选
+- `sort_order`: 排序值，越小越靠前
+- `is_enabled`: 是否启用
+- `notes`: 简短备注
+
+## 面板行为约定
+
+- `youtube`
+  - 优先使用 `youtube_video_id`
+  - 无法内嵌时至少保留 `youtube_channel` 或 `homepage_url` 供外部打开
+- `hls` / `video`
+  - 优先走 `stream_url`
+- `iframe`
+  - 优先走 `embed_url`
+- `external`
+  - 不尝试内嵌，只保留外部打开
+
+## 当前实现状态
+
+- 后台设置页可以手工维护频道目录
+- `Earth` 电视模块会合并：
+  - 手工配置源
+  - `news_live_streams` 采集器采集源
+- 当前默认兜底源为 `CCTV-4 中文国际`

docs/situational-awareness-foundation-plan.md

@@ -0,0 +1,309 @@
+# Situational Awareness Foundation Plan
+
+## 定位
+
+当前这套 AI 能力应被视为 `态势感知服务底座`，而不是完整的态势感知产品。
+
+也就是说，现阶段的目标不是：
+
+- 做一个“什么都能分析”的万能 AI 页面
+- 让模型在证据不足时替代人工研判
+- 过早把页面做成完整指挥大屏
+
+现阶段真正要做的是：
+
+- 先把 `model gateway / backend facade / evidence injection / page-specific brief` 这几层边界搭稳
+- 让系统能够在已有证据上稳定地产出“可读、可回看、可扩展”的摘要
+- 为后续更强的数据联动、agent 推理和 assessment 结构化输出预留好接口与数据模型
+
+## 当前现实约束
+
+### 1. 数据维度不足
+
+目前系统能提供的主要证据仍集中在：
+
+- BGP incidents / anomalies / events
+- collector coverage
+- datasource health / platform alerts
+- prefix geography 的部分归属信息
+
+当前明显还缺：
+
+- 流量异常与业务指标
+- 电商、支付、物流等业务侧指标
+- 更丰富的资产、链路、区域、行业画像
+- 外部舆情、公告、运营商状态、基础设施事件等背景信息
+
+这意味着：
+
+- 模型现在可以做“基于现有证据的摘要与归纳”
+- 但还不能可靠地做“跨维度因果研判”
+
+### 2. 维度之间联动还弱
+
+目前不同模块之间更多是“并列展示”，还不是“强关联分析”：
+
+- 系统告警和 BGP 事件还没有统一事件模型
+- collector bias 与真实区域热度还没有完全剥离
+- datasource health 与 BGP 风险、业务影响之间还没有稳定映射
+
+这意味着：
+
+- 当前更适合做 `brief / overview / operator notes`
+- 还不适合过度承诺“自动态势判断”
+
+### 3. 结构化 assessment 还未成为主输出
+
+虽然已经有 BGP brief、系统告警 brief、态势告警 brief，但目前主输出仍偏向：
+
+- 文本摘要
+- facts/context 附带证据
+
+后续真正要服务态势感知，需要更稳定的结构化输出，例如：
+
+- summary
+- key risks
+- evidence
+- confidence
+- recommendations
+- missing data
+
+## 当前基座已经具备的能力
+
+### 1. AI 调用边界已经明确
+
+- `aiprovider` 负责模型协议与 provider 兼容
+- `backend` 负责业务 API、证据整合和鉴权
+- `frontend` 负责页面入口与结果展示
+
+### 2. 页面级 AI 入口已经开始成型
+
+当前已经有或正在收口的入口：
+
+- `Playground`
+  - 用于链路验证与 provider 诊断
+- `BGP AI 简报`
+  - 用于 BGP 事实摘要和区域风险归纳
+- `Alerts`
+  - 用于系统告警、BGP 告警、态势告警三类入口
+
+### 3. 证据优先的方向已经建立
+
+已经不再只依赖人工在 Playground 中手填 prompt，系统开始具备：
+
+- 从真实业务数据生成事实输入
+- 保存 facts/context 快照
+- 回看 AI 输出时同时回看证据
+
+这一步非常关键，因为它决定后面能否从“玩具 demo”走向“有运维价值的系统”。
+
+## 近期收尾建议
+
+这些事情都属于“底座收口”，值得做，但不应该再继续重产品包装。
+
+### 1. 统一 Alerts 页面
+
+已采用：
+
+- 一个 `Alerts` 页面
+- 三个 tab：
+  - `系统告警`
+  - `BGP 告警`
+  - `态势告警`
+
+收尾重点：
+
+- 保持 tab 的文案、摘要卡和 AI 简报交互一致
+- 不额外扩展成多个独立二级页面
+
+### 2. 保持 Playground 为测试台
+
+原则：
+
+- Playground 只承担链路验证、provider 状态诊断、请求结果观察
+- 不继续堆“万能业务分析器”式交互
+
+### 3. 把 brief 能力当服务能力而不是页面特效
+
+页面现在能看到按钮和结果，这很好，但更重要的是：
+
+- 后端接口稳定
+- facts/context 可追踪
+- 输出结构后续可升级
+
+### 4. 导航结构先收口，不继续平铺一级菜单
+
+随着后续能力扩展，系统很可能继续新增：
+
+- 海缆
+- 算力中心
+- 战争信息
+- 电商分析
+- 其他专题观测页
+
+如果继续把这些入口全部平铺在左侧一级菜单中，会带来两个问题：
+
+- 一级菜单过长，用户难以判断先进入哪个上下文
+- `观测页 / 告警页 / 研判页 / 运维页` 的职责边界会被混在一起
+
+因此近期应明确采用分组导航，而不是继续扩展平铺菜单。
+
+推荐的导航分组如下：
+
+- `总览`
+  - 仪表盘
+  - Earth
+- `专题观测`
+  - BGP 观测
+  - 采集数据
+  - 后续可扩展：海缆、算力中心、战争信息、电商分析
+- `告警与研判`
+  - Alerts
+- `运维与配置`
+  - 数据源
+  - AI Playground
+  - 用户管理
+  - 系统配置
+
+这套结构的含义是：
+
+- `专题观测` 页面负责看某个维度本身
+- `Alerts` 负责跨模块风险与值班工作台
+- `Playground` 保持为测试台，不挤占业务导航语义
+
+短期收尾时，应优先重组现有入口，而不是继续增加新的一级菜单。
+
+## 后续路线
+
+## Phase 1：服务底座稳固
+
+目标：
+
+- 不追求“更炫的 AI 页面”
+- 先把当前接口、证据、存储和页面入口收稳
+
+工作项：
+
+- 统一页面级 AI 入口模式
+- 统一 brief response schema
+- 保证 facts/context 在前后端都可回看
+- 继续清理 mock 和临时分支逻辑
+
+完成标准：
+
+- 每个 AI 入口都是真实链路
+- 每个 AI 结果都能追溯到证据输入
+
+## Phase 2：Evidence-first Assessment
+
+目标：
+
+- 从“文本摘要”升级成“结构化 assessment”
+
+工作项：
+
+- 为 brief/assessment 定义统一 schema
+- 固化：
+  - summary
+  - key_risks
+  - evidence
+  - confidence
+  - recommendations
+  - missing_data
+- 页面以结构化区块展示，而不只是大段文本
+
+完成标准：
+
+- AI 输出可持久化、可比较、可审计
+
+## Phase 3：多维证据接入
+
+目标：
+
+- 让“态势感知”真正拥有更多维度，而不是只靠 BGP 与系统告警
+
+优先接入方向：
+
+- datasource health findings
+- 流量或业务指标
+- 区域/资产/链路映射
+- 外部事件与公告
+- 业务垂直数据，例如电商分析相关指标
+
+完成标准：
+
+- AI 能基于多个维度做交叉说明
+- 不再只围绕单一模块自说自话
+
+## Phase 4：Correlation Layer
+
+目标：
+
+- 不同来源的信号不再只是并列，而是形成统一的事件关联
+
+工作项：
+
+- 统一 signal/finding 模型
+- 跨模块事件聚合
+- 证据来源权重
+- collector bias 与真实热度分离
+
+完成标准：
+
+- 系统能回答“这些异常是不是同一件事”
+- 系统能回答“哪些结论只是观测偏差”
+
+## Phase 5：Agent-assisted Situational Awareness
+
+目标：
+
+- 在证据足够的前提下，再让 agent 负责更复杂的推理与建议
+
+工作项：
+
+- 复用现有 agent runtime 规划
+- 引入 web search / docs fetch / repair proposal 等能力
+- 但始终坚持：
+  - evidence first
+  - proposal before action
+  - no silent mutation of defaults
+
+完成标准：
+
+- agent 成为证据驱动的分析层
+- 而不是一个“万能猜测层”
+
+## 设计原则
+
+### 1. 先底座，后产品化
+
+先把服务链路和证据模型做好，再做更大的页面表达。
+
+### 2. 先证据，后判断
+
+事实输入应先稳定，再让模型做归纳。
+
+### 3. 先专用 brief，后统一态势层
+
+先让各业务页有各自可信的 AI 入口，再考虑统一态势页。
+
+### 4. 先 proposal，后自动动作
+
+涉及修复、覆盖、写配置、调任务的动作，都应经过 proposal 和审计。
+
+## 当前建议结论
+
+对现在这个项目，最合理的定位是：
+
+- `Playground` 是测试台
+- `BGP / Alerts` 是第一批业务 AI 入口
+- `aiprovider + backend AI facade + evidence snapshots` 是核心服务底座
+
+现阶段不需要追求“已经具备完整态势感知能力”。
+
+现阶段真正的成功标准是：
+
+- 这套底座可用
+- 可回看
+- 可扩展
+- 不自欺欺人

docs/version-history.md

@@ -16,12 +16,13 @@
 ## Current Version
 
 - `main` 当前主线历史推导到：`0.16.5`
-- `dev` 当前开发分支历史推导到：`0.23.0`
+- `dev` 当前开发分支历史推导到：`0.27.0`
 
 ## Timeline
 
 | Version | Type | Branch | Commit | Summary |
 | --- | --- | --- | --- | --- |
+| `0.27.0` | feature | `dev` | — | Earth HUD 重构：图层面板、信息卡片悬浮定位、Fresnel 大气层渲染 |
 | `0.0.1-beta` | bootstrap | `main` | `e7033775` | first commit |
 | `0.1.0` | feature | `main` | `6cb4398f` | Modularize 3D Earth page with ES Modules |
 | `0.2.0` | feature | `main` | `aaae6a53` | Add cable graph service and data collectors |
@@ -71,6 +72,23 @@
 | `0.21.6` | bugfix | `dev` | `pending` | improve Earth legend generation, info-card interactions, and HUD messaging polish |
 | `0.22.9` | bugfix | `dev` | `6bfcd053` | simplify `planet.sh` readiness messaging and only show retry counts on actual restart |
 | `0.23.0` | feature | `dev` | `pending` | add dedicated `aiprovider` service, multi-protocol AI adapters, uv-only Python runtime, and AI Provider restart controls |
+| `0.23.3` | bugfix | `dev` | `pending` | refine `planet.sh` zsh runtime compatibility, startup logging, and frontend readiness feedback |
+| `0.24.0` | feature | `dev` | `pending` | add AI Playground entry, provider diagnostics, and frontend layout guidance |
+| `0.24.1` | bugfix | `dev` | `pending` | refactor Earth HUD into class-first CSS layers, unify Bun-only frontend tooling guidance, and auto-bootstrap Bun/uv in `planet.sh` |
+| `0.24.2` | bugfix | `dev` | `pending` | restore public Earth entry, refresh Playground provider diagnostics correctly, fit help-card content, and split frontend bundles by route/vendor |
+| `0.24.3` | bugfix | `dev` | `pending` | expand Playground diagnostics presets and result inspection, and make `planet.sh` rebuild changed AI Provider images with explicit Compose fallback reporting |
+| `0.24.4` | bugfix | `dev` | `pending` | polish `planet.sh` AI Provider rebuild stage boundaries, hide raw Compose build logs on success, and add explicit image-build completion feedback |
+| `0.24.5` | bugfix | `dev` | `pending` | add persistent BGP AI briefs with Markdown history, lazy-load BGP tabs, and move BGP hot-path filtering and aggregation back into the database |
+| `0.24.6` | bugfix | `dev` | `pending` | batch datasource and visualization hot-path queries, fix BGP collector JSON extraction, and rebuild the BGP AI brief tab layout and markdown rendering |
+| `0.24.7` | bugfix | `dev` | `pending` | formalize release workflow and frontend layout constraints with repo rules and a reusable release skill |
+| `0.24.8` | bugfix | `dev` | `pending` | move BGP brief markdown into a dedicated modal, keep tab content metadata-focused, and constrain modal scrolling to the viewport |
+| `0.25.0` | feature | `dev` | `89a71e6f` | add persistent backend-backed AI Playground chat state, split alert workspaces into dedicated pages, and establish the situational-awareness foundation for later multi-signal analysis |
+| `0.25.1` | bugfix | `dev` | `pending` | clean duplicated Playground flow code, add reusable code-hygiene rules, and fix first-level sidebar menu expansion behavior across route navigation and refresh |
+| `0.25.2` | bugfix | `dev` | `pending` | refine Earth settings modal sizing, eliminate first-frame HUD scale flicker, and make dragged HUD panels animate cleanly through maximized layout transitions |
+| `0.25.3` | bugfix | `dev` | `pending` | refactor the Earth HUD visual system, extract the top-left Earth brand into a reusable language-driven component, and consolidate duplicated brand assets into a single canonical set |
+| `0.26.0` | feature | `dev` | `pending` | add the Earth TV live module with backend-configurable sources, a draggable/resizable TV HUD window, a first curated news channel catalog, and TV source management hooks in system settings |
+| `0.26.1` | bugfix | `dev` | `pending` | extract the dashboard sidebar scrollbar into a reusable component and clean duplicated Earth TV player reset logic after the first live-module rollout |
+| `0.26.2` | bugfix | `dev` | `pending` | stabilize the reusable sidebar scrollbar, restore automatic dual-axis floating tracks safely, and apply the same overlay scrollbar system to datasource tables |
 
 ## Maintenance Commits Not Counted as Version Bumps
 

frontend/package.json

@@ -1,7 +1,8 @@
 {
   "name": "planet-frontend",
-  "version": "0.23.0",
+  "version": "0.27.0",
   "private": true,
+  "packageManager": "bun@1",
   "dependencies": {
     "@ant-design/icons": "^5.2.6",
     "antd": "^5.12.5",

frontend/public/earth/assets/brand/README.md

@@ -0,0 +1,9 @@
+Earth brand runtime assets live in this folder.
+
+- `earth-logo.png`
+- `title-zh.png`
+- `title-en.png`
+
+These runtime images were generated from the uploaded SVG artwork so the Earth
+page can keep the intended look without tracking unusually large pixel-rect SVG
+exports in Git history.

frontend/public/earth/assets/icons/cables.svg

@@ -1,6 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <circle cx="12" cy="12" r="6.75" stroke="#4DB8FF" stroke-width="2.1"/>
-  <path d="M5.75 12H18.25" stroke="#4DB8FF" stroke-width="2.1" stroke-linecap="round"/>
-  <path d="M12 5.8C14.7 7.75 14.7 16.25 12 18.2" stroke="#4DB8FF" stroke-width="2.1" stroke-linecap="round"/>
-  <path d="M8 16C9.95 14.2 14.05 14.2 16 16" stroke="#4DB8FF" stroke-width="2.1" stroke-linecap="round"/>
-</svg>

frontend/public/earth/assets/icons/info.svg

@@ -1,5 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <circle cx="12" cy="12" r="8" stroke="#4DB8FF" stroke-width="2.2"/>
-  <path d="M12 10V16" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <circle cx="12" cy="7.25" r="1.25" fill="#4DB8FF"/>
-</svg>

frontend/public/earth/assets/icons/layout-collapse.svg

@@ -1,10 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <path d="M9 9L5 5" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M5 8V5H8" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round" stroke-linejoin="round"/>
-  <path d="M15 9L19 5" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M16 5H19V8" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round" stroke-linejoin="round"/>
-  <path d="M9 15L5 19" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M5 16V19H8" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round" stroke-linejoin="round"/>
-  <path d="M15 15L19 19" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M16 19H19V16" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round" stroke-linejoin="round"/>
-</svg>

frontend/public/earth/assets/icons/layout.svg

@@ -1,10 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <path d="M6 9.2L9.2 9.2L9.2 6" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round" stroke-linejoin="round"/>
-  <path d="M18 9.2L14.8 9.2L14.8 6" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round" stroke-linejoin="round"/>
-  <path d="M6 14.8L9.2 14.8L9.2 18" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round" stroke-linejoin="round"/>
-  <path d="M18 14.8L14.8 14.8L14.8 18" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round" stroke-linejoin="round"/>
-  <path d="M6 6L10 10" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M18 6L14 10" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M6 18L10 14" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M18 18L14 14" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-</svg>

frontend/public/earth/assets/icons/pause.svg

@@ -1,4 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <path d="M9 6.25V17.75" stroke="#4DB8FF" stroke-width="2.4" stroke-linecap="round"/>
-  <path d="M15 6.25V17.75" stroke="#4DB8FF" stroke-width="2.4" stroke-linecap="round"/>
-</svg>

frontend/public/earth/assets/icons/play.svg

@@ -1,3 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <path d="M8.5 6.5L17.5 12L8.5 17.5V6.5Z" fill="#4DB8FF"/>
-</svg>

frontend/public/earth/assets/icons/reload.svg

@@ -1,6 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <path d="M6.8 12.6C6.8 8.95 9.75 6 13.4 6C14.9 6 16.24 6.46 17.3 7.28" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M18.85 10.45C19.2 11.15 19.4 11.95 19.4 12.8C19.4 16.45 16.45 19.4 12.8 19.4C10.05 19.4 7.69 17.72 6.7 15.33" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M15.9 5.95H19.2V9.25" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round" stroke-linejoin="round"/>
-  <path d="M19.2 5.95L16.7 8.45" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-</svg>

frontend/public/earth/assets/icons/reset-view.svg

@@ -1,8 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <circle cx="12" cy="12" r="5" stroke="#4DB8FF" stroke-width="2.2"/>
-  <path d="M12 3V6.5" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M12 17.5V21" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M3 12H6.5" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M17.5 12H21" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <circle cx="12" cy="12" r="1.45" fill="#4DB8FF"/>
-</svg>

frontend/public/earth/assets/icons/satellite.svg

@@ -1,11 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <rect x="10" y="10" width="4" height="4" rx="0.8" stroke="#4DB8FF" stroke-width="2"/>
-  <rect x="4" y="9" width="4" height="6" rx="0.8" stroke="#4DB8FF" stroke-width="2"/>
-  <rect x="16" y="9" width="4" height="6" rx="0.8" stroke="#4DB8FF" stroke-width="2"/>
-  <path d="M8 12H10" stroke="#4DB8FF" stroke-width="2" stroke-linecap="round"/>
-  <path d="M14 12H16" stroke="#4DB8FF" stroke-width="2" stroke-linecap="round"/>
-  <path d="M12 8V6" stroke="#4DB8FF" stroke-width="2" stroke-linecap="round"/>
-  <path d="M10.75 6H13.25" stroke="#4DB8FF" stroke-width="2" stroke-linecap="round"/>
-  <path d="M12 14V18" stroke="#4DB8FF" stroke-width="2" stroke-linecap="round"/>
-  <path d="M10.25 18H13.75" stroke="#4DB8FF" stroke-width="2" stroke-linecap="round"/>
-</svg>

frontend/public/earth/assets/icons/search.svg

@@ -1,4 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <circle cx="10.5" cy="10.5" r="5.75" stroke="#4DB8FF" stroke-width="2.2"/>
-  <path d="M15.2 15.2L19.25 19.25" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-</svg>

frontend/public/earth/assets/icons/terrain.svg

@@ -1,5 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <path d="M3 18H21" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M4.5 18L9.5 11L12.5 15L16 9L19.5 18" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round" stroke-linejoin="round"/>
-  <path d="M11.25 18L13.05 15.35L14.55 17.25" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round" stroke-linejoin="round"/>
-</svg>

frontend/public/earth/assets/icons/trails.svg

@@ -1,7 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <path d="M5 17H12" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M7 13.5H15" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M10 10H16" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <circle cx="17.5" cy="8.5" r="2.2" fill="#4DB8FF"/>
-  <path d="M15.8 10.2L18.55 7.45" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-</svg>

frontend/public/earth/assets/icons/zoom.svg

@@ -1,6 +0,0 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <circle cx="10.5" cy="10.5" r="5.75" stroke="#4DB8FF" stroke-width="2.2"/>
-  <path d="M15.25 15.25L19.25 19.25" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M10.5 8V13" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-  <path d="M8 10.5H13" stroke="#4DB8FF" stroke-width="2.2" stroke-linecap="round"/>
-</svg>

frontend/public/earth/css/base.css

@@ -1,4 +1,46 @@
-/* base.css - 公共基础样式 */
+/* base.css - app shell, tokens, and global primitives */
+
+@property --float-offset {
+  syntax: "<length>";
+  inherits: false;
+  initial-value: 0px;
+}
+
+:root {
+  --hud-scale: 1;
+  --hud-offset: calc(20px * var(--hud-scale));
+  --hud-radius: calc(22px * var(--hud-scale));
+  --hud-panel-padding: calc(18px * var(--hud-scale));
+  --hud-panel-padding-sm: calc(13px * var(--hud-scale));
+  --hud-gap-xs: calc(7px * var(--hud-scale));
+  --hud-gap-sm: calc(11px * var(--hud-scale));
+  --hud-gap-md: calc(16px * var(--hud-scale));
+  --hud-gap-lg: calc(22px * var(--hud-scale));
+  --hud-font-size: calc(0.88rem * var(--hud-scale));
+  --hud-font-size-sm: calc(0.75rem * var(--hud-scale));
+  --hud-title-size: calc(1.02rem * var(--hud-scale));
+  --hud-kicker-size: calc(0.68rem * var(--hud-scale));
+  --hud-surface-top: rgba(17, 31, 53, 0.84);
+  --hud-surface-bottom: rgba(7, 17, 31, 0.76);
+  --hud-surface-overlay: rgba(157, 204, 255, 0.07);
+  --hud-border: rgba(201, 225, 247, 0.14);
+  --hud-border-hover: rgba(226, 238, 250, 0.24);
+  --hud-border-active: rgba(235, 244, 255, 0.32);
+  --hud-shadow: 0 18px 46px rgba(1, 7, 16, 0.34);
+  --hud-shadow-soft: 0 10px 28px rgba(3, 10, 22, 0.22);
+  --hud-highlight: rgba(248, 252, 255, 0.14);
+  --hud-line: rgba(197, 220, 242, 0.1);
+  --hud-title: #dbe8f5;
+  --hud-text: #eef4fb;
+  --hud-text-muted: #8ea3ba;
+  --hud-text-soft: #6f849b;
+  --hud-accent: #91baff;
+  --hud-accent-strong: #d7e8ff;
+  --glass-fill-top: rgba(255, 255, 255, 0.08);
+  --glass-fill-bottom: rgba(109, 157, 214, 0.04);
+  --glass-shadow: 0 16px 36px rgba(0, 0, 0, 0.2);
+  --glass-glow: 0 0 18px rgba(123, 176, 236, 0.08);
+}
 
 * {
   margin: 0;
@@ -6,788 +48,100 @@
   box-sizing: border-box;
 }
 
-body {
-  font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
+html,
+body,
+.earth-page {
+  width: 100%;
+  height: 100%;
+}
+
+body.earth-page {
+  font-family: "Segoe UI", Tahoma, Geneva, Verdana, sans-serif;
   background-color: #0a0a1a;
   color: #fff;
   overflow: hidden;
 }
 
-:root {
-  --hud-border: rgba(210, 237, 255, 0.32);
-  --hud-border-hover: rgba(232, 246, 255, 0.48);
-  --hud-border-active: rgba(245, 251, 255, 0.62);
-  --glass-fill-top: rgba(255, 255, 255, 0.18);
-  --glass-fill-bottom: rgba(115, 180, 255, 0.08);
-  --glass-sheen: rgba(255, 255, 255, 0.34);
-  --glass-shadow: 0 14px 30px rgba(0, 0, 0, 0.22);
-  --glass-glow: 0 0 26px rgba(120, 200, 255, 0.16);
-}
-
-@property --float-offset {
-  syntax: '<length>';
-  inherits: false;
-  initial-value: 0px;
-}
-
-#container {
+.earth-app {
   position: relative;
   width: 100vw;
   height: 100vh;
 }
 
-#container.dragging {
+.earth-app.dragging {
   cursor: grabbing;
 }
 
-/* Bottom Dock */
-#right-toolbar-group {
+.earth-filters {
   position: absolute;
-  bottom: 18px;
-  left: 50%;
-  transform: translateX(-50%);
-  display: flex;
-  flex-direction: row;
-  align-items: center;
-  justify-content: center;
-  z-index: 200;
-}
-
-#right-toolbar-group,
-#info-panel,
-#coordinates-display,
-#legend,
-#earth-stats {
-  transition:
-    top 0.45s ease,
-    right 0.45s ease,
-    bottom 0.45s ease,
-    left 0.45s ease,
-    transform 0.45s ease,
-    box-shadow 0.45s ease;
-}
-
-#info-panel,
-#coordinates-display,
-#legend,
-#earth-stats,
-#satellite-info {
-  position: absolute;
-  overflow: hidden;
-  isolation: isolate;
-  background:
-    radial-gradient(circle at 24% 12%, rgba(255, 255, 255, 0.12), transparent 28%),
-    radial-gradient(circle at 78% 115%, rgba(255, 255, 255, 0.06), transparent 32%),
-    linear-gradient(180deg, rgba(255, 255, 255, 0.14), rgba(110, 176, 255, 0.06)),
-    rgba(7, 18, 36, 0.28);
-  border: 1px solid rgba(225, 242, 255, 0.2);
-  box-shadow:
-    inset 0 1px 0 rgba(255, 255, 255, 0.14),
-    inset 0 -1px 0 rgba(255, 255, 255, 0.04),
-    0 18px 40px rgba(0, 0, 0, 0.24),
-    0 0 32px rgba(120, 200, 255, 0.12);
-  backdrop-filter: blur(20px) saturate(145%);
-  -webkit-backdrop-filter: blur(20px) saturate(145%);
-}
-
-#info-panel::before,
-#coordinates-display::before,
-#legend::before,
-#earth-stats::before,
-#satellite-info::before {
-  content: '';
-  position: absolute;
-  inset: 1px 1px 24% 1px;
-  border-radius: inherit;
-  background:
-    linear-gradient(180deg, rgba(255, 255, 255, 0.18), rgba(255, 255, 255, 0.05) 26%, transparent 70%);
-  opacity: 0.46;
+  width: 0;
+  height: 0;
   pointer-events: none;
 }
 
-#info-panel::after,
-#coordinates-display::after,
-#legend::after,
-#earth-stats::after,
-#satellite-info::after {
-  content: '';
-  position: absolute;
-  inset: -1px;
-  padding: 1.4px;
-  border-radius: inherit;
-  background:
-    linear-gradient(135deg, rgba(255, 255, 255, 0.3), rgba(170, 223, 255, 0.2) 34%, rgba(88, 169, 255, 0.14) 68%, rgba(255, 255, 255, 0.24));
-  opacity: 0.78;
-  pointer-events: none;
-  filter: url(#liquid-glass-distortion) blur(0.35px);
-  -webkit-mask:
-    linear-gradient(#000 0 0) content-box,
-    linear-gradient(#000 0 0);
-  -webkit-mask-composite: xor;
-  mask:
-    linear-gradient(#000 0 0) content-box,
-    linear-gradient(#000 0 0);
-  mask-composite: exclude;
-}
-
-#info-panel > *,
-#coordinates-display > *,
-#legend > *,
-#earth-stats > *,
-#satellite-info > * {
-  position: relative;
-  z-index: 1;
-}
-
-#loading {
+.earth-loading {
   position: absolute;
   top: 50%;
   left: 50%;
   transform: translate(-50%, -50%);
-  font-size: 1.2rem;
-  color: #4db8ff;
-  z-index: 100;
+  z-index: 240;
+  min-width: min(calc(320px * var(--hud-scale)), 78vw);
+  padding: calc(26px * var(--hud-scale));
+  border-radius: calc(18px * var(--hud-scale));
+  border: 1px solid rgba(77, 184, 255, 0.34);
+  background:
+    radial-gradient(circle at 50% 18%, rgba(255, 255, 255, 0.12), transparent 35%),
+    linear-gradient(180deg, rgba(13, 24, 46, 0.95), rgba(7, 14, 28, 0.94));
+  box-shadow:
+    0 0 30px rgba(77, 184, 255, 0.22),
+    0 16px 40px rgba(0, 0, 0, 0.28);
   text-align: center;
-  background-color: rgba(10, 10, 30, 0.95);
-  padding: 30px;
-  border-radius: 10px;
-  border: 1px solid #4db8ff;
-  box-shadow: 0 0 30px rgba(77,184,255,0.3);
+  color: #4db8ff;
 }
 
-#loading-spinner {
-  border: 4px solid rgba(77, 184, 255, 0.3);
+.earth-loading-text {
+  color: #4db8ff;
+}
+
+.earth-loading-title {
+  font-size: calc(1.15rem * var(--hud-scale));
+  font-weight: 600;
+}
+
+.earth-loading-subtitle {
+  margin-top: calc(10px * var(--hud-scale));
+  color: #9ab7d4;
+  font-size: calc(0.84rem * var(--hud-scale));
+  line-height: 1.45;
+}
+
+.earth-loading-spinner {
+  width: calc(40px * var(--hud-scale));
+  height: calc(40px * var(--hud-scale));
+  margin: 0 auto calc(15px * var(--hud-scale));
+  border: 4px solid rgba(77, 184, 255, 0.28);
   border-top: 4px solid #4db8ff;
   border-radius: 50%;
-  width: 40px;
-  height: 40px;
   animation: spin 1s linear infinite;
-  margin: 0 auto 15px;
 }
 
 @keyframes spin {
-  0% { transform: rotate(0deg); }
-  100% { transform: rotate(360deg); }
-}
+  0% {
+    transform: rotate(0deg);
+  }
 
-.error-message {
-  color: #ff4444;
-  margin-top: 10px;
-  font-size: 0.9rem;
-  display: none;
-  padding: 10px;
-  background-color: rgba(255, 68, 68, 0.1);
-  border-radius: 5px;
-  border-left: 3px solid #ff4444;
-}
-
-.terrain-controls {
-  margin-top: 15px;
-  padding-top: 15px;
-  border-top: 1px solid rgba(255, 255, 255, 0.1);
-}
-
-.slider-container {
-  margin-bottom: 10px;
-}
-
-.slider-label {
-  display: flex;
-  justify-content: space-between;
-  margin-bottom: 5px;
-  font-size: 0.9rem;
-}
-
-input[type="range"] {
-  width: 100%;
-  height: 8px;
-  -webkit-appearance: none;
-  background: rgba(0, 102, 204, 0.3);
-  border-radius: 4px;
-  outline: none;
-}
-
-input[type="range"]::-webkit-slider-thumb {
-  -webkit-appearance: none;
-  width: 16px;
-  height: 16px;
-  border-radius: 50%;
-  background: #4db8ff;
-  cursor: pointer;
-  box-shadow: 0 0 10px #4db8ff;
-}
-
-.status-message {
-  position: absolute;
-  top: 20px;
-  left: 50%;
-  transform: translate(-50%, -18px);
-  background-color: rgba(10, 10, 30, 0.85);
-  border-radius: 10px;
-  padding: 10px 15px;
-  z-index: 210;
-  box-shadow: 0 0 20px rgba(0, 150, 255, 0.3);
-  border: 1px solid rgba(0, 150, 255, 0.2);
-  font-size: 0.9rem;
-  display: none;
-  backdrop-filter: blur(5px);
-  text-align: center;
-  min-width: 180px;
-  opacity: 0;
-  transition:
-    transform 0.28s ease,
-    opacity 0.28s ease;
-}
-
-.status-message.visible {
-  transform: translate(-50%, 0);
-  opacity: 1;
-}
-
-.status-message.success {
-  color: #44ff44;
-  border-left: 3px solid #44ff44;
-}
-
-.status-message.warning {
-  color: #ffff44;
-  border-left: 3px solid #ffff44;
-}
-
-.status-message.error {
-  color: #ff4444;
-  border-left: 3px solid #ff4444;
-}
-
-.tooltip {
-  position: absolute;
-  background-color: rgba(10, 10, 30, 0.95);
-  border: 1px solid #4db8ff;
-  border-radius: 5px;
-  padding: 5px 10px;
-  font-size: 0.8rem;
-  color: #fff;
-  pointer-events: none;
-  z-index: 100;
-  box-shadow: 0 0 10px rgba(77, 184, 255, 0.3);
-  display: none;
-  user-select: none;
-}
-
-/* Floating toolbar dock */
-#control-toolbar {
-  position: relative;
-  display: flex;
-  align-items: center;
-  justify-content: center;
-  gap: 0;
-  background: transparent;
-  border: none;
-  box-shadow: none;
-  padding: 0;
-}
-
-.toolbar-items {
-  display: flex;
-  gap: 10px;
-  align-items: center;
-  flex-wrap: nowrap;
-}
-
-.floating-popover-group {
-  position: relative;
-}
-
-.floating-popover-group::before {
-  content: '';
-  position: absolute;
-  left: 50%;
-  bottom: 100%;
-  transform: translateX(-50%);
-  width: 56px;
-  height: 16px;
-  background: transparent;
-}
-
-.floating-popover-group > .stack-toolbar {
-  position: absolute;
-  left: 50%;
-  top: auto;
-  right: auto;
-  bottom: calc(100% + 12px);
-  transform: translate(-50%, 10px);
-  display: flex;
-  flex-direction: column;
-  align-items: center;
-  gap: 8px;
-  opacity: 0;
-  visibility: hidden;
-  pointer-events: none;
-  transition:
-    opacity 0.22s ease,
-    transform 0.22s ease,
-    visibility 0.22s ease;
-  z-index: 220;
-}
-
-.toolbar-btn.floating-btn {
-  width: 42px;
-  height: 42px;
-  min-width: 42px;
-  min-height: 42px;
-  border-radius: 50%;
-  overflow: hidden;
-}
-
-.liquid-glass-surface {
-  --elastic-x: 0px;
-  --elastic-y: 0px;
-  --tilt-x: 0deg;
-  --tilt-y: 0deg;
-  --btn-scale: 1;
-  --press-offset: 0px;
-  --float-offset: 0px;
-  --glow-opacity: 0.24;
-  --glow-x: 50%;
-  --glow-y: 22%;
-  position: relative;
-  isolation: isolate;
-  transform-style: preserve-3d;
-  overflow: hidden;
-  background:
-    radial-gradient(circle at var(--glow-x) var(--glow-y), rgba(255, 255, 255, 0.16), transparent 34%),
-    radial-gradient(circle at 50% 118%, rgba(255, 255, 255, 0.08), transparent 30%),
-    linear-gradient(180deg, var(--glass-fill-top), var(--glass-fill-bottom)),
-    rgba(8, 20, 38, 0.22);
-  border: 1px solid var(--hud-border);
-  box-shadow:
-    inset 0 1px 0 rgba(255, 255, 255, 0.14),
-    inset 0 -1px 0 rgba(255, 255, 255, 0.05),
-    var(--glass-shadow),
-    var(--glass-glow);
-  backdrop-filter: blur(18px) saturate(145%);
-  -webkit-backdrop-filter: blur(18px) saturate(145%);
-  transform:
-    translate3d(var(--elastic-x), calc(var(--float-offset) + var(--press-offset) + var(--elastic-y)), 0)
-    scale(var(--btn-scale));
-  transition:
-    transform 0.22s ease,
-    box-shadow 0.22s ease,
-    background 0.22s ease,
-    opacity 0.18s ease,
-    border-color 0.22s ease;
-  animation: floatDock 3.8s ease-in-out infinite;
-}
-
-.liquid-glass-surface::before {
-  content: '';
-  position: absolute;
-  inset: 1px 1px 18px 1px;
-  border-radius: inherit;
-  background:
-    linear-gradient(180deg, rgba(255, 255, 255, 0.18), rgba(255, 255, 255, 0.05) 28%, transparent 68%);
-  opacity: 0.5;
-  pointer-events: none;
-  transform:
-    perspective(120px)
-    rotateX(calc(var(--tilt-x) * 0.7))
-    rotateY(calc(var(--tilt-y) * 0.7))
-    translate3d(calc(var(--elastic-x) * 0.22), calc(var(--elastic-y) * 0.22), 0);
-  transition: opacity 0.18s ease, transform 0.18s ease;
-}
-
-.liquid-glass-surface::after {
-  content: '';
-  position: absolute;
-  inset: -1px;
-  padding: 1.35px;
-  border-radius: inherit;
-  background:
-    linear-gradient(135deg, rgba(255, 255, 255, 0.36), rgba(168, 222, 255, 0.22) 34%, rgba(96, 175, 255, 0.16) 66%, rgba(255, 255, 255, 0.28));
-  opacity: 0.82;
-  pointer-events: none;
-  filter: url(#liquid-glass-distortion) blur(0.35px);
-  transform:
-    perspective(120px)
-    rotateX(calc(var(--tilt-x) * 0.5))
-    rotateY(calc(var(--tilt-y) * 0.5))
-    translate3d(calc(var(--elastic-x) * 0.16), calc(var(--elastic-y) * 0.16), 0);
-  -webkit-mask:
-    linear-gradient(#000 0 0) content-box,
-    linear-gradient(#000 0 0);
-  -webkit-mask-composite: xor;
-  mask:
-    linear-gradient(#000 0 0) content-box,
-    linear-gradient(#000 0 0);
-  mask-composite: exclude;
-  transition: opacity 0.18s ease, transform 0.18s ease;
-}
-
-.toolbar-items > :nth-child(2n).floating-btn,
-.toolbar-items > :nth-child(2n) .floating-btn {
-  animation-delay: 0.18s;
-}
-
-.toolbar-items > :nth-child(3n).floating-btn,
-.toolbar-items > :nth-child(3n) .floating-btn {
-  animation-delay: 0.34s;
+  100% {
+    transform: rotate(360deg);
+  }
 }
 
 @keyframes floatDock {
-  0%, 100% {
+  0%,
+  100% {
     --float-offset: 0px;
   }
+
   50% {
-    --float-offset: -4px;
+    --float-offset: -2px;
   }
 }
-
-.toolbar-btn {
-  position: relative;
-  width: 28px;
-  height: 28px;
-  border: none;
-  border-radius: 0;
-  background: transparent;
-  color: #4db8ff;
-  font-size: 14px;
-  cursor: pointer;
-  display: flex;
-  align-items: center;
-  justify-content: center;
-  box-sizing: border-box;
-  padding: 0;
-  margin: 0;
-  overflow: visible;
-  appearance: none;
-  -webkit-appearance: none;
-}
-
-.toolbar-btn:not(.liquid-glass-surface)::after {
-  content: none;
-}
-
-.toolbar-btn .icon {
-  display: inline-flex;
-  align-items: center;
-  justify-content: center;
-  position: relative;
-  z-index: 1;
-  transform: translateZ(0);
-  transition: transform 0.16s ease, opacity 0.16s ease;
-  backface-visibility: hidden;
-  -webkit-backface-visibility: hidden;
-  line-height: 1;
-}
-
-.liquid-glass-surface:hover {
-  --btn-scale: 1.035;
-  --press-offset: -1px;
-  --glow-opacity: 0.32;
-  background:
-    radial-gradient(circle at var(--glow-x) var(--glow-y), rgba(255, 255, 255, 0.18), transparent 34%),
-    radial-gradient(circle at 50% 118%, rgba(255, 255, 255, 0.1), transparent 30%),
-    linear-gradient(180deg, rgba(255, 255, 255, 0.18), rgba(128, 198, 255, 0.1)),
-    rgba(8, 20, 38, 0.2);
-  border-color: var(--hud-border-hover);
-  box-shadow:
-    inset 0 1px 0 rgba(255, 255, 255, 0.2),
-    inset 0 -1px 0 rgba(255, 255, 255, 0.08),
-    0 18px 36px rgba(0, 0, 0, 0.24),
-    0 0 28px rgba(145, 214, 255, 0.22);
-}
-
-.liquid-glass-surface:hover::before {
-  opacity: 0.62;
-  transform: scale(1.01);
-}
-
-.liquid-glass-surface:hover::after {
-  opacity: 0.96;
-  transform: scale(1.01);
-}
-
-.liquid-glass-surface:active,
-.liquid-glass-surface.is-pressed {
-  --btn-scale: 0.942;
-  --press-offset: 2px;
-  --glow-opacity: 0.2;
-  background:
-    radial-gradient(circle at var(--glow-x) var(--glow-y), rgba(255, 255, 255, 0.24), transparent 34%),
-    radial-gradient(circle at 50% 118%, rgba(255, 255, 255, 0.14), transparent 30%),
-    linear-gradient(180deg, rgba(255, 255, 255, 0.24), rgba(146, 210, 255, 0.16)),
-    rgba(10, 24, 44, 0.24);
-  border-color: rgba(240, 249, 255, 0.58);
-  box-shadow:
-    inset 0 2px 10px rgba(0, 0, 0, 0.2),
-    inset 0 1px 0 rgba(255, 255, 255, 0.16),
-    0 4px 10px rgba(0, 0, 0, 0.18),
-    0 0 14px rgba(176, 226, 255, 0.18);
-}
-
-.liquid-glass-surface:active::before,
-.liquid-glass-surface.is-pressed::before {
-  opacity: 0.46;
-  transform: translateY(2px) scale(0.985);
-}
-
-.liquid-glass-surface:active::after,
-.liquid-glass-surface.is-pressed::after {
-  opacity: 0.78;
-  transform: scale(0.985);
-}
-
-.liquid-glass-surface:active .icon,
-.liquid-glass-surface.is-pressed .icon {
-  transform: translateY(1.5px);
-}
-
-.liquid-glass-surface:active img,
-.liquid-glass-surface.is-pressed img,
-.liquid-glass-surface:active .material-symbols-rounded,
-.liquid-glass-surface.is-pressed .material-symbols-rounded {
-  transform: translateY(1.5px);
-  transition: transform 0.16s ease, opacity 0.16s ease;
-}
-
-#zoom-control-group #zoom-toolbar .zoom-btn:active,
-#zoom-control-group #zoom-toolbar .zoom-btn.is-pressed,
-#zoom-control-group #zoom-toolbar .zoom-percent:active,
-#zoom-control-group #zoom-toolbar .zoom-percent.is-pressed {
-  letter-spacing: -0.01em;
-}
-
-.liquid-glass-surface.active {
-  background:
-    radial-gradient(circle at var(--glow-x) var(--glow-y), rgba(255, 255, 255, 0.18), transparent 34%),
-    linear-gradient(180deg, rgba(255, 255, 255, 0.2), rgba(118, 200, 255, 0.14)),
-    rgba(11, 34, 58, 0.26);
-  border-color: var(--hud-border-active);
-  box-shadow:
-    inset 0 1px 0 rgba(255, 255, 255, 0.22),
-    inset 0 0 18px rgba(160, 220, 255, 0.14),
-    0 18px 34px rgba(0, 0, 0, 0.24),
-    0 0 30px rgba(145, 214, 255, 0.24);
-}
-
-.toolbar-btn svg {
-  width: 20px;
-  height: 20px;
-  stroke: currentColor;
-  stroke-width: 2.1;
-  fill: none;
-  stroke-linecap: round;
-  stroke-linejoin: round;
-}
-
-.toolbar-btn .material-symbols-rounded {
-  font-size: 21px;
-  line-height: 1;
-  font-variation-settings:
-    'FILL' 0,
-    'wght' 500,
-    'GRAD' 0,
-    'opsz' 24;
-  color: currentColor;
-  display: inline-flex;
-  align-items: center;
-  justify-content: center;
-  user-select: none;
-  pointer-events: none;
-  text-rendering: geometricPrecision;
-  -webkit-font-smoothing: antialiased;
-  -moz-osx-font-smoothing: grayscale;
-}
-
-.toolbar-btn img {
-  width: 20px;
-  height: 20px;
-  display: block;
-  user-select: none;
-  pointer-events: none;
-  shape-rendering: geometricPrecision;
-  image-rendering: -webkit-optimize-contrast;
-  backface-visibility: hidden;
-  -webkit-backface-visibility: hidden;
-}
-
-#rotate-toggle .icon-play,
-#rotate-toggle.is-stopped .icon-pause,
-#layout-toggle .layout-collapse,
-#layout-toggle.active .layout-expand {
-  display: none;
-}
-
-#rotate-toggle.is-stopped .icon-play,
-#layout-toggle.active .layout-collapse {
-  display: inline-flex;
-}
-
-#zoom-control-group:hover #zoom-toolbar,
-#zoom-control-group:focus-within #zoom-toolbar,
-#zoom-control-group.open #zoom-toolbar,
-#info-control-group:hover #info-toolbar,
-#info-control-group:focus-within #info-toolbar,
-#info-control-group.open #info-toolbar {
-  opacity: 1;
-  visibility: visible;
-  pointer-events: auto;
-  transform: translate(-50%, 0);
-}
-
-#zoom-control-group.force-closed #zoom-toolbar,
-#info-control-group.force-closed #info-toolbar {
-  opacity: 0;
-  visibility: hidden;
-  pointer-events: none;
-  transform: translate(-50%, 8px);
-}
-
-#zoom-control-group #zoom-toolbar .zoom-percent {
-  min-width: 0;
-  width: 42px;
-  display: inline-flex;
-  align-items: center;
-  justify-content: center;
-  height: 42px;
-  padding: 0;
-  font-size: 0.68rem;
-  border-radius: 50%;
-  color: #4db8ff;
-  animation: floatDock 3.8s ease-in-out infinite;
-  animation-delay: 0.18s;
-}
-
-#zoom-control-group #zoom-toolbar .zoom-percent:hover {
-}
-
-#zoom-control-group #zoom-toolbar,
-#info-control-group #info-toolbar {
-  top: auto;
-  right: auto;
-  left: 50%;
-  bottom: calc(100% + 12px);
-  display: flex;
-  flex-direction: column;
-  align-items: center;
-  justify-content: flex-start;
-  gap: 8px;
-}
-
-#info-toolbar .toolbar-btn:nth-child(1) {
-  animation-delay: 0.34s;
-}
-
-#info-toolbar .toolbar-btn:nth-child(2) {
-  animation-delay: 0.18s;
-}
-
-#info-toolbar .toolbar-btn:nth-child(3) {
-  animation-delay: 0.1s;
-}
-
-#info-toolbar .toolbar-btn:nth-child(4) {
-  animation-delay: 0s;
-}
-
-#zoom-control-group #zoom-toolbar .zoom-btn {
-  width: 42px;
-  height: 42px;
-  min-width: 42px;
-  border-radius: 50%;
-  color: #4db8ff;
-  animation: floatDock 3.8s ease-in-out infinite;
-}
-
-#zoom-toolbar .zoom-btn:nth-child(1) {
-  animation-delay: 0s;
-}
-
-#zoom-toolbar .zoom-btn:nth-child(3) {
-  animation-delay: 0.34s;
-}
-
-#zoom-control-group #zoom-toolbar .zoom-btn:hover {
-}
-
-#zoom-control-group #zoom-toolbar .zoom-btn:active,
-#zoom-control-group #zoom-toolbar .zoom-percent:active {
-}
-
-#zoom-control-group #zoom-toolbar .tooltip {
-  bottom: calc(100% + 10px);
-}
-
-#zoom-control-group #zoom-toolbar .tooltip::after {
-  top: 100%;
-  left: 50%;
-  transform: translateX(-50%);
-  border: 6px solid transparent;
-  border-top-color: rgba(77, 184, 255, 0.4);
-}
-
-#container.layout-expanded #info-panel {
-  top: 20px;
-  left: 20px;
-  transform: translate(calc(-100% + 20px), calc(-100% + 20px));
-}
-
-#container.layout-expanded #coordinates-display {
-  top: 20px;
-  right: 20px;
-  transform: translate(calc(100% - 20px), calc(-100% + 20px));
-}
-
-#container.layout-expanded #legend {
-  left: 20px;
-  bottom: 20px;
-  transform: translate(calc(-100% + 20px), calc(100% - 20px));
-}
-
-#container.layout-expanded #earth-stats {
-  right: 20px;
-  bottom: 20px;
-  transform: translate(calc(100% - 20px), calc(100% - 20px));
-}
-
-#container.layout-expanded #right-toolbar-group {
-  bottom: 18px;
-  transform: translateX(-50%);
-}
-
-.toolbar-btn .tooltip {
-  position: absolute;
-  bottom: 56px;
-  left: 50%;
-  transform: translateX(-50%);
-  background: rgba(10, 10, 30, 0.95);
-  color: #fff;
-  padding: 6px 12px;
-  border-radius: 6px;
-  font-size: 12px;
-  white-space: nowrap;
-  opacity: 0;
-  visibility: hidden;
-  transition: all 0.2s ease;
-  border: 1px solid rgba(77, 184, 255, 0.4);
-  pointer-events: none;
-  z-index: 100;
-}
-
-.toolbar-btn:hover .tooltip,
-.floating-popover-group:hover > .toolbar-btn .tooltip,
-.floating-popover-group:focus-within > .toolbar-btn .tooltip {
-  opacity: 1;
-  visibility: visible;
-  bottom: 58px;
-}
-
-.toolbar-btn .tooltip::after {
-  content: '';
-  position: absolute;
-  top: 100%;
-  left: 50%;
-  transform: translateX(-50%);
-  border: 6px solid transparent;
-  border-top-color: rgba(77, 184, 255, 0.4);
-}

frontend/public/earth/css/coordinates-display.css

@@ -1,42 +1,33 @@
 /* coordinates-display */
 
-#coordinates-display {
-  top: 20px;
-  right: 20px;
-  border-radius: 18px;
-  padding: 10px 15px;
+.hud-panel-coordinates {
+  top: var(--hud-offset);
+  right: var(--hud-offset);
+  border-radius: var(--hud-radius);
+  padding: calc(14px * var(--hud-scale)) var(--hud-panel-padding);
   z-index: 10;
-  font-size: 0.9rem;
-  min-width: 180px;
+  font-size: var(--hud-font-size);
+  min-width: calc(196px * var(--hud-scale));
 }
 
-#coordinates-display .coord-item {
-  margin-bottom: 5px;
-  display: flex;
-  justify-content: space-between;
+.hud-panel-coordinates .coord-item {
+  margin-bottom: calc(8px * var(--hud-scale));
 }
 
-#coordinates-display .coord-label {
-  color: #aaa;
-}
-
-#coordinates-display .coord-value {
-  color: #4db8ff;
-  font-weight: 500;
-}
-
-#coordinates-display #zoom-level {
-  margin-top: 5px;
-  color: #ffff44;
-  font-weight: 500;
+.hud-panel-coordinates .earth-zoom-level {
+  margin-top: calc(10px * var(--hud-scale));
+  padding-top: calc(10px * var(--hud-scale));
+  border-top: 1px solid var(--hud-line);
+  color: var(--hud-accent);
+  font-weight: 600;
   text-align: center;
-  font-size: 1rem;
+  font-size: calc(0.94rem * var(--hud-scale));
+  letter-spacing: 0.02em;
 }
 
-#coordinates-display .mouse-coords {
-  font-size: 0.8rem;
-  color: #aaa;
-  margin-top: 5px;
-  padding-top: 5px;
-  border-top: 1px solid rgba(255, 255, 255, 0.1);
+.hud-panel-coordinates .earth-mouse-coords {
+  font-size: var(--hud-font-size-sm);
+  color: var(--hud-text-muted);
+  margin-top: calc(8px * var(--hud-scale));
+  line-height: 1.45;
 }

frontend/public/earth/css/earth-stats.css

@@ -1,51 +1,133 @@
-/* earth-stats */
+/* earth-stats.css — compact KPI grid panel */
 
-#earth-stats {
-  bottom: 20px;
-  right: 20px;
-  border-radius: 18px;
-  padding: 15px;
-  width: 250px;
+.hud-panel-stats {
+  top: var(--hud-offset);
+  right: var(--hud-offset);
+  border-radius: 0;           /* square / angular — matches layer panel */
+  padding: 0;
+  width: min(calc(240px * var(--hud-scale)), calc(100vw - 32px));
   z-index: 10;
-  font-size: 0.9rem;
+  overflow: hidden;
 }
 
-#earth-stats .stats-item {
-  margin-bottom: 8px;
+/* ── Thin drag bar ────────────────────────────────────────────── */
+
+.stats-drag-bar {
   display: flex;
+  align-items: center;
   justify-content: space-between;
+  padding: calc(8px * var(--hud-scale)) calc(12px * var(--hud-scale));
+  border-bottom: 1px solid var(--hud-line);
+  cursor: grab;
+  user-select: none;
 }
 
-#earth-stats .stats-label {
-  color: #aaa;
+.stats-drag-bar:active {
+  cursor: grabbing;
 }
 
-#earth-stats .stats-value {
-  color: #4db8ff;
-  font-weight: 500;
+.stats-kicker {
+  color: var(--hud-text-soft);
+  font-size: calc(0.64rem * var(--hud-scale));
+  letter-spacing: 0.16em;
+  text-transform: uppercase;
 }
 
-#satellite-info {
-  bottom: 20px;
-  right: 290px;
-  border-radius: 18px;
-  padding: 15px;
-  width: 220px;
-  z-index: 10;
-  font-size: 0.85rem;
+/* Reuse hud-panel-close — just override size to match kicker line */
+.stats-drag-bar .hud-panel-close {
+  width: calc(20px * var(--hud-scale));
+  height: calc(20px * var(--hud-scale));
+  min-width: calc(20px * var(--hud-scale));
 }
 
-#satellite-info .stats-item {
-  margin-bottom: 6px;
+.stats-drag-bar .hud-panel-close .material-symbols-rounded {
+  font-size: calc(12px * var(--hud-scale));
+}
+
+/* ── 2-column KPI grid ────────────────────────────────────────── */
+
+.stats-grid {
+  display: grid;
+  grid-template-columns: 1fr 1fr;
+}
+
+.stat-cell {
   display: flex;
-  justify-content: space-between;
+  flex-direction: column;
+  gap: calc(2px * var(--hud-scale));
+  padding: calc(10px * var(--hud-scale)) calc(12px * var(--hud-scale));
+  border-right: 1px solid var(--hud-line);
+  border-bottom: 1px solid var(--hud-line);
 }
 
-#satellite-info .stats-label {
-  color: #aaa;
+/* Right column cells: remove right border */
+.stat-cell:nth-child(even) {
+  border-right: none;
 }
 
-#satellite-info .stats-value {
-  color: #00e5ff;
-  font-weight: 500;
+/* Bottom row cells: remove bottom border */
+.stat-cell:nth-last-child(-n+2) {
+  border-bottom: none;
+}
+
+.stat-num {
+  color: var(--hud-title);
+  font-size: calc(1.3rem * var(--hud-scale));
+  font-weight: 700;
+  letter-spacing: -0.01em;
+  line-height: 1.1;
+  font-variant-numeric: tabular-nums;
+}
+
+/* Smaller number for multi-word values like "256/314" */
+.stat-num--sm {
+  font-size: calc(0.94rem * var(--hud-scale));
+  font-weight: 600;
+  letter-spacing: 0;
+}
+
+.stat-label {
+  color: var(--hud-text-soft);
+  font-size: calc(0.62rem * var(--hud-scale));
+  letter-spacing: 0.1em;
+  text-transform: uppercase;
+  line-height: 1.2;
+}
+
+/* ── BGP status footer ────────────────────────────────────────── */
+
+.stats-footer {
+  display: flex;
+  align-items: center;
+  gap: calc(6px * var(--hud-scale));
+  padding: calc(7px * var(--hud-scale)) calc(12px * var(--hud-scale));
+  border-top: 1px solid var(--hud-line);
+}
+
+.stats-footer-dot {
+  flex-shrink: 0;
+  width: calc(5px * var(--hud-scale));
+  height: calc(5px * var(--hud-scale));
+  border-radius: 50%;
+  background: var(--hud-accent);
+  box-shadow: 0 0 5px var(--hud-accent);
+}
+
+.stats-footer-text {
+  color: var(--hud-text-muted);
+  font-size: calc(0.7rem * var(--hud-scale));
+  line-height: 1.3;
+  flex: 1 1 auto;
+  min-width: 0;
+  overflow: hidden;
+  text-overflow: ellipsis;
+  white-space: nowrap;
+}
+
+/* ── Layout-expanded: slide off-screen ───────────────────────── */
+
+.earth-app.layout-expanded .hud-panel-stats:not([data-dragged="true"]) {
+  top: var(--hud-offset);
+  right: var(--hud-offset);
+  transform: translate(calc(100% - var(--hud-offset)), calc(-100% + var(--hud-offset)));
 }

frontend/public/earth/css/hud.css

@@ -0,0 +1,459 @@
+/* hud.css - HUD surfaces and shared overlays */
+
+.hud-panel {
+  position: absolute;
+  overflow: hidden;
+  isolation: isolate;
+  background:
+    radial-gradient(circle at 18% 0%, rgba(255, 255, 255, 0.08), transparent 30%),
+    radial-gradient(circle at 86% 115%, rgba(145, 186, 255, 0.08), transparent 36%),
+    linear-gradient(180deg, rgba(255, 255, 255, 0.05), transparent 26%),
+    linear-gradient(180deg, var(--hud-surface-top), var(--hud-surface-bottom));
+  border: 1px solid var(--hud-border);
+  box-shadow:
+    inset 0 1px 0 var(--hud-highlight),
+    inset 0 -1px 0 rgba(255, 255, 255, 0.03),
+    var(--hud-shadow),
+    0 0 0 1px rgba(255, 255, 255, 0.02);
+  backdrop-filter: blur(18px) saturate(125%);
+  -webkit-backdrop-filter: blur(18px) saturate(125%);
+}
+
+.hud-panel::before {
+  content: '';
+  position: absolute;
+  inset: 1px 1px 52% 1px;
+  border-radius: inherit;
+  background:
+    linear-gradient(180deg, rgba(255, 255, 255, 0.12), rgba(255, 255, 255, 0.02) 58%, transparent 100%);
+  opacity: 0.52;
+  pointer-events: none;
+}
+
+.hud-panel::after {
+  content: '';
+  position: absolute;
+  inset: -1px;
+  padding: 1px;
+  border-radius: inherit;
+  background:
+    linear-gradient(135deg, rgba(244, 249, 255, 0.22), rgba(164, 194, 226, 0.08) 36%, rgba(90, 123, 161, 0.04) 70%, rgba(255, 255, 255, 0.16));
+  opacity: 0.72;
+  pointer-events: none;
+  filter: blur(0.2px);
+  -webkit-mask:
+    linear-gradient(#000 0 0) content-box,
+    linear-gradient(#000 0 0);
+  -webkit-mask-composite: xor;
+  mask:
+    linear-gradient(#000 0 0) content-box,
+    linear-gradient(#000 0 0);
+  mask-composite: exclude;
+}
+
+.hud-panel > * {
+  position: relative;
+  z-index: 1;
+}
+
+.hud-panel-title {
+  color: var(--hud-title);
+  margin: 0 0 var(--hud-gap-sm);
+  font-size: var(--hud-title-size);
+  font-weight: 600;
+  letter-spacing: 0.01em;
+  line-height: 1.2;
+}
+
+.hud-panel-header {
+  display: flex;
+  align-items: flex-start;
+  justify-content: space-between;
+  gap: var(--hud-gap-sm);
+  margin-bottom: var(--hud-gap-sm);
+  padding-bottom: var(--hud-gap-sm);
+  border-bottom: 1px solid var(--hud-line);
+}
+
+.hud-panel-header .hud-panel-title {
+  margin-bottom: 0;
+}
+
+.hud-panel-drag-handle {
+  cursor: grab;
+  user-select: none;
+}
+
+.hud-panel-drag-handle:active {
+  cursor: grabbing;
+}
+
+.hud-panel-close {
+  align-self: flex-start;
+  width: calc(var(--hud-title-size) * 1.24);
+  height: calc(var(--hud-title-size) * 1.24);
+  min-width: calc(var(--hud-title-size) * 1.24);
+  padding: 0;
+  border: 1px solid transparent;
+  border-radius: calc(4px * var(--hud-scale));
+  background: transparent;
+  color: var(--hud-text-muted);
+  display: inline-flex;
+  align-items: center;
+  justify-content: center;
+  cursor: pointer;
+  transition:
+    background 0.18s ease,
+    border-color 0.18s ease,
+    color 0.18s ease,
+    transform 0.18s ease,
+    opacity 0.18s ease;
+}
+
+.hud-panel-close .material-symbols-rounded {
+  font-size: calc(var(--hud-title-size) * 0.8);
+  line-height: 1;
+}
+
+.hud-panel-close:hover {
+  background: rgba(255, 255, 255, 0.08);
+  border-color: rgba(225, 239, 255, 0.14);
+  color: var(--hud-accent-strong);
+}
+
+.hud-panel.is-dragging {
+  transition: none !important;
+  box-shadow:
+    inset 0 1px 0 rgba(255, 255, 255, 0.14),
+    inset 0 -1px 0 rgba(255, 255, 255, 0.04),
+    0 24px 52px rgba(0, 0, 0, 0.28),
+    0 0 36px rgba(120, 200, 255, 0.16);
+}
+
+.hud-panel.is-layout-animating {
+  transition: none !important;
+}
+
+.hud-panel-hidden {
+  display: none !important;
+}
+
+.hud-panel-row {
+  display: flex;
+  justify-content: space-between;
+  gap: var(--hud-gap-sm);
+  align-items: baseline;
+}
+
+.hud-panel-label {
+  color: var(--hud-text-soft);
+  font-size: var(--hud-font-size-sm);
+  letter-spacing: 0.08em;
+  text-transform: uppercase;
+}
+
+.hud-panel-value {
+  color: var(--hud-text);
+  font-weight: 600;
+}
+
+.hud-error-message {
+  color: #ff4444;
+  margin-top: 10px;
+  font-size: 0.9rem;
+  display: none;
+  padding: 10px;
+  background-color: rgba(255, 68, 68, 0.1);
+  border-radius: 5px;
+  border-left: 3px solid #ff4444;
+}
+
+.earth-status-message {
+  position: absolute;
+  top: 20px;
+  left: 50%;
+  transform: translate(-50%, -18px);
+  background:
+    linear-gradient(180deg, rgba(18, 31, 52, 0.92), rgba(8, 18, 32, 0.9));
+  border-radius: 14px;
+  padding: 11px 15px;
+  z-index: 210;
+  box-shadow: var(--hud-shadow-soft);
+  border: 1px solid var(--hud-border);
+  font-size: 0.9rem;
+  display: none;
+  backdrop-filter: blur(10px);
+  text-align: center;
+  min-width: 180px;
+  opacity: 0;
+  transition:
+    transform 0.28s ease,
+    opacity 0.28s ease;
+}
+
+.earth-status-message.visible {
+  transform: translate(-50%, 0);
+  opacity: 1;
+}
+
+.earth-status-message.success {
+  color: #d8f7df;
+  border-left: 3px solid #66d18f;
+}
+
+.earth-status-message.warning {
+  color: #fff2c3;
+  border-left: 3px solid #e4c464;
+}
+
+.earth-status-message.error {
+  color: #ffd4d7;
+  border-left: 3px solid #ff7b86;
+}
+
+.earth-tooltip {
+  position: absolute;
+  background:
+    linear-gradient(180deg, rgba(21, 37, 61, 0.96), rgba(8, 18, 31, 0.95));
+  border: 1px solid rgba(214, 230, 247, 0.14);
+  border-radius: 10px;
+  padding: 8px 12px;
+  font-size: 0.78rem;
+  color: var(--hud-text);
+  pointer-events: none;
+  z-index: 100;
+  box-shadow: var(--hud-shadow-soft);
+  display: none;
+  user-select: none;
+}
+
+.earth-settings-modal {
+  position: fixed;
+  inset: 0;
+  z-index: 260;
+  display: none;
+}
+
+.earth-settings-modal.is-open {
+  display: block;
+}
+
+.earth-settings-backdrop {
+  position: fixed;
+  inset: 0;
+  background: rgba(2, 8, 20, 0.46);
+  backdrop-filter: blur(14px);
+  -webkit-backdrop-filter: blur(14px);
+}
+
+.earth-settings-sheet {
+  position: fixed;
+  top: max(32px, 9vh);
+  right: 16px;
+  left: 16px;
+  width: min(560px, calc(100vw - 32px));
+  max-width: 560px;
+  max-height: calc(100vh - max(64px, 18vh));
+  margin-inline: auto;
+  transform: none;
+  border-radius: calc(24px * var(--hud-scale));
+  padding: calc(20px * var(--hud-scale));
+  display: flex;
+  flex-direction: column;
+  gap: var(--hud-gap-md);
+  overflow: hidden;
+}
+
+.earth-settings-sheet.liquid-glass-surface {
+  animation: none;
+  background:
+    radial-gradient(circle at 18% 0%, rgba(255, 255, 255, 0.09), transparent 30%),
+    linear-gradient(180deg, rgba(255, 255, 255, 0.05), transparent 28%),
+    linear-gradient(180deg, rgba(19, 34, 56, 0.92), rgba(8, 18, 31, 0.9));
+  border-color: rgba(207, 224, 243, 0.12);
+  box-shadow:
+    inset 0 1px 0 rgba(255, 255, 255, 0.08),
+    0 24px 56px rgba(2, 7, 15, 0.4),
+    0 0 0 1px rgba(255, 255, 255, 0.025);
+}
+
+.earth-settings-sheet.liquid-glass-surface:hover,
+.earth-settings-sheet.liquid-glass-surface:active,
+.earth-settings-sheet.liquid-glass-surface.is-pressed {
+  --btn-scale: 1;
+  --press-offset: 0px;
+  --glow-opacity: 0.24;
+  transform: none;
+}
+
+.earth-settings-header,
+.earth-settings-content {
+  position: relative;
+  z-index: 1;
+}
+
+.earth-settings-header {
+  display: flex;
+  align-items: flex-start;
+  justify-content: space-between;
+  gap: var(--hud-gap-md);
+  padding-bottom: var(--hud-gap-sm);
+  border-bottom: 1px solid var(--hud-line);
+}
+
+.earth-settings-kicker {
+  color: var(--hud-text-soft);
+  font-size: var(--hud-kicker-size);
+  letter-spacing: 0.16em;
+  text-transform: uppercase;
+}
+
+.earth-settings-title {
+  margin: 4px 0 0;
+  color: var(--hud-title);
+}
+
+.earth-settings-close {
+  margin-top: 4px;
+}
+
+.earth-settings-content {
+  overflow-y: auto;
+  padding-right: 4px;
+  scrollbar-width: thin;
+  scrollbar-color: rgba(160, 186, 216, 0.34) transparent;
+}
+
+.earth-settings-content::-webkit-scrollbar {
+  width: 6px;
+}
+
+.earth-settings-content::-webkit-scrollbar-track {
+  background: transparent;
+}
+
+.earth-settings-content::-webkit-scrollbar-thumb {
+  background: linear-gradient(180deg, rgba(210, 225, 242, 0.2), rgba(126, 154, 185, 0.28));
+  border-radius: 999px;
+}
+
+.earth-settings-section {
+  padding: 12px 0 20px;
+}
+
+.earth-settings-section-title {
+  margin-bottom: 12px;
+  color: var(--hud-text-soft);
+  font-size: var(--hud-kicker-size);
+  letter-spacing: 0.16em;
+  text-transform: uppercase;
+}
+
+.earth-settings-list {
+  display: flex;
+  flex-direction: column;
+  gap: 10px;
+}
+
+.earth-settings-item {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  gap: 16px;
+  padding: 15px 16px;
+  border-radius: 16px;
+  background:
+    linear-gradient(180deg, rgba(255, 255, 255, 0.04), transparent),
+    rgba(255, 255, 255, 0.025);
+  border: 1px solid rgba(212, 227, 244, 0.08);
+  box-shadow: inset 0 1px 0 rgba(255, 255, 255, 0.04);
+  cursor: pointer;
+  transition: background 0.18s ease, border-color 0.18s ease, transform 0.18s ease;
+}
+
+.earth-settings-item:hover {
+  background:
+    linear-gradient(180deg, rgba(255, 255, 255, 0.06), transparent),
+    rgba(255, 255, 255, 0.04);
+  border-color: rgba(224, 236, 249, 0.14);
+  transform: translateY(-1px);
+}
+
+.earth-settings-copy {
+  display: flex;
+  flex-direction: column;
+  gap: 4px;
+}
+
+.earth-settings-item-title {
+  color: var(--hud-text);
+  font-size: calc(0.98rem * var(--hud-scale));
+  font-weight: 600;
+}
+
+.earth-settings-item-subtitle {
+  color: var(--hud-text-muted);
+  font-size: 0.82rem;
+  line-height: 1.4;
+}
+
+.earth-settings-switch {
+  position: relative;
+  display: inline-flex;
+  align-items: center;
+}
+
+.earth-settings-switch input {
+  position: absolute;
+  opacity: 0;
+  pointer-events: none;
+}
+
+.earth-settings-switch-track {
+  width: 48px;
+  height: 30px;
+  border-radius: 999px;
+  background: rgba(255, 255, 255, 0.08);
+  border: 1px solid rgba(215, 229, 242, 0.12);
+  position: relative;
+  transition: background 0.18s ease, border-color 0.18s ease;
+}
+
+.earth-settings-switch-track::after {
+  content: "";
+  position: absolute;
+  top: 3px;
+  left: 3px;
+  width: 22px;
+  height: 22px;
+  border-radius: 50%;
+  background: #edf4fc;
+  box-shadow: 0 6px 14px rgba(1, 8, 18, 0.26);
+  transition: transform 0.18s ease;
+}
+
+.earth-settings-switch input:checked + .earth-settings-switch-track {
+  background: linear-gradient(180deg, rgba(143, 185, 255, 0.72), rgba(104, 147, 221, 0.78));
+  border-color: rgba(223, 236, 252, 0.28);
+}
+
+.earth-settings-switch input:checked + .earth-settings-switch-track::after {
+  transform: translateX(18px);
+}
+
+@media (max-width: 960px) {
+  .earth-settings-sheet {
+    top: 24px;
+    right: 12px;
+    left: 12px;
+    width: auto;
+    max-width: none;
+    max-height: calc(100vh - 48px);
+  }
+}
+
+/* .earth-left-column layout-expanded rule lives in info-panel.css */
+/* .hud-panel-legend layout-expanded rule lives in legend.css */
+/* .hud-panel-stats layout-expanded rule lives in earth-stats.css */
+/* .hud-panel-layers layout-expanded rule lives in layer-panel.css */
+/* .hud-panel-tv layout-expanded rule lives in tv-panel.css */

frontend/public/earth/css/info-panel.css

@@ -1,248 +1,202 @@
-/* info-panel */
+/* info-panel.css — brand panel + detail card */
 
-#info-panel {
-  top: 20px;
-  left: 20px;
-  border-radius: 18px;
-  padding: 20px;
-  width: 320px;
-  z-index: 10;
-}
+/* ── Left column wrapper ──────────────────────────────────────── */
 
-#info-panel h1 {
-  font-size: 1.8rem;
-  margin-bottom: 5px;
-  color: #4db8ff;
-  text-shadow: 0 0 10px rgba(77, 184, 255, 0.5);
-  text-align: center;
-}
-
-#info-panel .subtitle {
-  margin-bottom: 20px;
-  border-bottom: 1px solid rgba(255,255,255,0.1);
-  padding-bottom: 12px;
-  text-align: center;
+.earth-left-column {
+  position: absolute;
+  top: var(--hud-offset);
+  left: var(--hud-offset);
   display: flex;
   flex-direction: column;
+  z-index: 10;
+  pointer-events: none;
+  /* Width is set by the widest child (brand or info card) */
+  max-width: min(calc(340px * var(--hud-scale)), calc(100vw - 32px));
+}
+
+.earth-left-column > * {
+  position: relative;   /* override .hud-panel position:absolute */
+  pointer-events: auto;
+  width: 100%;
+  box-sizing: border-box;
+}
+
+/* ── Brand panel ──────────────────────────────────────────────── */
+
+.hud-panel-brand {
+  border-radius: 0;
+  padding: calc(12px * var(--hud-scale)) calc(14px * var(--hud-scale));
+  display: flex;
   align-items: center;
-  gap: 4px;
+  justify-content: center;
+  /* Reserve full panel height before brand images load */
+  min-height: calc(66px * var(--hud-scale));
 }
 
-#info-panel .subtitle-main {
-  color: #d7e7f5;
-  font-size: 0.95rem;
-  line-height: 1.35;
+.hud-panel-brand .earth-brand {
+  display: flex;
+  align-items: center;
+  gap: calc(10px * var(--hud-scale));
+  min-width: 0;
+}
+
+.hud-panel-brand .earth-brand__logo {
+  display: block;
+  flex: 0 0 auto;
+  width: calc(128px * var(--hud-scale));
+  height: calc(128px * var(--hud-scale));
+  object-fit: contain;
+}
+
+.hud-panel-brand .earth-brand__copy {
+  display: flex;
+  flex: 1 1 auto;
+  min-width: 0;
+  flex-direction: column;
+  justify-content: center;
+  gap: calc(5px * var(--hud-scale));
+}
+
+.hud-panel-brand .earth-brand__title {
+  display: block;
+  width: min(100%, calc(160px * var(--hud-scale)));
+  max-width: 100%;
+  height: auto;
+  min-height: calc(20px * var(--hud-scale));
+  object-fit: contain;
+}
+
+.hud-panel-brand .earth-brand__meta {
+  display: flex;
+  flex-direction: column;
+  gap: calc(2px * var(--hud-scale));
+  width: fit-content;
+}
+
+.hud-panel-brand .earth-brand__subtitle {
+  color: var(--hud-text-muted);
+  font-size: calc(0.74rem * var(--hud-scale));
+  line-height: 1.3;
   font-weight: 500;
-  letter-spacing: 0.02em;
+  letter-spacing: 0.01em;
+  /* Prevent text from pushing brand wider than logo column */
+  width: fit-content;
+  max-width: 100%;
+  overflow: hidden;
+  white-space: nowrap;
+  text-overflow: ellipsis;
 }
 
-#info-panel .subtitle-meta {
-  color: #8ea5bc;
-  font-size: 0.74rem;
+.hud-panel-brand .earth-brand__description {
+  color: var(--hud-text-soft);
+  font-size: calc(0.6rem * var(--hud-scale));
   line-height: 1.3;
   letter-spacing: 0.08em;
-  text-transform: uppercase;
+  width: fit-content;
+  max-width: 100%;
+  overflow: hidden;
+  white-space: nowrap;
+  text-overflow: ellipsis;
 }
 
-#info-panel .cable-info {
-  margin-top: 15px;
-  padding-top: 15px;
-  border-top: 1px solid rgba(255, 255, 255, 0.1);
+.hud-panel-brand .earth-brand--en .earth-brand__title {
+  width: min(100%, calc(172px * var(--hud-scale)));
 }
 
-#info-panel .cable-info h3 {
-  color: #4db8ff;
-  margin-bottom: 8px;
-  font-size: 1.2rem;
+.hud-panel-brand .earth-brand--en .earth-brand__subtitle,
+.hud-panel-brand .earth-brand--en .earth-brand__description {
+  font-family: "Roboto Condensed", "Arial Narrow", "Trebuchet MS", "Segoe UI", Tahoma, Geneva, Verdana, sans-serif;
 }
 
-#info-panel .cable-property {
-  display: flex;
-  justify-content: space-between;
-  margin-bottom: 5px;
-  font-size: 0.9rem;
-}
+/* ── Info detail panel (floating, positioned near click by JS) ── */
 
-#info-panel .property-label {
-  color: #aaa;
-}
-
-#info-panel .property-value {
-  color: #fff;
-  font-weight: 500;
-}
-
-#info-panel .controls {
-  display: flex;
-  justify-content: space-between;
-  margin-top: 20px;
-  flex-wrap: wrap;
-  gap: 10px;
-}
-
-#info-panel button {
-  background: linear-gradient(135deg, #0066cc, #004c99);
-  color: white;
-  border: none;
-  padding: 8px 15px;
-  border-radius: 5px;
-  cursor: pointer;
-  font-size: 0.9rem;
-  transition: all 0.3s;
-  flex: 1;
-  min-width: 120px;
-  box-shadow: 0 2px 5px rgba(0,0,0,0.3);
-}
-
-#info-panel button:hover {
-  background: linear-gradient(135deg, #0088ff, #0066cc);
-  transform: translateY(-2px);
-  box-shadow: 0 5px 15px rgba(0,102,204,0.4);
-}
-
-#info-panel .zoom-controls {
-  display: flex;
-  align-items: center;
-  margin-top: 15px;
-  padding-top: 15px;
-  border-top: 1px solid rgba(255, 255, 255, 0.1);
-}
-
-#info-panel .zoom-buttons {
-  display: flex;
-  align-items: center;
-  justify-content: center;
-  gap: 15px;
-  margin-top: 10px;
-  width: 100%;
-}
-
-#info-panel .zoom-percent-container {
-  display: flex;
-  align-items: center;
-  justify-content: center;
-  gap: 15px;
-}
-
-#info-panel .zoom-percent {
-  font-size: 1.4rem;
-  font-weight: 600;
-  color: #4db8ff;
-  min-width: 70px;
-  text-align: center;
-  cursor: pointer;
-  padding: 5px 10px;
-  border-radius: 5px;
-  transition: all 0.2s ease;
-}
-
-#info-panel .zoom-percent:hover {
-  background: rgba(77, 184, 255, 0.2);
-  box-shadow: 0 0 10px rgba(77, 184, 255, 0.3);
-}
-
-#info-panel .zoom-buttons .zoom-btn {
-  width: 36px;
-  height: 36px;
-  min-width: 36px;
-  border: none;
-  border-radius: 50%;
-  background: rgba(77, 184, 255, 0.2);
-  color: #4db8ff;
-  font-size: 22px;
-  font-weight: bold;
-  cursor: pointer;
-  display: flex;
-  align-items: center;
-  justify-content: center;
-  transition: all 0.2s ease;
-  padding: 0;
-  flex: 0 0 auto;
-}
-
-#info-panel .zoom-buttons .zoom-btn:hover {
-  background: rgba(77, 184, 255, 0.4);
-  transform: scale(1.1);
-  box-shadow: 0 0 10px rgba(77, 184, 255, 0.5);
-}
-
-#info-panel .zoom-buttons button {
-  flex: 1;
-  min-width: 60px;
-}
-
-/* Info Card - Unified details panel (inside info-panel) */
-.info-card {
-  margin-top: 15px;
-  background:
-    linear-gradient(180deg, rgba(255, 255, 255, 0.08), rgba(110, 176, 255, 0.04)),
-    rgba(7, 18, 36, 0.2);
-  border-radius: 14px;
-  border: 1px solid rgba(225, 242, 255, 0.12);
-  box-shadow:
-    inset 0 1px 0 rgba(255, 255, 255, 0.08),
-    0 10px 24px rgba(0, 0, 0, 0.16);
+.hud-panel-info {
+  position: absolute;
+  z-index: 50;
+  width: min(calc(300px * var(--hud-scale)), calc(100vw - 32px));
+  border-radius: 0;
   padding: 0;
   overflow: hidden;
+  /* Start hidden */
+  opacity: 0;
+  transform: scale(0.94) translateY(4px);
+  pointer-events: none;
+  transition:
+    opacity 0.22s ease,
+    transform 0.22s ease;
+}
+
+.hud-panel-info.is-visible {
+  opacity: 1;
+  transform: scale(1) translateY(0);
   pointer-events: auto;
 }
 
-.info-card.no-border {
-  background: transparent;
-  border: none;
+/* ── Info Card ────────────────────────────────────────────────── */
+
+.info-card {
+  display: block;
 }
 
 .info-card-header {
   display: flex;
   align-items: center;
-  padding: 10px 12px;
-  background: linear-gradient(180deg, rgba(255, 255, 255, 0.09), rgba(77, 184, 255, 0.06));
-  gap: 8px;
+  padding: calc(10px * var(--hud-scale)) calc(12px * var(--hud-scale));
+  background: linear-gradient(180deg, rgba(255, 255, 255, 0.05), rgba(255, 255, 255, 0.015));
+  border-bottom: 1px solid var(--hud-line);
+  gap: var(--hud-gap-xs);
 }
 
 .info-card-icon {
-  font-size: 18px;
+  font-size: calc(16px * var(--hud-scale));
+  flex-shrink: 0;
 }
 
 .info-card-header h3 {
   flex: 1;
   margin: 0;
-  font-size: 1rem;
-  color: #4db8ff;
+  font-size: calc(0.92rem * var(--hud-scale));
+  color: var(--hud-title);
+  font-weight: 600;
+  white-space: nowrap;
+  overflow: hidden;
+  text-overflow: ellipsis;
 }
 
-#info-card-content {
-  padding: 10px 12px;
-  max-height: 40vh;
+.info-card-close {
+  flex-shrink: 0;
+}
+
+.info-card-content {
+  padding: calc(8px * var(--hud-scale)) calc(12px * var(--hud-scale));
+  max-height: 58vh;
   overflow-y: auto;
   scrollbar-width: thin;
-  scrollbar-color: rgba(160, 220, 255, 0.45) transparent;
+  scrollbar-color: rgba(160, 186, 216, 0.34) transparent;
   pointer-events: auto;
 }
 
-#info-card-content::-webkit-scrollbar {
-  width: 6px;
+.info-card-content::-webkit-scrollbar {
+  width: 4px;
 }
 
-#info-card-content::-webkit-scrollbar-track {
+.info-card-content::-webkit-scrollbar-track {
   background: transparent;
 }
 
-#info-card-content::-webkit-scrollbar-thumb {
-  background: linear-gradient(180deg, rgba(210, 237, 255, 0.32), rgba(110, 176, 255, 0.34));
+.info-card-content::-webkit-scrollbar-thumb {
+  background: linear-gradient(180deg, rgba(210, 225, 242, 0.2), rgba(126, 154, 185, 0.28));
   border-radius: 999px;
 }
 
-#info-card-content::-webkit-scrollbar-thumb:hover {
-  background: linear-gradient(180deg, rgba(232, 246, 255, 0.42), rgba(128, 198, 255, 0.46));
-}
-
 .info-card-property {
   display: flex;
   justify-content: space-between;
-  padding: 6px;
-  border-bottom: 1px solid rgba(255, 255, 255, 0.05);
+  align-items: flex-start;
+  padding: calc(6px * var(--hud-scale)) 0;
+  border-bottom: 1px solid rgba(214, 229, 245, 0.06);
   pointer-events: auto;
+  gap: var(--hud-gap-sm);
 }
 
 .info-card-property:last-child {
@@ -250,47 +204,50 @@
 }
 
 .info-card-label {
-  color: #aaa;
-  font-size: 0.85rem;
+  color: var(--hud-text-soft);
+  font-size: calc(0.68rem * var(--hud-scale));
+  letter-spacing: 0.1em;
+  text-transform: uppercase;
   cursor: pointer;
+  flex-shrink: 0;
   transition: color 0.18s ease;
 }
 
 .info-card-label:hover {
-  color: #d9f1ff;
+  color: var(--hud-text-muted);
 }
 
 .info-card-value {
-  color: #4db8ff;
-  font-weight: 500;
-  font-size: 0.9rem;
+  color: var(--hud-text);
+  font-weight: 600;
+  font-size: calc(0.82rem * var(--hud-scale));
+  line-height: 1.45;
   text-align: right;
-  max-width: 180px;
+  max-width: calc(180px * var(--hud-scale));
   word-break: break-word;
 }
 
-/* Cable type */
-.info-card.cable {
-  border-color: rgba(255, 200, 0, 0.4);
-}
-
+/* Type-specific header accent colors */
 .info-card.cable .info-card-header {
-  background: rgba(255, 200, 0, 0.15);
-}
-
-.info-card.cable .info-card-header h3 {
-  color: #ffc800;
-}
-
-/* Satellite type */
-.info-card.satellite {
-  border-color: rgba(0, 229, 255, 0.4);
+  background: rgba(255, 200, 0, 0.12);
+  border-bottom-color: rgba(255, 200, 0, 0.15);
 }
+.info-card.cable .info-card-header h3 { color: #ffc800; }
 
 .info-card.satellite .info-card-header {
-  background: rgba(0, 229, 255, 0.15);
+  background: rgba(0, 229, 255, 0.12);
+  border-bottom-color: rgba(0, 229, 255, 0.15);
 }
+.info-card.satellite .info-card-header h3 { color: #00e5ff; }
 
-.info-card.satellite .info-card-header h3 {
-  color: #00e5ff;
+.info-card.bgp .info-card-header {
+  background: rgba(120, 180, 255, 0.12);
+  border-bottom-color: rgba(120, 180, 255, 0.15);
+}
+.info-card.bgp .info-card-header h3 { color: var(--hud-accent-strong); }
+
+/* ── Layout-expanded: slide left column off-screen ────────────── */
+
+.earth-app.layout-expanded .earth-left-column {
+  transform: translate(calc(-100% + var(--hud-offset)), 0);
 }

frontend/public/earth/css/layer-panel.css

@@ -0,0 +1,260 @@
+/* layer-panel.css — layer toggle panel (below brand, in left column) */
+
+.hud-panel-layers {
+  /* Lives inside .earth-left-column — position is relative via column rule */
+  border-radius: 0;
+  padding: 0;
+  width: 100%;
+  z-index: 10;
+  overflow: hidden;
+  margin-top: calc(6px * var(--hud-scale));
+}
+
+/* ── Header / drag handle ─────────────────────────────────────── */
+
+.layer-panel-header {
+  display: flex;
+  align-items: center;
+  gap: calc(6px * var(--hud-scale));
+  padding: calc(8px * var(--hud-scale)) calc(10px * var(--hud-scale));
+  border-bottom: 1px solid var(--hud-line);
+  cursor: grab;
+  user-select: none;
+}
+
+.layer-panel-header:active {
+  cursor: grabbing;
+}
+
+.layer-panel-icon {
+  flex-shrink: 0;
+  font-size: calc(15px * var(--hud-scale));
+  color: var(--hud-text-soft);
+  line-height: 1;
+  font-variation-settings: 'FILL' 0, 'wght' 400, 'GRAD' 0, 'opsz' 20;
+  pointer-events: none;
+}
+
+.layer-panel-title {
+  flex: 1 1 auto;
+  margin: 0;
+  color: var(--hud-title);
+  font-size: calc(0.82rem * var(--hud-scale));
+  font-weight: 600;
+  letter-spacing: 0.04em;
+  line-height: 1.2;
+}
+
+/* ── Collapse / generic icon button ──────────────────────────── */
+
+.layer-panel-btn {
+  display: inline-flex;
+  align-items: center;
+  justify-content: center;
+  width: calc(22px * var(--hud-scale));
+  height: calc(22px * var(--hud-scale));
+  min-width: calc(22px * var(--hud-scale));
+  padding: 0;
+  border: none;
+  border-radius: calc(4px * var(--hud-scale));
+  background: transparent;
+  color: var(--hud-text-muted);
+  cursor: pointer;
+  flex-shrink: 0;
+  transition: background 0.14s ease, color 0.14s ease;
+}
+
+.layer-panel-btn:hover {
+  background: rgba(255, 255, 255, 0.07);
+  color: var(--hud-text);
+}
+
+.layer-panel-btn .material-symbols-rounded {
+  font-size: calc(14px * var(--hud-scale));
+  line-height: 1;
+  pointer-events: none;
+  transition: transform 0.22s ease;
+}
+
+/* Chevron rotates when collapsed */
+.layer-panel--collapsed .layer-panel-btn .material-symbols-rounded {
+  transform: rotate(180deg);
+}
+
+/* ── Search bar ───────────────────────────────────────────────── */
+
+.layer-panel-search {
+  display: flex;
+  align-items: center;
+  gap: calc(5px * var(--hud-scale));
+  padding: calc(6px * var(--hud-scale)) calc(10px * var(--hud-scale));
+  border-bottom: 1px solid var(--hud-line);
+}
+
+.layer-panel-search-icon {
+  flex-shrink: 0;
+  font-size: calc(14px * var(--hud-scale));
+  color: var(--hud-text-soft);
+  line-height: 1;
+  font-variation-settings: 'FILL' 0, 'wght' 400, 'GRAD' 0, 'opsz' 20;
+  pointer-events: none;
+}
+
+.layer-panel-search-input {
+  flex: 1 1 auto;
+  min-width: 0;
+  border: none;
+  background: transparent;
+  color: var(--hud-text);
+  font-size: calc(0.78rem * var(--hud-scale));
+  line-height: 1.4;
+  outline: none;
+  padding: 0;
+}
+
+.layer-panel-search-input::placeholder {
+  color: var(--hud-text-soft);
+}
+
+/* ── Empty search state ───────────────────────────────────────── */
+
+.layer-panel-empty {
+  padding: calc(12px * var(--hud-scale)) calc(10px * var(--hud-scale));
+  color: var(--hud-text-soft);
+  font-size: calc(0.78rem * var(--hud-scale));
+  text-align: center;
+}
+
+/* ── Collapsible body ─────────────────────────────────────────── */
+
+.layer-panel-body {
+  overflow: hidden;
+  max-height: 600px;
+  transition: max-height 0.24s ease, opacity 0.18s ease;
+  opacity: 1;
+}
+
+.layer-panel--collapsed .layer-panel-body {
+  max-height: 0;
+  opacity: 0;
+}
+
+/* ── Layer rows ───────────────────────────────────────────────── */
+
+.layer-panel-list {
+  display: flex;
+  flex-direction: column;
+}
+
+.layer-row {
+  display: flex;
+  align-items: center;
+  gap: calc(8px * var(--hud-scale));
+  padding: calc(9px * var(--hud-scale)) calc(10px * var(--hud-scale));
+  border-bottom: 1px solid var(--hud-line);
+  transition: background 0.14s ease;
+}
+
+.layer-row:last-child {
+  border-bottom: none;
+}
+
+.layer-row:hover {
+  background: rgba(255, 255, 255, 0.03);
+}
+
+.layer-row-icon {
+  flex-shrink: 0;
+  font-size: calc(15px * var(--hud-scale));
+  color: var(--hud-text-soft);
+  line-height: 1;
+  font-variation-settings: 'FILL' 0, 'wght' 400, 'GRAD' 0, 'opsz' 20;
+  transition: color 0.18s ease;
+}
+
+/* Icon brightens when layer is active */
+.layer-row:has(.layer-row-toggle.active) .layer-row-icon {
+  color: var(--hud-accent);
+}
+
+.layer-row-copy {
+  flex: 1 1 auto;
+  min-width: 0;
+  display: flex;
+  flex-direction: column;
+  gap: calc(1px * var(--hud-scale));
+}
+
+.layer-row-label {
+  color: var(--hud-text);
+  font-size: calc(0.82rem * var(--hud-scale));
+  font-weight: 500;
+  line-height: 1.2;
+  white-space: nowrap;
+  overflow: hidden;
+  text-overflow: ellipsis;
+  transition: color 0.18s ease;
+}
+
+.layer-row:has(.layer-row-toggle.active) .layer-row-label {
+  color: var(--hud-title);
+}
+
+.layer-row-meta {
+  color: var(--hud-text-soft);
+  font-size: calc(0.58rem * var(--hud-scale));
+  letter-spacing: 0.08em;
+  text-transform: uppercase;
+  line-height: 1.2;
+}
+
+/* ── Toggle switch ────────────────────────────────────────────── */
+
+.layer-row-toggle {
+  flex-shrink: 0;
+  position: relative;
+  width: calc(34px * var(--hud-scale));
+  height: calc(20px * var(--hud-scale));
+  padding: 0;
+  border: none;
+  background: transparent;
+  cursor: pointer;
+}
+
+.layer-row-toggle-track {
+  display: block;
+  position: absolute;
+  inset: 0;
+  border-radius: 999px;
+  background: rgba(255, 255, 255, 0.08);
+  border: 1px solid rgba(215, 229, 242, 0.12);
+  transition: background 0.18s ease, border-color 0.18s ease;
+}
+
+/* Thumb */
+.layer-row-toggle-track::after {
+  content: "";
+  position: absolute;
+  top: calc(2px * var(--hud-scale));
+  left: calc(2px * var(--hud-scale));
+  width: calc(14px * var(--hud-scale));
+  height: calc(14px * var(--hud-scale));
+  border-radius: 50%;
+  background: #c8d8ea;
+  box-shadow: 0 2px 6px rgba(1, 8, 18, 0.3);
+  transition: transform 0.18s ease, background 0.18s ease;
+}
+
+/* Active (ON) state */
+.layer-row-toggle.active .layer-row-toggle-track {
+  background: linear-gradient(180deg, rgba(143, 185, 255, 0.72), rgba(104, 147, 221, 0.78));
+  border-color: rgba(223, 236, 252, 0.28);
+}
+
+.layer-row-toggle.active .layer-row-toggle-track::after {
+  background: #f0f6ff;
+  transform: translateX(calc(14px * var(--hud-scale)));
+}
+
+/* Layout-expanded: layer panel slides off with .earth-left-column — no
+   individual rule needed since the whole column translates together. */

frontend/public/earth/css/legend.css

@@ -1,59 +1,177 @@
-/* legend */
+/* legend.css — compact tab-strip legend */
 
-#legend {
-  bottom: 20px;
-  left: 20px;
-  border-radius: 18px;
-  padding: 15px;
-  width: 220px;
+.hud-panel-legend {
+  bottom: var(--hud-offset);
+  left: var(--hud-offset);
+  border-radius: 0;
+  padding: 0;
+  width: min(calc(200px * var(--hud-scale)), calc(100vw - 32px));
   z-index: 10;
+  overflow: hidden;
 }
 
-#legend .legend-title {
-  color: #4db8ff;
-  margin-bottom: 10px;
-  font-size: 1.1rem;
+/* ── Drag bar ─────────────────────────────────────────────────── */
+
+.legend-bar {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  gap: calc(6px * var(--hud-scale));
+  padding: calc(6px * var(--hud-scale)) calc(8px * var(--hud-scale));
+  border-bottom: 1px solid var(--hud-line);
+  cursor: grab;
+  user-select: none;
 }
 
-#legend .legend-list {
+.legend-bar:active {
+  cursor: grabbing;
+}
+
+/* ── Mode tabs ────────────────────────────────────────────────── */
+
+.legend-tabs {
+  display: flex;
+  gap: calc(2px * var(--hud-scale));
+  flex: 1 1 auto;
+  min-width: 0;
+}
+
+.legend-tab {
+  padding: calc(3px * var(--hud-scale)) calc(7px * var(--hud-scale));
+  border-radius: calc(4px * var(--hud-scale));
+  border: 1px solid transparent;
+  background: transparent;
+  color: var(--hud-text-muted);
+  font-size: calc(0.68rem * var(--hud-scale));
+  font-family: inherit;
+  letter-spacing: 0.08em;
+  cursor: pointer;
+  transition: background 0.14s ease, color 0.14s ease, border-color 0.14s ease;
+  white-space: nowrap;
+}
+
+.legend-tab:hover {
+  background: rgba(255, 255, 255, 0.06);
+  color: var(--hud-text);
+}
+
+.legend-tab--active {
+  background: rgba(120, 180, 255, 0.12);
+  border-color: rgba(120, 180, 255, 0.2);
+  color: var(--hud-accent-strong);
+}
+
+/* ── Bar action buttons ───────────────────────────────────────── */
+
+.legend-bar-actions {
+  display: flex;
+  align-items: center;
+  gap: calc(2px * var(--hud-scale));
+  flex-shrink: 0;
+}
+
+.legend-bar-btn {
+  display: inline-flex;
+  align-items: center;
+  justify-content: center;
+  width: calc(20px * var(--hud-scale));
+  height: calc(20px * var(--hud-scale));
+  min-width: calc(20px * var(--hud-scale));
+  padding: 0;
+  border: none;
+  border-radius: calc(4px * var(--hud-scale));
+  background: transparent;
+  color: var(--hud-text-muted);
+  cursor: pointer;
+  transition: background 0.14s ease, color 0.14s ease;
+}
+
+.legend-bar-btn:hover {
+  background: rgba(255, 255, 255, 0.07);
+  color: var(--hud-text);
+}
+
+.legend-bar-btn .material-symbols-rounded {
+  font-size: calc(13px * var(--hud-scale));
+  line-height: 1;
+  font-variation-settings: 'FILL' 0, 'wght' 400, 'GRAD' 0, 'opsz' 20;
+  pointer-events: none;
+}
+
+/* Collapse chevron */
+#legend-collapse .material-symbols-rounded {
+  transition: transform 0.22s ease;
+}
+
+.legend--collapsed #legend-collapse .material-symbols-rounded {
+  transform: rotate(180deg);
+}
+
+/* ── Collapsible list body ────────────────────────────────────── */
+
+.legend-body {
+  max-height: calc(220px * var(--hud-scale));
+  overflow: hidden;
+  transition:
+    max-height 0.26s cubic-bezier(0.4, 0, 0.2, 1),
+    opacity 0.2s ease;
+  opacity: 1;
+}
+
+.legend--collapsed .legend-body {
+  max-height: 0;
+  opacity: 0;
+  pointer-events: none;
+}
+
+/* ── Item list ────────────────────────────────────────────────── */
+
+.legend-list {
   display: flex;
   flex-direction: column;
-  gap: 8px;
-  max-height: 202px;
+  padding: calc(4px * var(--hud-scale)) 0;
   overflow-y: auto;
-  padding-right: 4px;
+  max-height: calc(220px * var(--hud-scale));
   scrollbar-width: thin;
-  scrollbar-color: rgba(160, 220, 255, 0.4) transparent;
+  scrollbar-color: rgba(160, 186, 216, 0.28) transparent;
 }
 
-#legend .legend-list::-webkit-scrollbar {
-  width: 6px;
-}
-
-#legend .legend-list::-webkit-scrollbar-track {
-  background: transparent;
-}
-
-#legend .legend-list::-webkit-scrollbar-thumb {
-  background: linear-gradient(180deg, rgba(210, 237, 255, 0.28), rgba(110, 176, 255, 0.34));
+.legend-list::-webkit-scrollbar { width: 3px; }
+.legend-list::-webkit-scrollbar-track { background: transparent; }
+.legend-list::-webkit-scrollbar-thumb {
+  background: rgba(160, 186, 216, 0.26);
   border-radius: 999px;
 }
 
-#legend .legend-item {
+.legend-item {
   display: flex;
   align-items: center;
-  padding: 6px 8px;
-  border-radius: 10px;
-  background: linear-gradient(180deg, rgba(255, 255, 255, 0.05), rgba(120, 180, 255, 0.02));
-  border: 1px solid rgba(225, 242, 255, 0.06);
+  gap: calc(8px * var(--hud-scale));
+  padding: calc(5px * var(--hud-scale)) calc(10px * var(--hud-scale));
 }
 
-#legend .legend-color {
-  width: 20px;
-  height: 20px;
-  border-radius: 6px;
-  margin-right: 10px;
-  box-shadow:
-    inset 0 1px 0 rgba(255, 255, 255, 0.2),
-    0 0 12px rgba(77, 184, 255, 0.18);
+.legend-dot {
+  flex-shrink: 0;
+  width: calc(7px * var(--hud-scale));
+  height: calc(7px * var(--hud-scale));
+  border-radius: 50%;
+  box-shadow: 0 0 4px currentColor;
+}
+
+.legend-label {
+  color: var(--hud-text);
+  font-size: calc(0.78rem * var(--hud-scale));
+  font-weight: 400;
+  line-height: 1.3;
+  white-space: nowrap;
+  overflow: hidden;
+  text-overflow: ellipsis;
+}
+
+/* ── Layout-expanded ──────────────────────────────────────────── */
+
+.earth-app.layout-expanded .hud-panel-legend:not([data-dragged="true"]) {
+  left: var(--hud-offset);
+  bottom: var(--hud-offset);
+  transform: translate(calc(-100% + var(--hud-offset)), calc(100% - var(--hud-offset)));
 }

frontend/public/earth/css/toolbar.css

@@ -0,0 +1,588 @@
+/* toolbar.css - bottom dock and floating toolbar primitives */
+
+.earth-toolbar-group {
+  position: absolute;
+  bottom: 18px;
+  left: 50%;
+  transform: translateX(-50%);
+  display: flex;
+  flex-direction: row;
+  align-items: center;
+  justify-content: center;
+  z-index: 200;
+}
+
+.earth-toolbar-group,
+.hud-panel {
+  transition:
+    top 0.45s ease,
+    right 0.45s ease,
+    bottom 0.45s ease,
+    left 0.45s ease,
+    transform 0.45s ease,
+    box-shadow 0.45s ease;
+}
+
+.earth-toolbar {
+  position: relative;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  gap: 0;
+  background: transparent;
+  border: none;
+  box-shadow: none;
+  padding: 0;
+}
+
+.earth-toolbar-items {
+  display: flex;
+  gap: 10px;
+  align-items: center;
+  flex-wrap: nowrap;
+}
+
+.earth-toolbar-popover {
+  position: relative;
+}
+
+.earth-toolbar-popover::before {
+  content: '';
+  position: absolute;
+  left: 50%;
+  bottom: 100%;
+  transform: translateX(-50%);
+  width: 56px;
+  height: 16px;
+  background: transparent;
+}
+
+.earth-toolbar-popover > .earth-stack-toolbar {
+  position: absolute;
+  left: 50%;
+  top: auto;
+  right: auto;
+  bottom: calc(100% + 12px);
+  transform: translate(-50%, 10px);
+  display: flex;
+  flex-direction: column;
+  align-items: center;
+  gap: 8px;
+  opacity: 0;
+  visibility: hidden;
+  pointer-events: none;
+  transition:
+    opacity 0.22s ease,
+    transform 0.22s ease,
+    visibility 0.22s ease;
+  z-index: 220;
+}
+
+.earth-toolbar-btn {
+  position: relative;
+  width: 28px;
+  height: 28px;
+  border: none;
+  border-radius: 0;
+  background: transparent;
+  color: #4db8ff;
+  font-size: 14px;
+  cursor: pointer;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  box-sizing: border-box;
+  padding: 0;
+  margin: 0;
+  overflow: visible;
+  appearance: none;
+  -webkit-appearance: none;
+}
+
+.earth-toolbar-btn.floating-btn {
+  width: 42px;
+  height: 42px;
+  min-width: 42px;
+  min-height: 42px;
+  border-radius: 50%;
+  overflow: hidden;
+}
+
+.earth-toolbar-btn:not(.liquid-glass-surface)::after {
+  content: none;
+}
+
+.earth-toolbar-btn .icon {
+  display: inline-flex;
+  align-items: center;
+  justify-content: center;
+  position: relative;
+  z-index: 1;
+  transform: translateZ(0);
+  transition: transform 0.16s ease, opacity 0.16s ease;
+  backface-visibility: hidden;
+  -webkit-backface-visibility: hidden;
+  line-height: 1;
+}
+
+.earth-toolbar-btn svg {
+  width: 20px;
+  height: 20px;
+  stroke: currentColor;
+  stroke-width: 2.1;
+  fill: none;
+  stroke-linecap: round;
+  stroke-linejoin: round;
+}
+
+.earth-toolbar-btn .material-symbols-rounded {
+  font-size: 21px;
+  line-height: 1;
+  font-variation-settings:
+    'FILL' 0,
+    'wght' 500,
+    'GRAD' 0,
+    'opsz' 24;
+  color: currentColor;
+  display: inline-flex;
+  align-items: center;
+  justify-content: center;
+  user-select: none;
+  pointer-events: none;
+  text-rendering: geometricPrecision;
+  -webkit-font-smoothing: antialiased;
+  -moz-osx-font-smoothing: grayscale;
+}
+
+.earth-toolbar-btn img {
+  width: 20px;
+  height: 20px;
+  display: block;
+  user-select: none;
+  pointer-events: none;
+  shape-rendering: geometricPrecision;
+  image-rendering: -webkit-optimize-contrast;
+  backface-visibility: hidden;
+  -webkit-backface-visibility: hidden;
+}
+
+.earth-toolbar-items > :nth-child(2n).floating-btn,
+.earth-toolbar-items > :nth-child(2n) .floating-btn {
+  animation-delay: 0.18s;
+}
+
+.earth-toolbar-items > :nth-child(3n).floating-btn,
+.earth-toolbar-items > :nth-child(3n) .floating-btn {
+  animation-delay: 0.34s;
+}
+
+.liquid-glass-surface {
+  --elastic-x: 0px;
+  --elastic-y: 0px;
+  --tilt-x: 0deg;
+  --tilt-y: 0deg;
+  --btn-scale: 1;
+  --press-offset: 0px;
+  --float-offset: 0px;
+  --glow-opacity: 0.24;
+  --glow-x: 50%;
+  --glow-y: 22%;
+  position: relative;
+  isolation: isolate;
+  transform-style: preserve-3d;
+  overflow: hidden;
+  background:
+    radial-gradient(circle at var(--glow-x) var(--glow-y), rgba(255, 255, 255, 0.16), transparent 34%),
+    radial-gradient(circle at 50% 118%, rgba(255, 255, 255, 0.08), transparent 30%),
+    linear-gradient(180deg, var(--glass-fill-top), var(--glass-fill-bottom)),
+    rgba(8, 20, 38, 0.22);
+  border: 1px solid var(--hud-border);
+  box-shadow:
+    inset 0 1px 0 rgba(255, 255, 255, 0.14),
+    inset 0 -1px 0 rgba(255, 255, 255, 0.05),
+    var(--glass-shadow),
+    var(--glass-glow);
+  backdrop-filter: blur(18px) saturate(145%);
+  -webkit-backdrop-filter: blur(18px) saturate(145%);
+  transform:
+    translate3d(var(--elastic-x), calc(var(--float-offset) + var(--press-offset) + var(--elastic-y)), 0)
+    scale(var(--btn-scale));
+  transition:
+    transform 0.22s ease,
+    box-shadow 0.22s ease,
+    background 0.22s ease,
+    opacity 0.18s ease,
+    border-color 0.22s ease;
+  animation: floatDock 3.8s ease-in-out infinite;
+}
+
+.liquid-glass-surface::before {
+  content: '';
+  position: absolute;
+  inset: 1px 1px 18px 1px;
+  border-radius: inherit;
+  background:
+    linear-gradient(180deg, rgba(255, 255, 255, 0.18), rgba(255, 255, 255, 0.05) 28%, transparent 68%);
+  opacity: 0.5;
+  pointer-events: none;
+  transform:
+    perspective(120px)
+    rotateX(calc(var(--tilt-x) * 0.7))
+    rotateY(calc(var(--tilt-y) * 0.7))
+    translate3d(calc(var(--elastic-x) * 0.22), calc(var(--elastic-y) * 0.22), 0);
+  transition: opacity 0.18s ease, transform 0.18s ease;
+}
+
+.liquid-glass-surface::after {
+  content: '';
+  position: absolute;
+  inset: -1px;
+  padding: 1.35px;
+  border-radius: inherit;
+  background:
+    linear-gradient(135deg, rgba(255, 255, 255, 0.36), rgba(168, 222, 255, 0.22) 34%, rgba(96, 175, 255, 0.16) 66%, rgba(255, 255, 255, 0.28));
+  opacity: 0.82;
+  pointer-events: none;
+  filter: url(#liquid-glass-distortion) blur(0.35px);
+  transform:
+    perspective(120px)
+    rotateX(calc(var(--tilt-x) * 0.5))
+    rotateY(calc(var(--tilt-y) * 0.5))
+    translate3d(calc(var(--elastic-x) * 0.16), calc(var(--elastic-y) * 0.16), 0);
+  -webkit-mask:
+    linear-gradient(#000 0 0) content-box,
+    linear-gradient(#000 0 0);
+  -webkit-mask-composite: xor;
+  mask:
+    linear-gradient(#000 0 0) content-box,
+    linear-gradient(#000 0 0);
+  mask-composite: exclude;
+  transition: opacity 0.18s ease, transform 0.18s ease;
+}
+
+.liquid-glass-surface:hover {
+  --btn-scale: 1.035;
+  --press-offset: -1px;
+  --glow-opacity: 0.32;
+  background:
+    radial-gradient(circle at var(--glow-x) var(--glow-y), rgba(255, 255, 255, 0.18), transparent 34%),
+    radial-gradient(circle at 50% 118%, rgba(255, 255, 255, 0.1), transparent 30%),
+    linear-gradient(180deg, rgba(255, 255, 255, 0.18), rgba(128, 198, 255, 0.1)),
+    rgba(8, 20, 38, 0.2);
+  border-color: var(--hud-border-hover);
+  box-shadow:
+    inset 0 1px 0 rgba(255, 255, 255, 0.2),
+    inset 0 -1px 0 rgba(255, 255, 255, 0.08),
+    0 18px 36px rgba(0, 0, 0, 0.24),
+    0 0 28px rgba(145, 214, 255, 0.22);
+}
+
+.liquid-glass-surface:hover::before {
+  opacity: 0.62;
+  transform: scale(1.01);
+}
+
+.liquid-glass-surface:hover::after {
+  opacity: 0.96;
+  transform: scale(1.01);
+}
+
+.liquid-glass-surface:active,
+.liquid-glass-surface.is-pressed {
+  --btn-scale: 0.942;
+  --press-offset: 2px;
+  --glow-opacity: 0.2;
+  background:
+    radial-gradient(circle at var(--glow-x) var(--glow-y), rgba(255, 255, 255, 0.24), transparent 34%),
+    radial-gradient(circle at 50% 118%, rgba(255, 255, 255, 0.14), transparent 30%),
+    linear-gradient(180deg, rgba(255, 255, 255, 0.24), rgba(146, 210, 255, 0.16)),
+    rgba(10, 24, 44, 0.24);
+  border-color: rgba(240, 249, 255, 0.58);
+  box-shadow:
+    inset 0 2px 10px rgba(0, 0, 0, 0.2),
+    inset 0 1px 0 rgba(255, 255, 255, 0.16),
+    0 4px 10px rgba(0, 0, 0, 0.18),
+    0 0 14px rgba(176, 226, 255, 0.18);
+}
+
+.liquid-glass-surface:active::before,
+.liquid-glass-surface.is-pressed::before {
+  opacity: 0.46;
+  transform: translateY(2px) scale(0.985);
+}
+
+.liquid-glass-surface:active::after,
+.liquid-glass-surface.is-pressed::after {
+  opacity: 0.78;
+  transform: scale(0.985);
+}
+
+.liquid-glass-surface:active .icon,
+.liquid-glass-surface.is-pressed .icon {
+  transform: translateY(1.5px);
+}
+
+.liquid-glass-surface:active img,
+.liquid-glass-surface.is-pressed img,
+.liquid-glass-surface:active .material-symbols-rounded,
+.liquid-glass-surface.is-pressed .material-symbols-rounded {
+  transform: translateY(1.5px);
+  transition: transform 0.16s ease, opacity 0.16s ease;
+}
+
+.liquid-glass-surface.active {
+  background:
+    radial-gradient(circle at var(--glow-x) var(--glow-y), rgba(255, 255, 255, 0.18), transparent 34%),
+    linear-gradient(180deg, rgba(255, 255, 255, 0.2), rgba(118, 200, 255, 0.14)),
+    rgba(11, 34, 58, 0.26);
+  border-color: var(--hud-border-active);
+  box-shadow:
+    inset 0 1px 0 rgba(255, 255, 255, 0.22),
+    inset 0 0 18px rgba(160, 220, 255, 0.14),
+    0 18px 34px rgba(0, 0, 0, 0.24),
+    0 0 30px rgba(145, 214, 255, 0.24);
+}
+
+.earth-rotate-toggle .icon-play,
+.earth-rotate-toggle.is-stopped .icon-pause,
+.earth-layout-toggle .layout-collapse,
+.earth-layout-toggle.active .layout-expand {
+  display: none;
+}
+
+.earth-rotate-toggle.is-stopped .icon-play,
+.earth-layout-toggle.active .layout-collapse {
+  display: inline-flex;
+}
+
+.earth-zoom-group:hover > .earth-zoom-toolbar,
+.earth-zoom-group:focus-within > .earth-zoom-toolbar,
+.earth-zoom-group.open > .earth-zoom-toolbar,
+.earth-info-group:hover > .earth-info-toolbar,
+.earth-info-group:focus-within > .earth-info-toolbar,
+.earth-info-group.open > .earth-info-toolbar {
+  opacity: 1;
+  visibility: visible;
+  pointer-events: auto;
+  transform: translate(-50%, 0);
+}
+
+.earth-zoom-group.force-closed > .earth-zoom-toolbar,
+.earth-info-group.force-closed > .earth-info-toolbar {
+  opacity: 0;
+  visibility: hidden;
+  pointer-events: none;
+  transform: translate(-50%, 8px);
+}
+
+.earth-zoom-group > .earth-zoom-toolbar,
+.earth-info-group > .earth-info-toolbar {
+  top: auto;
+  right: auto;
+  left: 50%;
+  bottom: calc(100% + 12px);
+  display: flex;
+  flex-direction: column;
+  align-items: center;
+  justify-content: flex-start;
+  gap: 8px;
+}
+
+.earth-info-toolbar {
+  width: min(280px, calc(100vw - 36px));
+  padding: 12px;
+  border-radius: 22px;
+  background:
+    radial-gradient(circle at top, rgba(255, 255, 255, 0.12), transparent 34%),
+    linear-gradient(180deg, rgba(16, 29, 48, 0.96), rgba(8, 18, 33, 0.94));
+  border: 1px solid rgba(211, 228, 246, 0.14);
+  box-shadow:
+    0 20px 40px rgba(0, 0, 0, 0.28),
+    inset 0 1px 0 rgba(255, 255, 255, 0.08);
+  backdrop-filter: blur(18px) saturate(135%);
+  -webkit-backdrop-filter: blur(18px) saturate(135%);
+}
+
+.earth-layer-toolbar-header {
+  width: 100%;
+  padding: 2px 4px 8px;
+  border-bottom: 1px solid rgba(201, 225, 247, 0.08);
+  margin-bottom: 2px;
+}
+
+.earth-layer-toolbar-title {
+  display: block;
+  color: var(--hud-accent-strong);
+  font-size: 0.86rem;
+  font-weight: 600;
+  letter-spacing: 0.08em;
+  text-transform: uppercase;
+}
+
+.earth-layer-toolbar-subtitle {
+  display: block;
+  margin-top: 4px;
+  color: var(--hud-text-soft);
+  font-size: 0.68rem;
+  letter-spacing: 0.08em;
+  text-transform: uppercase;
+}
+
+.earth-layer-btn {
+  width: 100%;
+  min-width: 0;
+  min-height: 52px;
+  height: auto;
+  border-radius: 16px;
+  padding: 12px 14px;
+  justify-content: space-between;
+  align-items: center;
+  gap: 12px;
+  overflow: hidden;
+  animation: none;
+}
+
+.earth-layer-btn__copy {
+  min-width: 0;
+  display: flex;
+  flex-direction: column;
+  align-items: flex-start;
+  gap: 3px;
+}
+
+.earth-layer-btn__label {
+  color: var(--hud-text);
+  font-size: 0.92rem;
+  font-weight: 600;
+  line-height: 1.15;
+}
+
+.earth-layer-btn__meta {
+  color: var(--hud-text-soft);
+  font-size: 0.68rem;
+  letter-spacing: 0.08em;
+  text-transform: uppercase;
+  line-height: 1.2;
+}
+
+.earth-layer-btn__state {
+  flex: 0 0 auto;
+  min-width: 42px;
+  padding: 5px 10px;
+  border-radius: 999px;
+  border: 1px solid rgba(201, 225, 247, 0.12);
+  color: var(--hud-text-soft);
+  font-size: 0.67rem;
+  font-weight: 700;
+  letter-spacing: 0.12em;
+  text-align: center;
+  text-transform: uppercase;
+  background: rgba(255, 255, 255, 0.04);
+}
+
+.earth-layer-btn.active .earth-layer-btn__state {
+  color: #dff4ff;
+  border-color: rgba(220, 240, 255, 0.24);
+  background: rgba(131, 197, 255, 0.14);
+}
+
+.earth-layer-btn .earth-toolbar-tooltip {
+  display: none;
+}
+
+.earth-zoom-toolbar .earth-zoom-btn,
+.earth-zoom-toolbar .earth-zoom-value {
+  width: 42px;
+  min-width: 42px;
+  border-radius: 50%;
+  color: #4db8ff;
+  animation: floatDock 3.8s ease-in-out infinite;
+}
+
+.earth-zoom-toolbar .earth-zoom-btn {
+  height: 42px;
+  font-size: 20px;
+  font-weight: 500;
+  line-height: 1;
+}
+
+.earth-zoom-toolbar .earth-zoom-value {
+  display: inline-flex;
+  align-items: center;
+  justify-content: center;
+  height: 42px;
+  padding: 0;
+  font-size: 0.68rem;
+  letter-spacing: normal;
+  animation-delay: 0.18s;
+}
+
+.earth-zoom-toolbar .earth-zoom-btn:active,
+.earth-zoom-toolbar .earth-zoom-btn.is-pressed,
+.earth-zoom-toolbar .earth-zoom-value:active,
+.earth-zoom-toolbar .earth-zoom-value.is-pressed {
+  letter-spacing: -0.01em;
+}
+
+.earth-zoom-toolbar .earth-zoom-btn:nth-child(1) {
+  animation-delay: 0s;
+}
+
+.earth-zoom-toolbar .earth-zoom-btn:nth-child(3) {
+  animation-delay: 0.34s;
+}
+
+.earth-zoom-toolbar .earth-toolbar-tooltip {
+  bottom: calc(100% + 10px);
+}
+
+.earth-zoom-toolbar .earth-toolbar-tooltip::after {
+  top: 100%;
+  left: 50%;
+  transform: translateX(-50%);
+  border: 6px solid transparent;
+  border-top-color: rgba(77, 184, 255, 0.4);
+}
+
+
+.earth-app.layout-expanded .earth-toolbar-group {
+  bottom: 18px;
+  transform: translateX(-50%);
+}
+
+.earth-toolbar-btn .earth-toolbar-tooltip {
+  position: absolute;
+  bottom: 56px;
+  left: 50%;
+  transform: translateX(-50%);
+  background: rgba(10, 10, 30, 0.95);
+  color: #fff;
+  padding: 6px 12px;
+  border-radius: 6px;
+  font-size: 12px;
+  white-space: nowrap;
+  opacity: 0;
+  visibility: hidden;
+  transition: all 0.2s ease;
+  border: 1px solid rgba(77, 184, 255, 0.4);
+  pointer-events: none;
+  z-index: 100;
+}
+
+.earth-toolbar-btn:hover .earth-toolbar-tooltip,
+.earth-toolbar-popover:hover > .earth-toolbar-btn .earth-toolbar-tooltip,
+.earth-toolbar-popover:focus-within > .earth-toolbar-btn .earth-toolbar-tooltip {
+  opacity: 1;
+  visibility: visible;
+  bottom: 58px;
+}
+
+.earth-toolbar-btn .earth-toolbar-tooltip::after {
+  content: '';
+  position: absolute;
+  top: 100%;
+  left: 50%;
+  transform: translateX(-50%);
+  border: 6px solid transparent;
+  border-top-color: rgba(77, 184, 255, 0.4);
+}

frontend/public/earth/css/tv-panel.css

@@ -0,0 +1,207 @@
+/* tv-panel */
+
+.hud-panel-tv {
+  bottom: var(--hud-offset);
+  right: var(--hud-offset);
+  width: calc(420px * var(--hud-scale));
+  max-width: calc(100vw - 32px);
+  min-width: calc(300px * var(--hud-scale));
+  min-height: calc(340px * var(--hud-scale));
+  padding: calc(18px * var(--hud-scale));
+  display: flex;
+  flex-direction: column;
+  gap: var(--hud-gap-sm);
+  z-index: 18;
+}
+
+.tv-panel-header-copy {
+  display: grid;
+  gap: calc(3px * var(--hud-scale));
+}
+
+.tv-panel-status {
+  color: var(--hud-text-soft);
+  font-size: calc(0.68rem * var(--hud-scale));
+  letter-spacing: 0.08em;
+  text-transform: uppercase;
+}
+
+.tv-panel-controls {
+  display: flex;
+  gap: var(--hud-gap-sm);
+  align-items: center;
+  min-width: 0;
+}
+
+.tv-panel-select {
+  flex: 1 1 auto;
+  min-width: 0;
+  appearance: none;
+  -webkit-appearance: none;
+  color-scheme: dark;
+  border: 1px solid rgba(201, 225, 247, 0.14);
+  border-radius: calc(12px * var(--hud-scale));
+  background: rgba(255, 255, 255, 0.04);
+  color: var(--hud-text);
+  padding: calc(10px * var(--hud-scale)) calc(12px * var(--hud-scale));
+  font-size: calc(0.84rem * var(--hud-scale));
+}
+
+.tv-panel-select option,
+.tv-panel-select optgroup {
+  background: #0a1422;
+  color: #eef5fc;
+}
+
+.tv-panel-actions {
+  display: flex;
+  gap: var(--hud-gap-xs);
+}
+
+.tv-panel-action {
+  border: 1px solid rgba(201, 225, 247, 0.12);
+  border-radius: calc(12px * var(--hud-scale));
+  background: rgba(255, 255, 255, 0.05);
+  color: var(--hud-text);
+  padding: calc(10px * var(--hud-scale)) calc(12px * var(--hud-scale));
+  display: inline-flex;
+  align-items: center;
+  justify-content: center;
+  white-space: nowrap;
+  font-size: calc(0.84rem * var(--hud-scale));
+  line-height: 1;
+  cursor: pointer;
+  transition: background 0.18s ease, border-color 0.18s ease, color 0.18s ease;
+}
+
+.tv-panel-action--icon {
+  padding: calc(9px * var(--hud-scale));
+  border-radius: calc(10px * var(--hud-scale));
+}
+
+.tv-panel-action--icon .material-symbols-rounded {
+  font-size: calc(18px * var(--hud-scale));
+  line-height: 1;
+  font-variation-settings: 'FILL' 0, 'wght' 400, 'GRAD' 0, 'opsz' 20;
+  pointer-events: none;
+}
+
+.tv-panel-action:hover:not(:disabled) {
+  background: rgba(255, 255, 255, 0.08);
+  border-color: rgba(225, 239, 255, 0.2);
+  color: var(--hud-accent-strong);
+}
+
+.tv-panel-action:disabled {
+  opacity: 0.45;
+  cursor: not-allowed;
+}
+
+.tv-panel-meta {
+  display: grid;
+  gap: calc(4px * var(--hud-scale));
+}
+
+.tv-panel-title {
+  color: var(--hud-text);
+  font-size: calc(0.96rem * var(--hud-scale));
+  font-weight: 600;
+}
+
+.tv-panel-subtitle {
+  color: var(--hud-text-muted);
+  font-size: calc(0.74rem * var(--hud-scale));
+  line-height: 1.35;
+}
+
+.tv-panel-notes {
+  color: var(--hud-text-soft);
+  font-size: calc(0.68rem * var(--hud-scale));
+  line-height: 1.45;
+}
+
+.tv-panel-catalog {
+  color: var(--hud-text-soft);
+  font-size: calc(0.66rem * var(--hud-scale));
+  letter-spacing: 0.04em;
+  text-transform: uppercase;
+}
+
+.tv-panel-player {
+  position: relative;
+  flex: 1 1 auto;
+  min-height: calc(220px * var(--hud-scale));
+  border-radius: calc(16px * var(--hud-scale));
+  overflow: hidden;
+  border: 1px solid rgba(201, 225, 247, 0.1);
+  background:
+    linear-gradient(180deg, rgba(9, 18, 32, 0.94), rgba(5, 11, 22, 0.94));
+}
+
+.tv-panel-empty,
+.tv-panel-iframe,
+.tv-panel-video {
+  position: absolute;
+  inset: 0;
+  width: 100%;
+  height: 100%;
+}
+
+.tv-panel-empty {
+  display: grid;
+  place-items: center;
+  padding: calc(18px * var(--hud-scale));
+  text-align: center;
+  color: var(--hud-text-muted);
+  font-size: calc(0.84rem * var(--hud-scale));
+  line-height: 1.5;
+}
+
+.tv-panel-iframe,
+.tv-panel-video {
+  border: 0;
+  background: #050a14;
+}
+
+.tv-panel-resize-handle {
+  position: absolute;
+  right: calc(8px * var(--hud-scale));
+  bottom: calc(8px * var(--hud-scale));
+  width: calc(18px * var(--hud-scale));
+  height: calc(18px * var(--hud-scale));
+  border: 0;
+  padding: 0;
+  background: transparent;
+  cursor: nwse-resize;
+  z-index: 2;
+}
+
+.tv-panel-resize-handle::before {
+  content: "";
+  position: absolute;
+  inset: 0;
+  border-right: 2px solid rgba(223, 235, 248, 0.46);
+  border-bottom: 2px solid rgba(223, 235, 248, 0.46);
+  border-bottom-right-radius: calc(10px * var(--hud-scale));
+  opacity: 0.78;
+  transition: opacity 0.18s ease, border-color 0.18s ease;
+}
+
+.tv-panel-resize-handle:hover::before {
+  opacity: 1;
+  border-color: rgba(244, 249, 255, 0.78);
+}
+
+.hud-panel-tv.is-resizing {
+  transition: none !important;
+  user-select: none;
+}
+
+.earth-app.layout-expanded .hud-panel-tv:not([data-dragged="true"]) {
+  bottom: var(--hud-offset);
+  right: var(--hud-offset);
+  transform: translate(calc(100% - var(--hud-offset)), calc(100% - var(--hud-offset)));
+}
+
+/* TV panel keeps its fixed width on all screen sizes.
+   Responsive stretching removed — width only changes if user manually resizes. */

frontend/public/earth/index.html

@@ -9,19 +9,38 @@
     "imports": {
       "three": "https://esm.sh/three@0.128.0",
       "simplex-noise": "https://esm.sh/simplex-noise@4.0.1",
-      "satellite.js": "https://esm.sh/satellite.js@5.0.0"
+      "satellite.js": "https://esm.sh/satellite.js@5.0.0",
+      "hls.js": "https://esm.sh/hls.js@1.6.15"
     }
   }
   </script>
+  <script>
+    (function applyInitialHudScale() {
+      var referenceWidth = 1920;
+      var referenceHeight = 1080;
+      var minScale = 0.7;
+      var maxScale = 1;
+      var widthScale = window.innerWidth / referenceWidth;
+      var heightScale = window.innerHeight / referenceHeight;
+      var scale = Math.min(widthScale, heightScale);
+      var clampedScale = Math.max(minScale, Math.min(maxScale, scale));
+
+      document.documentElement.style.setProperty("--hud-scale", clampedScale.toFixed(3));
+    })();
+  </script>
   <link rel="stylesheet" href="css/base.css">
+  <link rel="stylesheet" href="css/hud.css">
+  <link rel="stylesheet" href="css/toolbar.css">
   <link rel="stylesheet" href="css/info-panel.css">
-  <link rel="stylesheet" href="css/coordinates-display.css">
   <link rel="stylesheet" href="css/legend.css">
   <link rel="stylesheet" href="css/earth-stats.css">
+  <link rel="stylesheet" href="css/tv-panel.css">
+  <link rel="stylesheet" href="css/layer-panel.css">
+  <link rel="stylesheet" href="https://fonts.googleapis.com/css2?family=Roboto+Condensed:wght@400;500;600&display=swap">
   <link rel="stylesheet" href="https://fonts.googleapis.com/css2?family=Material+Symbols+Rounded:opsz,wght,FILL,GRAD@24,500,0,0">
 </head>
-<body>
-  <svg aria-hidden="true" width="0" height="0" style="position:absolute; width:0; height:0; pointer-events:none;">
+<body class="earth-page">
+  <svg class="earth-filters" aria-hidden="true" width="0" height="0">
     <defs>
       <filter id="liquid-glass-distortion" x="-20%" y="-20%" width="140%" height="140%">
         <feTurbulence type="fractalNoise" baseFrequency="0.012 0.02" numOctaves="2" seed="7" result="noise" />
@@ -30,208 +49,372 @@
       </filter>
     </defs>
   </svg>
-  <div id="container">
-    <div id="info-panel">
-      <h1>智能星球计划</h1>
-      <div class="subtitle">
-        <span class="subtitle-main">现实层宇宙全息感知系统</span>
-        <span class="subtitle-meta">卫星 · 海底光缆 · 算力基础设施</span>
+  <div id="container" class="earth-app">
+    <div class="earth-left-column">
+      <div id="brand-panel" class="hud-panel hud-panel-brand">
+        <div id="brand-root"></div>
       </div>
-      
-      <div id="info-card" class="info-card" style="display: none;">
-        <div class="info-card-header">
-          <span class="info-card-icon" id="info-card-icon">🛰️</span>
-          <h3 id="info-card-title">详情</h3>
+
+      <div id="layer-toggles" class="hud-panel hud-panel-layers hud-panel-draggable" data-panel-key="layer-toggles">
+        <!-- Header / drag handle -->
+        <div class="layer-panel-header hud-panel-drag-handle">
+          <span class="material-symbols-rounded layer-panel-icon">layers</span>
+          <span class="layer-panel-title">图层</span>
+          <button id="layer-panel-collapse" class="layer-panel-btn" type="button" aria-label="折叠图层列表" title="折叠">
+            <span class="material-symbols-rounded">expand_more</span>
+          </button>
+        </div>
+
+        <!-- Collapsible body -->
+        <div class="layer-panel-body" id="layer-panel-body">
+          <!-- Search -->
+          <div class="layer-panel-search">
+            <span class="material-symbols-rounded layer-panel-search-icon">search</span>
+            <input
+              type="search"
+              id="layer-search-input"
+              class="layer-panel-search-input"
+              placeholder="搜索图层..."
+              autocomplete="off"
+              spellcheck="false"
+            >
+            <button id="layer-search-clear" class="layer-panel-btn layer-search-clear" type="button" aria-label="清除搜索" title="清除" hidden>
+              <span class="material-symbols-rounded">close</span>
+            </button>
+          </div>
+
+          <!-- Layer rows -->
+          <div class="layer-panel-list" id="layer-panel-list">
+            <div class="layer-row" data-layer-name="地形 terrain">
+              <span class="material-symbols-rounded layer-row-icon">landscape</span>
+              <div class="layer-row-copy">
+                <span class="layer-row-label">地形</span>
+                <span class="layer-row-meta">Terrain</span>
+              </div>
+              <button id="toggle-terrain" class="layer-row-toggle" type="button" role="switch" aria-checked="false" title="切换地形显示">
+                <span class="layer-row-toggle-track"></span>
+              </button>
+            </div>
+            <div class="layer-row" data-layer-name="卫星 satellites">
+              <span class="material-symbols-rounded layer-row-icon">satellite_alt</span>
+              <div class="layer-row-copy">
+                <span class="layer-row-label">卫星</span>
+                <span class="layer-row-meta">Satellites</span>
+              </div>
+              <button id="toggle-satellites" class="layer-row-toggle" type="button" role="switch" aria-checked="false" title="切换卫星显示">
+                <span class="layer-row-toggle-track"></span>
+              </button>
+            </div>
+            <div class="layer-row" data-layer-name="轨迹 trails">
+              <span class="material-symbols-rounded layer-row-icon">timeline</span>
+              <div class="layer-row-copy">
+                <span class="layer-row-label">轨迹</span>
+                <span class="layer-row-meta">Trails</span>
+              </div>
+              <button id="toggle-trails" class="layer-row-toggle active" type="button" role="switch" aria-checked="true" title="切换轨迹显示">
+                <span class="layer-row-toggle-track"></span>
+              </button>
+            </div>
+            <div class="layer-row" data-layer-name="海缆 subsea cables">
+              <span class="material-symbols-rounded layer-row-icon">cable</span>
+              <div class="layer-row-copy">
+                <span class="layer-row-label">海缆</span>
+                <span class="layer-row-meta">Subsea Cables</span>
+              </div>
+              <button id="toggle-cables" class="layer-row-toggle active" type="button" role="switch" aria-checked="true" title="切换海缆显示">
+                <span class="layer-row-toggle-track"></span>
+              </button>
+            </div>
+            <div class="layer-row" data-layer-name="bgp观测 routing signals">
+              <span class="material-symbols-rounded layer-row-icon">hub</span>
+              <div class="layer-row-copy">
+                <span class="layer-row-label">BGP观测</span>
+                <span class="layer-row-meta">Routing Signals</span>
+              </div>
+              <button id="toggle-bgp" class="layer-row-toggle active" type="button" role="switch" aria-checked="true" title="切换BGP观测显示">
+                <span class="layer-row-toggle-track"></span>
+              </button>
+            </div>
+          </div>
+
+          <!-- Empty search state -->
+          <div class="layer-panel-empty" id="layer-panel-empty" hidden>无匹配图层</div>
         </div>
-        <div id="info-card-content"></div>
       </div>
-      
-      <div id="error-message" class="error-message"></div>
     </div>
 
-    <div id="right-toolbar-group">
-      <div id="control-toolbar">
-        <div class="toolbar-items">
-          <button id="search-action" class="toolbar-btn floating-btn liquid-glass-surface" title="搜索功能（待开发）">
+    <!-- Floating detail panel — positioned near click by JS -->
+    <div id="info-panel" class="hud-panel hud-panel-info hud-panel-draggable" aria-live="polite">
+      <div id="info-card" class="info-card">
+        <div class="info-card-header hud-panel-drag-handle">
+          <span class="info-card-icon" id="info-card-icon">🛰️</span>
+          <h3 id="info-card-title">详情</h3>
+          <button class="info-card-close hud-panel-close" type="button" aria-label="关闭详情">
+            <span class="material-symbols-rounded">close</span>
+          </button>
+        </div>
+        <div id="info-card-content" class="info-card-content"></div>
+      </div>
+      <div id="error-message" class="hud-error-message"></div>
+    </div>
+
+    <div id="right-toolbar-group" class="earth-toolbar-group">
+      <div id="control-toolbar" class="earth-toolbar">
+        <div class="earth-toolbar-items">
+          <button id="search-action" class="floating-btn liquid-glass-surface earth-toolbar-btn" title="搜索功能（待开发）">
             <span class="icon" aria-hidden="true">
               <span class="material-symbols-rounded">search</span>
             </span>
-            <span class="tooltip">搜索功能（待开发）</span>
+            <span class="tooltip earth-toolbar-tooltip">搜索功能（待开发）</span>
           </button>
-          <button id="rotate-toggle" class="toolbar-btn floating-btn liquid-glass-surface" title="自动旋转">
+          <button id="rotate-toggle" class="floating-btn liquid-glass-surface earth-toolbar-btn earth-rotate-toggle" title="自动旋转">
             <span class="icon rotate-icon icon-pause" aria-hidden="true">
               <span class="material-symbols-rounded">pause</span>
             </span>
             <span class="icon rotate-icon icon-play" aria-hidden="true">
               <span class="material-symbols-rounded">play_arrow</span>
             </span>
-            <span class="tooltip">自动旋转</span>
+            <span class="tooltip earth-toolbar-tooltip">自动旋转</span>
           </button>
-          <div id="info-control-group" class="floating-popover-group">
-            <button id="info-trigger" class="toolbar-btn floating-btn liquid-glass-surface" title="显示控制">
-              <span class="icon" aria-hidden="true">
-                <span class="material-symbols-rounded">info</span>
-              </span>
-              <span class="tooltip">显示控制</span>
-            </button>
-            <div id="info-toolbar" class="stack-toolbar">
-              <button id="toggle-terrain" class="toolbar-btn floating-btn liquid-glass-surface" title="显示/隐藏地形">
-                <span class="icon" aria-hidden="true">
-                  <span class="material-symbols-rounded">terrain</span>
-                </span>
-                <span class="tooltip">显示/隐藏地形</span>
-              </button>
-              <button id="toggle-trails" class="toolbar-btn floating-btn liquid-glass-surface active" title="显示/隐藏轨迹">
-                <span class="icon" aria-hidden="true">
-                  <span class="material-symbols-rounded">timeline</span>
-                </span>
-                <span class="tooltip">隐藏轨迹</span>
-              </button>
-              <button id="toggle-satellites" class="toolbar-btn floating-btn liquid-glass-surface" title="显示/隐藏卫星">
-                <span class="icon" aria-hidden="true">
-                  <span class="material-symbols-rounded">satellite_alt</span>
-                </span>
-                <span class="tooltip">显示卫星</span>
-              </button>
-              <button id="toggle-bgp" class="toolbar-btn floating-btn liquid-glass-surface active" title="显示/隐藏BGP观测">
-                <span class="icon" aria-hidden="true">
-                  <span class="material-symbols-rounded">radar</span>
-                </span>
-                <span class="tooltip">隐藏BGP观测</span>
-              </button>
-              <button id="toggle-cables" class="toolbar-btn floating-btn liquid-glass-surface active" title="显示/隐藏线缆">
-                <span class="icon" aria-hidden="true">
-                  <span class="material-symbols-rounded">cable</span>
-                </span>
-                <span class="tooltip">隐藏线缆</span>
-              </button>
-            </div>
-          </div>
-          <button id="reload-data" class="toolbar-btn floating-btn liquid-glass-surface" title="重新加载数据">
+          <button id="toggle-tv" class="floating-btn liquid-glass-surface earth-toolbar-btn" title="新闻直播">
+            <span class="icon" aria-hidden="true">
+              <span class="material-symbols-rounded">live_tv</span>
+            </span>
+            <span class="tooltip earth-toolbar-tooltip">打开新闻直播</span>
+          </button>
+          <button id="reload-data" class="floating-btn liquid-glass-surface earth-toolbar-btn" title="重新加载数据">
             <span class="icon" aria-hidden="true">
               <span class="material-symbols-rounded">refresh</span>
             </span>
-            <span class="tooltip">重新加载数据</span>
+            <span class="tooltip earth-toolbar-tooltip">重新加载数据</span>
           </button>
-          <div id="zoom-control-group" class="floating-popover-group">
-            <button id="zoom-trigger" class="toolbar-btn floating-btn liquid-glass-surface" title="缩放控制">
+          <div id="zoom-control-group" class="earth-toolbar-popover earth-zoom-group">
+            <button id="zoom-trigger" class="floating-btn liquid-glass-surface earth-toolbar-btn" title="缩放控制">
               <span class="icon" aria-hidden="true">
                 <span class="material-symbols-rounded">zoom_in</span>
               </span>
-              <span class="tooltip">缩放控制</span>
+              <span class="tooltip earth-toolbar-tooltip">缩放控制</span>
             </button>
-            <div id="zoom-toolbar" class="stack-toolbar">
-              <button id="zoom-in" class="zoom-btn liquid-glass-surface" title="放大">+<span class="tooltip">放大</span></button>
-              <span id="zoom-value" class="zoom-percent liquid-glass-surface" title="重置缩放到100%">100%<span class="tooltip">重置缩放到100%</span></span>
-              <button id="zoom-out" class="zoom-btn liquid-glass-surface" title="缩小">−<span class="tooltip">缩小</span></button>
+            <div id="zoom-toolbar" class="earth-stack-toolbar earth-zoom-toolbar">
+              <button id="zoom-in" class="liquid-glass-surface earth-zoom-btn" title="放大" aria-label="放大"><span aria-hidden="true">+</span></button>
+              <span id="zoom-value" class="liquid-glass-surface earth-zoom-value" title="重置缩放到100%">100%<span class="tooltip earth-toolbar-tooltip">重置缩放到100%</span></span>
+              <button id="zoom-out" class="liquid-glass-surface earth-zoom-btn" title="缩小" aria-label="缩小"><span aria-hidden="true">−</span></button>
             </div>
           </div>
-          <button id="reset-view" class="toolbar-btn floating-btn liquid-glass-surface" title="重置视角">
+          <button id="settings-trigger" class="floating-btn liquid-glass-surface earth-toolbar-btn" title="设置">
+            <span class="icon" aria-hidden="true">
+              <span class="material-symbols-rounded">settings</span>
+            </span>
+            <span class="tooltip earth-toolbar-tooltip">设置</span>
+          </button>
+          <button id="reset-view" class="floating-btn liquid-glass-surface earth-toolbar-btn" title="重置视角">
             <span class="icon" aria-hidden="true">
               <span class="material-symbols-rounded">my_location</span>
             </span>
-            <span class="tooltip">重置视角</span>
+            <span class="tooltip earth-toolbar-tooltip">重置视角</span>
           </button>
-          <button id="layout-toggle" class="toolbar-btn floating-btn liquid-glass-surface" title="最大化布局">
+          <button id="layout-toggle" class="floating-btn liquid-glass-surface earth-toolbar-btn earth-layout-toggle" title="最大化布局">
             <span class="icon layout-icon layout-expand" aria-hidden="true">
               <span class="material-symbols-rounded">open_in_full</span>
             </span>
             <span class="icon layout-icon layout-collapse" aria-hidden="true">
               <span class="material-symbols-rounded">close_fullscreen</span>
             </span>
-            <span class="tooltip">最大化布局</span>
+            <span class="tooltip earth-toolbar-tooltip">最大化布局</span>
           </button>
         </div>
       </div>
     </div>
 
-    <div id="coordinates-display">
-      <h3 style="color:#4db8ff; margin-bottom:8px; font-size:1.1rem;">坐标信息</h3>
-      <div class="coord-item">
-        <span class="coord-label">经度:</span>
-        <span id="longitude-value" class="coord-value">0.00°</span>
-      </div>
-      <div class="coord-item">
-        <span class="coord-label">纬度:</span>
-        <span id="latitude-value" class="coord-value">0.00°</span>
-      </div>
-      <div id="zoom-level">缩放: 1.0x</div>
-      <div class="mouse-coords" id="mouse-coords">鼠标位置: 无</div>
-    </div>
 
-    <div id="legend">
-      <h3 class="legend-title">线缆图例</h3>
-      <div class="legend-list">
-        <div class="legend-item">
-          <div class="legend-color" style="background-color: #ff4444;"></div>
-          <span>Americas II</span>
+    <div id="legend" class="hud-panel hud-panel-legend hud-panel-draggable" data-panel-key="legend">
+      <!-- Drag bar: mode tabs + collapse + close -->
+      <div class="legend-bar hud-panel-drag-handle">
+        <div class="legend-tabs" id="legend-tabs">
+          <button class="legend-tab legend-tab--active" data-legend-mode="cables">海缆</button>
+          <button class="legend-tab" data-legend-mode="satellites">卫星</button>
+          <button class="legend-tab" data-legend-mode="bgp">BGP</button>
         </div>
-        <div class="legend-item">
-          <div class="legend-color" style="background-color: #44ff44;"></div>
-          <span>AU Aleutian A</span>
-        </div>
-        <div class="legend-item">
-          <div class="legend-color" style="background-color: #4444ff;"></div>
-          <span>AU Aleutian B</span>
-        </div>
-        <div class="legend-item">
-          <div class="legend-color" style="background-color: #ffff44;"></div>
-          <span>其他电缆</span>
+        <div class="legend-bar-actions">
+          <button id="legend-collapse" class="legend-bar-btn" title="折叠">
+            <span class="material-symbols-rounded">expand_less</span>
+          </button>
+          <button class="legend-bar-btn hud-panel-close" type="button" data-close-panel="legend" aria-label="关闭图例">
+            <span class="material-symbols-rounded">close</span>
+          </button>
         </div>
       </div>
+      <!-- Collapsible list -->
+      <div id="legend-body" class="legend-body">
+        <div class="legend-list"></div>
+      </div>
     </div>
 
-    <div id="earth-stats">
-      <h3 style="color:#4db8ff; margin-bottom:10px; font-size:1.1rem;">地球信息</h3>
-      <div class="stats-item">
-        <span class="stats-label">电缆系统:</span>
-        <span class="stats-value" id="cable-count">0个</span>
+    <div id="earth-stats" class="hud-panel hud-panel-stats hud-panel-draggable" data-panel-key="earth-stats">
+      <!-- Thin drag bar with kicker + close -->
+      <div class="stats-drag-bar hud-panel-drag-handle">
+        <span class="stats-kicker">全球态势</span>
+        <button class="hud-panel-close" type="button" data-close-panel="earth-stats" aria-label="关闭地球信息">
+          <span class="material-symbols-rounded">close</span>
+        </button>
       </div>
-      <div class="stats-item">
-        <span class="stats-label">状态:</span>
-        <span class="stats-value" id="cable-status-summary">-</span>
+
+      <!-- 2-col KPI grid -->
+      <div class="stats-grid">
+        <div class="stat-cell">
+          <span class="stat-num" id="cable-count">—</span>
+          <span class="stat-label">海缆系统</span>
+        </div>
+        <div class="stat-cell">
+          <span class="stat-num" id="landing-point-count">—</span>
+          <span class="stat-label">登陆点</span>
+        </div>
+        <div class="stat-cell">
+          <span class="stat-num" id="satellite-count">—</span>
+          <span class="stat-label">在轨卫星</span>
+        </div>
+        <div class="stat-cell">
+          <span class="stat-num" id="bgp-anomaly-count">—</span>
+          <span class="stat-label">BGP 事件</span>
+        </div>
+        <div class="stat-cell">
+          <span class="stat-num" id="bgp-collector-count">—</span>
+          <span class="stat-label">BGP 观测站</span>
+        </div>
+        <div class="stat-cell">
+          <span class="stat-num stat-num--sm" id="cable-status-summary">—</span>
+          <span class="stat-label">运行中</span>
+        </div>
       </div>
-      <div class="stats-item">
-        <span class="stats-label">登陆点:</span>
-        <span class="stats-value" id="landing-point-count">0个</span>
-      </div>
-      <div class="stats-item">
-        <span class="stats-label">地形:</span>
-        <span class="stats-value" id="terrain-status">开启</span>
-      </div>
-      <div class="stats-item">
-        <span class="stats-label">卫星:</span>
-        <span class="stats-value" id="satellite-count">0 颗</span>
-      </div>
-      <div class="stats-item">
-        <span class="stats-label">BGP事件:</span>
-        <span class="stats-value" id="bgp-anomaly-count">0 条</span>
-      </div>
-      <div class="stats-item">
-        <span class="stats-label">观测站:</span>
-        <span class="stats-value" id="bgp-collector-count">0 个</span>
-      </div>
-      <div class="stats-item">
-        <span class="stats-label">BGP态势:</span>
-        <span class="stats-value" id="bgp-status-summary">暂无观测数据</span>
-      </div>
-      <div class="stats-item">
-        <span class="stats-label">视角距离:</span>
-        <span class="stats-value" id="camera-distance">300 km</span>
-      </div>
-      <div class="stats-item">
-        <span class="stats-label">纹理质量:</span>
-        <span class="stats-value" id="texture-quality">8K 卫星图</span>
+
+      <!-- BGP status footer -->
+      <div class="stats-footer">
+        <span class="stats-footer-dot"></span>
+        <span id="bgp-status-summary" class="stats-footer-text">暂无观测数据</span>
       </div>
+
+      <!-- hidden elements kept for JS compatibility -->
+      <span id="terrain-status" hidden></span>
+      <span id="texture-quality" hidden></span>
+      <span id="camera-distance" hidden></span>
     </div>
 
-    <div id="loading">
-      <div id="loading-spinner"></div>
-      <div id="loading-title">正在初始化全球态势数据...</div>
-      <div id="loading-subtitle" style="font-size:0.9rem; margin-top:10px; color:#aaa;">同步卫星、海底光缆、登陆点与BGP态势数据</div>
+    <div id="tv-panel" class="hud-panel hud-panel-tv hud-panel-draggable" data-panel-key="tv-panel">
+      <div class="hud-panel-header hud-panel-drag-handle">
+        <div class="tv-panel-header-copy">
+          <h3 class="hud-panel-title">新闻直播</h3>
+          <span id="tv-source-status" class="tv-panel-status">等待加载直播源</span>
+        </div>
+        <button class="hud-panel-close" type="button" data-close-panel="tv-panel" aria-label="关闭电视直播">
+          <span class="material-symbols-rounded">close</span>
+        </button>
+      </div>
+      <div class="tv-panel-controls">
+        <select id="tv-source-select" class="tv-panel-select" aria-label="选择新闻直播源"></select>
+        <div class="tv-panel-actions">
+          <button id="tv-refresh" class="tv-panel-action tv-panel-action--icon" type="button" title="刷新直播源" aria-label="刷新直播源">
+            <span class="material-symbols-rounded">refresh</span>
+          </button>
+          <button id="tv-open-external" class="tv-panel-action tv-panel-action--icon" type="button" title="访问官网" aria-label="访问官网">
+            <span class="material-symbols-rounded">open_in_new</span>
+          </button>
+        </div>
+      </div>
+      <div class="tv-panel-meta">
+        <div id="tv-source-title" class="tv-panel-title">暂无可用频道</div>
+        <div id="tv-source-meta" class="tv-panel-subtitle">当前未配置可播放新闻直播源</div>
+        <div id="tv-source-catalog" class="tv-panel-catalog">频道目录待同步</div>
+        <div id="tv-source-notes" class="tv-panel-notes">支持后台配置默认源与采集器补充源。</div>
+      </div>
+      <div class="tv-panel-player">
+        <div id="tv-empty-state" class="tv-panel-empty">暂无可播放直播源，请先在系统配置中添加频道。</div>
+        <iframe
+          id="tv-iframe"
+          class="tv-panel-iframe"
+          hidden
+          title="新闻直播"
+          referrerpolicy="strict-origin-when-cross-origin"
+          allow="autoplay; fullscreen; picture-in-picture"
+        ></iframe>
+        <video id="tv-video" class="tv-panel-video" hidden controls autoplay muted playsinline></video>
+      </div>
+      <button
+        id="tv-resize-handle"
+        class="tv-panel-resize-handle"
+        type="button"
+        aria-label="调整电视直播窗口大小"
+        title="调整大小"
+      ></button>
+    </div>
+
+    <div id="loading" class="earth-loading">
+      <div id="loading-spinner" class="earth-loading-spinner"></div>
+      <div id="loading-title" class="earth-loading-title earth-loading-text">正在初始化全球态势数据...</div>
+      <div id="loading-subtitle" class="earth-loading-subtitle">同步卫星、海底光缆、登陆点与BGP态势数据</div>
+    </div>
+    <div id="status-message" class="earth-status-message"></div>
+    <div id="tooltip" class="earth-tooltip"></div>
+    <div id="settings-modal" class="earth-settings-modal" aria-hidden="true">
+      <div id="settings-backdrop" class="earth-settings-backdrop"></div>
+      <div class="earth-settings-sheet liquid-glass-surface" role="dialog" aria-modal="true" aria-labelledby="settings-title">
+        <div class="earth-settings-header">
+          <div>
+            <div class="earth-settings-kicker">设置</div>
+            <h3 id="settings-title" class="earth-settings-title hud-panel-title">显示与视图</h3>
+          </div>
+          <button id="settings-close" class="earth-settings-close hud-panel-close" type="button" aria-label="关闭设置">
+            <span class="material-symbols-rounded">close</span>
+          </button>
+        </div>
+        <div class="earth-settings-content">
+          <section class="earth-settings-section">
+            <div class="earth-settings-section-title">视图</div>
+            <div class="earth-settings-list">
+              <label class="earth-settings-item" for="toggle-view-layers">
+                <div class="earth-settings-copy">
+                  <span class="earth-settings-item-title">图层控制</span>
+                  <span class="earth-settings-item-subtitle">控制右侧图层控制面板显示</span>
+                </div>
+                <span class="earth-settings-switch">
+                  <input id="toggle-view-layers" type="checkbox" data-settings-panel="layer-toggles" checked>
+                  <span class="earth-settings-switch-track"></span>
+                </span>
+              </label>
+              <label class="earth-settings-item" for="toggle-view-legend">
+                <div class="earth-settings-copy">
+                  <span class="earth-settings-item-title">图例</span>
+                  <span class="earth-settings-item-subtitle">控制左下角图例面板显示</span>
+                </div>
+                <span class="earth-settings-switch">
+                  <input id="toggle-view-legend" type="checkbox" data-settings-panel="legend">
+                  <span class="earth-settings-switch-track"></span>
+                </span>
+              </label>
+              <label class="earth-settings-item" for="toggle-view-stats">
+                <div class="earth-settings-copy">
+                  <span class="earth-settings-item-title">全球态势</span>
+                  <span class="earth-settings-item-subtitle">控制右上角全球态势统计面板显示</span>
+                </div>
+                <span class="earth-settings-switch">
+                  <input id="toggle-view-stats" type="checkbox" data-settings-panel="earth-stats">
+                  <span class="earth-settings-switch-track"></span>
+                </span>
+              </label>
+              <label class="earth-settings-item" for="toggle-view-tv">
+                <div class="earth-settings-copy">
+                  <span class="earth-settings-item-title">电视直播</span>
+                  <span class="earth-settings-item-subtitle">控制新闻直播窗口显示</span>
+                </div>
+                <span class="earth-settings-switch">
+                  <input id="toggle-view-tv" type="checkbox" data-settings-panel="tv-panel">
+                  <span class="earth-settings-switch-track"></span>
+                </span>
+              </label>
+            </div>
+          </section>
+        </div>
+      </div>
     </div>
-    <div id="status-message" class="status-message" style="display: none;"></div>
-    <div id="tooltip" class="tooltip"></div>
   </div>
 
   <script type="module" src="js/main.js"></script>

frontend/public/earth/js/brand.js

@@ -0,0 +1,57 @@
+const DEFAULT_BRAND_LANGUAGE = "zh";
+
+const BRANDS = {
+  zh: {
+    ariaLabel: "智能星球计划品牌标识",
+    titleAlt: "智能星球计划",
+    titleSrc: "assets/brand/title-zh.png",
+    subtitle: "现实层宇宙全息感知系统",
+    description: "卫星 · 海底光缆 · 算力基础设施",
+  },
+  en: {
+    ariaLabel: "Intelligent Planet Program brand banner",
+    titleAlt: "Intelligent Planet Program",
+    titleSrc: "assets/brand/title-en.png",
+    subtitle: "Physical-Universe Holography",
+    description: "Satellites · Cables · Compute Infra",
+  },
+};
+
+function getBrandConfig(variant = DEFAULT_BRAND_LANGUAGE) {
+  return BRANDS[variant] ?? BRANDS[DEFAULT_BRAND_LANGUAGE];
+}
+
+export function renderBrand(variant = DEFAULT_BRAND_LANGUAGE) {
+  const config = getBrandConfig(variant);
+
+  return `
+    <div class="earth-brand earth-brand--${variant}" aria-label="${config.ariaLabel}">
+      <img
+        class="earth-brand__logo"
+        src="assets/brand/earth-logo.png"
+        alt=""
+        aria-hidden="true"
+        fetchpriority="high"
+        loading="eager"
+      >
+      <div class="earth-brand__copy">
+        <img
+          class="earth-brand__title"
+          src="${config.titleSrc}"
+          alt="${config.titleAlt}"
+          fetchpriority="high"
+          loading="eager"
+        >
+        <div class="earth-brand__meta">
+          <span class="earth-brand__subtitle">${config.subtitle}</span>
+          <span class="earth-brand__description">${config.description}</span>
+        </div>
+      </div>
+    </div>
+  `.trim();
+}
+
+export function mountBrand(target, variant = DEFAULT_BRAND_LANGUAGE) {
+  if (!target) return;
+  target.innerHTML = renderBrand(variant);
+}

frontend/public/earth/js/cables.js

@@ -327,7 +327,9 @@ export async function loadGeoJSONFromPath(scene, earthObj) {
   const cableCount = data.features.length;
   const inServiceCount = data.features.filter(
     (feature) =>
-      feature.properties && feature.properties.status === "In Service",
+      feature.properties &&
+      (feature.properties.status === "active" ||
+        feature.properties.status === "In Service"),
   ).length;
 
   const cableCountEl = document.getElementById("cable-count");

frontend/public/earth/js/constants.js

@@ -12,6 +12,14 @@ export const CONFIG = {
   dragRotationScaleMax: 2.0,
 };
 
+export const HUD_CONFIG = {
+  scaleReferenceWidth: 1920,
+  scaleReferenceHeight: 1080,
+  minScale: 0.7,
+  maxScale: 1,
+  brandLanguage: "zh",
+};
+
 // Earth coordinate constants
 export const EARTH_CONFIG = {
   tilt: 23.5,  // earth tilt angle (degrees)
@@ -214,3 +222,38 @@ export const GRID_CONFIG = {
   longitudeStep: 30,
   gridStep: 5
 };
+
+export const EARTH_MATERIAL_CONFIG = {
+  // Diffuse color multiplies with texture — pure white = full saturation,
+  // slightly grey-blue pulls perceived saturation down without a custom shader.
+  color: 0xcdd8e6,
+  specular: 0x1a2d45,
+  shininess: 12,
+  emissive: 0x050a12,
+  opacity: 0.96,
+
+  // Depth-mask occluder keeps far-side objects hidden behind the earth
+  occluderRadiusFactor: 0.999,
+  occluderSegments: 48,
+
+  // Fresnel atmosphere glow — inner rim
+  atmosInnerRadiusFactor: 1.018,
+  atmosInnerSegments: 64,
+  atmosInnerColor: [0.25, 0.62, 1.0],
+  atmosInnerRimPower: 3.2,
+  atmosInnerIntensity: 0.72,
+
+  // Fresnel atmosphere glow — outer corona
+  atmosOuterRadiusFactor: 1.07,
+  atmosOuterSegments: 48,
+  atmosOuterColor: [0.18, 0.45, 0.9],
+  atmosOuterRimPower: 5.0,
+  atmosOuterIntensity: 0.28,
+
+  // Texture candidates — tried in order, first success wins
+  textureUrls: [
+    './assets/8k_earth_daymap.jpg',
+    'https://raw.githubusercontent.com/mrdoob/three.js/dev/examples/textures/planets/earth_atmos_2048.jpg',
+    'https://threejs.org/examples/textures/planets/earth_atmos_2048.jpg',
+  ],
+};

frontend/public/earth/js/controls.js

@@ -17,6 +17,7 @@ import {
 } from "./satellites.js";
 import { getShowCables } from "./cables.js";
 import { toggleBGP, getShowBGP, getBGPCount } from "./bgp.js";
+import { ensureTVPanelReady } from "./tv.js";
 
 export let autoRotate = true;
 export let zoomLevel = 1.0;
@@ -26,11 +27,18 @@ export let layoutExpanded = false;
 let earthObj = null;
 let listeners = [];
 let cleanupFns = [];
+const HUD_PANEL_IDS = [
+  "legend",
+  "earth-stats",
+  "tv-panel",
+  "layer-toggles",
+];
+const DRAGGABLE_PANEL_SELECTOR = ".hud-panel-draggable";
+const PANEL_LAYOUT_ANIMATION_MS = 420;
 
 function getFloatingGroups() {
   return [
     document.getElementById("zoom-control-group"),
-    document.getElementById("info-control-group"),
   ].filter(Boolean);
 }
 
@@ -44,6 +52,12 @@ function isFloatingMenuVisible() {
   });
 }
 
+function isSettingsModalOpen() {
+  return document
+    .getElementById("settings-modal")
+    ?.classList.contains("is-open");
+}
+
 function closeFloatingMenus() {
   getFloatingGroups().forEach((group) => {
     group.classList.remove("open");
@@ -55,6 +69,184 @@ function closeFloatingMenus() {
   }
 }
 
+function openSettingsModal() {
+  const modal = document.getElementById("settings-modal");
+  if (!modal) return;
+  closeFloatingMenus();
+  modal.classList.add("is-open");
+  modal.setAttribute("aria-hidden", "false");
+}
+
+function closeSettingsModal() {
+  const modal = document.getElementById("settings-modal");
+  if (!modal) return;
+  modal.classList.remove("is-open");
+  modal.setAttribute("aria-hidden", "true");
+}
+
+function setHudPanelVisibility(panelId, visible) {
+  const panel = document.getElementById(panelId);
+  if (!panel) return;
+  panel.classList.toggle("hud-panel-hidden", !visible);
+  syncSettingsToggle(panelId, visible);
+  if (panelId === "tv-panel") {
+    updateTVToggleUI(visible);
+    if (visible) {
+      ensureTVPanelReady().catch((error) => {
+        console.error("初始化电视直播面板失败:", error);
+      });
+    }
+  }
+}
+
+function syncSettingsToggle(panelId, visible) {
+  const input = document.querySelector(
+    `[data-settings-panel="${panelId}"]`,
+  );
+  if (input instanceof HTMLInputElement) {
+    input.checked = visible;
+  }
+}
+
+function syncAllHudPanelToggles() {
+  HUD_PANEL_IDS.forEach((panelId) => {
+    const panel = document.getElementById(panelId);
+    syncSettingsToggle(panelId, !panel?.classList.contains("hud-panel-hidden"));
+  });
+}
+
+function setupSettingsControls() {
+  const settingsTrigger = document.getElementById("settings-trigger");
+  const settingsClose = document.getElementById("settings-close");
+  const settingsBackdrop = document.getElementById("settings-backdrop");
+  const settingsModal = document.getElementById("settings-modal");
+
+  bindListener(settingsTrigger, "click", () => {
+    openSettingsModal();
+  });
+
+  bindListener(settingsClose, "click", () => {
+    closeSettingsModal();
+  });
+
+  bindListener(settingsBackdrop, "click", () => {
+    closeSettingsModal();
+  });
+
+  bindListener(settingsModal, "click", (event) => {
+    const sheet = event.target.closest(".earth-settings-sheet");
+    if (!sheet) {
+      closeSettingsModal();
+    }
+  });
+
+  const toggleInputs = document.querySelectorAll("[data-settings-panel]");
+  toggleInputs.forEach((input) => {
+    bindListener(input, "change", (event) => {
+      const target = event.currentTarget;
+      if (!(target instanceof HTMLInputElement)) return;
+      const panelId = target.dataset.settingsPanel;
+      if (!panelId) return;
+      setHudPanelVisibility(panelId, target.checked);
+    });
+  });
+
+  syncAllHudPanelToggles();
+}
+
+function setupHudPanelControls() {
+  const closeButtons = document.querySelectorAll("[data-close-panel]");
+  closeButtons.forEach((button) => {
+    bindListener(button, "click", (event) => {
+      event.stopPropagation();
+      const target = event.currentTarget;
+      if (!(target instanceof HTMLElement)) return;
+      const panelId = target.dataset.closePanel;
+      if (!panelId) return;
+      setHudPanelVisibility(panelId, false);
+    });
+  });
+}
+
+function setupDraggableHudPanels() {
+  const app = document.getElementById("container");
+  const draggablePanels = document.querySelectorAll(DRAGGABLE_PANEL_SELECTOR);
+  if (!app || draggablePanels.length === 0) return;
+
+  draggablePanels.forEach((panel) => {
+    const handle = panel.querySelector(".hud-panel-drag-handle");
+    if (!handle) return;
+
+    let isDragging = false;
+    let startPointerX = 0;
+    let startPointerY = 0;
+    let startLeft = 0;
+    let startTop = 0;
+
+    const stopDragging = () => {
+      isDragging = false;
+      panel.classList.remove("is-dragging");
+      document.body.style.userSelect = "";
+    };
+
+    const onMove = (event) => {
+      if (!isDragging) return;
+      const appRect = app.getBoundingClientRect();
+      const panelRect = panel.getBoundingClientRect();
+      const nextLeft = Math.min(
+        Math.max(startLeft + (event.clientX - startPointerX), 0),
+        appRect.width - panelRect.width,
+      );
+      const nextTop = Math.min(
+        Math.max(startTop + (event.clientY - startPointerY), 0),
+        appRect.height - panelRect.height,
+      );
+
+      panel.style.left = `${nextLeft}px`;
+      panel.style.top = `${nextTop}px`;
+      panel.style.right = "auto";
+      panel.style.bottom = "auto";
+      panel.style.transform = "none";
+      panel.dataset.dragged = "true";
+    };
+
+    bindListener(handle, "pointerdown", (event) => {
+      if (event.target.closest(".hud-panel-close, .layer-panel-btn, .info-card-close")) return;
+      isDragging = true;
+      startPointerX = event.clientX;
+      startPointerY = event.clientY;
+      const appRect = app.getBoundingClientRect();
+      const panelRect = panel.getBoundingClientRect();
+
+      // If panel is inside a flow container (not a direct child of app), reparent
+      // it so absolute positioning is relative to the app container.
+      if (panel.parentElement !== app) {
+        const capturedWidth = panelRect.width;
+        panel.style.position = "absolute";
+        panel.style.width = `${capturedWidth}px`;
+        app.appendChild(panel);
+      }
+
+      startLeft = panelRect.left - appRect.left;
+      startTop = panelRect.top - appRect.top;
+      panel.style.left = `${startLeft}px`;
+      panel.style.top = `${startTop}px`;
+      panel.style.right = "auto";
+      panel.style.bottom = "auto";
+      panel.style.transform = "none";
+      panel.dataset.dragged = "true";
+      panel.classList.add("is-dragging");
+      document.body.style.userSelect = "none";
+      handle.setPointerCapture?.(event.pointerId);
+    });
+
+    bindListener(handle, "pointermove", onMove);
+    bindListener(handle, "pointerup", stopDragging);
+    bindListener(handle, "pointercancel", stopDragging);
+    bindListener(handle, "lostpointercapture", stopDragging);
+  });
+}
+
 function clearForcedFloatingClose() {
   getFloatingGroups().forEach((group) => {
     group.classList.remove("force-closed");
@@ -76,12 +268,23 @@ function setFloatingMenuOpen(group, shouldOpen) {
 }
 
 function setButtonTooltip(button, text) {
-  const tooltip = button?.querySelector(".tooltip");
+  const tooltip = button?.querySelector(".earth-toolbar-tooltip");
   if (tooltip) {
     tooltip.textContent = text;
   }
 }
 
+export function updateLayerButtonState(button, isActive) {
+  if (!button) return;
+  button.classList.toggle("active", isActive);
+  button.setAttribute("aria-checked", isActive ? "true" : "false");
+  // Legacy badge text (kept for compatibility)
+  const state = button.querySelector(".earth-layer-btn__state");
+  if (state) {
+    state.textContent = isActive ? "ON" : "OFF";
+  }
+}
+
 function clearSelectionIfHiding(shouldHide) {
   if (shouldHide) {
     clearLockedObject();
@@ -100,6 +303,13 @@ function bindFloatingMenu(trigger, group) {
   });
 }
 
+function updateTVToggleUI(visible) {
+  const btn = document.getElementById("toggle-tv");
+  if (!btn) return;
+  btn.classList.toggle("active", visible);
+  setButtonTooltip(btn, visible ? "关闭新闻直播" : "打开新闻直播");
+}
+
 function bindListener(element, eventName, handler, options) {
   if (!element) return;
   element.addEventListener(eventName, handler, options);
@@ -365,11 +575,91 @@ function setupRotateControls(camera) {
   });
 }
 
+function filterLayerRows(query, emptyStateEl, clearBtn) {
+  const rows = document.querySelectorAll("#layer-panel-list .layer-row");
+  let visibleCount = 0;
+  rows.forEach((row) => {
+    const name = (row.dataset.layerName || "").toLowerCase();
+    const matches = !query || name.includes(query);
+    row.hidden = !matches;
+    if (matches) visibleCount++;
+  });
+  if (emptyStateEl) emptyStateEl.hidden = visibleCount > 0;
+  if (clearBtn) clearBtn.hidden = !query;
+}
+
+function setupLayerPanel() {
+  const panel = document.getElementById("layer-toggles");
+  const collapseBtn = document.getElementById("layer-panel-collapse");
+  const searchInput = document.getElementById("layer-search-input");
+  const searchClear = document.getElementById("layer-search-clear");
+  const emptyState = document.getElementById("layer-panel-empty");
+  if (!panel) return;
+
+  bindListener(collapseBtn, "click", (e) => {
+    e.stopPropagation();
+    const isCollapsed = panel.classList.toggle("layer-panel--collapsed");
+    collapseBtn.title = isCollapsed ? "展开" : "折叠";
+    collapseBtn.setAttribute("aria-label", isCollapsed ? "展开图层列表" : "折叠图层列表");
+    const icon = collapseBtn.querySelector(".material-symbols-rounded");
+    if (icon) icon.textContent = isCollapsed ? "expand_less" : "expand_more";
+  });
+
+  if (searchInput) {
+    bindListener(searchInput, "input", () => {
+      const query = searchInput.value.trim().toLowerCase();
+      filterLayerRows(query, emptyState, searchClear);
+    });
+
+    if (searchClear) {
+      bindListener(searchClear, "click", () => {
+        searchInput.value = "";
+        filterLayerRows("", emptyState, searchClear);
+        searchInput.focus();
+      });
+    }
+  }
+}
+
+function createLayerRow({ id, icon, label, meta, defaultActive }) {
+  const row = document.createElement("div");
+  row.className = "layer-row";
+  row.dataset.layerName = `${label} ${meta || ""}`.trim().toLowerCase();
+  row.innerHTML = `
+    <span class="material-symbols-rounded layer-row-icon">${icon}</span>
+    <div class="layer-row-copy">
+      <span class="layer-row-label">${label}</span>
+      ${meta ? `<span class="layer-row-meta">${meta}</span>` : ""}
+    </div>
+    <button id="${id}" class="layer-row-toggle${defaultActive ? " active" : ""}" type="button"
+      role="switch" aria-checked="${defaultActive ? "true" : "false"}" title="切换${label}显示">
+      <span class="layer-row-toggle-track"></span>
+    </button>
+  `;
+  return row;
+}
+
+export function registerLayer({ id, icon, label, meta = "", defaultActive = false, onToggle }) {
+  const list = document.getElementById("layer-panel-list");
+  if (!list) return;
+  if (document.getElementById(id)) return; // avoid duplicates
+
+  const row = createLayerRow({ id, icon, label, meta, defaultActive });
+  list.appendChild(row);
+
+  const btn = row.querySelector("button");
+  if (btn && typeof onToggle === "function") {
+    btn.addEventListener("click", function () {
+      const isActive = this.classList.toggle("active");
+      this.setAttribute("aria-checked", isActive ? "true" : "false");
+      onToggle(isActive);
+    });
+  }
+}
+
 function setupTerrainControls() {
   const container = document.getElementById("container");
   const searchBtn = document.getElementById("search-action");
-  const infoGroup = document.getElementById("info-control-group");
-  const infoTrigger = document.getElementById("info-trigger");
   const terrainBtn = document.getElementById("toggle-terrain");
   const satellitesBtn = document.getElementById("toggle-satellites");
   const bgpBtn = document.getElementById("toggle-bgp");
@@ -379,9 +669,13 @@ function setupTerrainControls() {
   const reloadBtn = document.getElementById("reload-data");
   const zoomGroup = document.getElementById("zoom-control-group");
   const zoomTrigger = document.getElementById("zoom-trigger");
+  setupSettingsControls();
+  setupHudPanelControls();
+  setupDraggableHudPanels();
+  setupLayerPanel();
 
   if (trailsBtn) {
-    trailsBtn.classList.add("active");
+    updateLayerButtonState(trailsBtn, true);
     setButtonTooltip(trailsBtn, "隐藏轨迹");
   }
 
@@ -392,7 +686,7 @@ function setupTerrainControls() {
   bindListener(terrainBtn, "click", function () {
     showTerrain = !showTerrain;
     toggleTerrain(showTerrain);
-    this.classList.toggle("active", showTerrain);
+    updateLayerButtonState(this, showTerrain);
     setButtonTooltip(this, showTerrain ? "隐藏地形" : "显示地形");
     const terrainStatus = document.getElementById("terrain-status");
     if (terrainStatus)
@@ -422,7 +716,7 @@ function setupTerrainControls() {
     const showNextBGP = !getShowBGP();
     clearSelectionIfHiding(!showNextBGP);
     toggleBGP(showNextBGP);
-    this.classList.toggle("active", showNextBGP);
+    updateLayerButtonState(this, showNextBGP);
     setButtonTooltip(this, showNextBGP ? "隐藏BGP观测" : "显示BGP观测");
     const bgpCountEl = document.getElementById("bgp-anomaly-count");
     if (bgpCountEl) {
@@ -435,7 +729,7 @@ function setupTerrainControls() {
     const isActive = this.classList.contains("active");
     const nextShowTrails = !isActive;
     toggleTrails(nextShowTrails);
-    this.classList.toggle("active", nextShowTrails);
+    updateLayerButtonState(this, nextShowTrails);
     setButtonTooltip(this, nextShowTrails ? "隐藏轨迹" : "显示轨迹");
     showStatusMessage(nextShowTrails ? "轨迹已显示" : "轨迹已隐藏", "info");
   });
@@ -455,10 +749,9 @@ function setupTerrainControls() {
   });
 
   bindFloatingMenu(zoomTrigger, zoomGroup);
-  bindFloatingMenu(infoTrigger, infoGroup);
 
   bindListener(document, "click", (event) => {
-    const openGroups = [zoomGroup, infoGroup].filter((group) =>
+    const openGroups = [zoomGroup].filter((group) =>
       group?.classList.contains("open"),
     );
     if (openGroups.length === 0) return;
@@ -480,6 +773,13 @@ function setupTerrainControls() {
     showStatusMessage(expanded ? "布局已最大化" : "布局已恢复", "info");
   });
 
+  const tvVisible = !document.getElementById("tv-panel")?.classList.contains("hud-panel-hidden");
+  updateTVToggleUI(tvVisible);
+  if (tvVisible) {
+    ensureTVPanelReady().catch((error) => {
+      console.error("初始化电视直播面板失败:", error);
+    });
+  }
   updateLayoutUI(container);
 }
 
@@ -487,6 +787,11 @@ function setupKeyboardControls() {
   bindListener(document, "keydown", (event) => {
     if (event.key !== "Escape") return;
 
+    if (isSettingsModalOpen()) {
+      closeSettingsModal();
+      return;
+    }
+
     if (isFloatingMenuVisible()) {
       closeFloatingMenus();
       return;
@@ -566,7 +871,7 @@ function updateRotateUI() {
   if (btn) {
     btn.classList.toggle("active", autoRotate);
     btn.classList.toggle("is-stopped", !autoRotate);
-    const tooltip = btn.querySelector(".tooltip");
+    const tooltip = btn.querySelector(".earth-toolbar-tooltip");
     if (tooltip) tooltip.textContent = autoRotate ? "暂停旋转" : "开始旋转";
   }
 }
@@ -599,15 +904,81 @@ function updateLayoutUI(container) {
   const btn = document.getElementById("layout-toggle");
   if (btn) {
     btn.classList.toggle("active", layoutExpanded);
-    const tooltip = btn.querySelector(".tooltip");
+    const tooltip = btn.querySelector(".earth-toolbar-tooltip");
     const nextLabel = layoutExpanded ? "恢复布局" : "最大化布局";
     btn.title = nextLabel;
     if (tooltip) tooltip.textContent = nextLabel;
   }
 }
 
-function toggleLayoutExpanded(container) {
-  layoutExpanded = !layoutExpanded;
-  updateLayoutUI(container);
-  return layoutExpanded;
+function resetPanelInlineLayout(panel) {
+  panel.style.left = "";
+  panel.style.top = "";
+  panel.style.right = "";
+  panel.style.bottom = "";
+  panel.style.transform = "";
+  delete panel.dataset.dragged;
+}
+
+function isPanelVisible(panel) {
+  return !panel.classList.contains("hud-panel-hidden");
+}
+
+function animatePanelLayoutTransition(container, expand) {
+  const panels = Array.from(
+    container.querySelectorAll(DRAGGABLE_PANEL_SELECTOR),
+  );
+  if (panels.length === 0) {
+    layoutExpanded = expand;
+    updateLayoutUI(container);
+    return expand;
+  }
+
+  const visiblePanels = panels.filter(isPanelVisible);
+  const firstRects = new Map(
+    visiblePanels.map((panel) => [panel, panel.getBoundingClientRect()]),
+  );
+
+  panels.forEach((panel) => panel.classList.add("is-layout-animating"));
+  layoutExpanded = expand;
+  panels.forEach(resetPanelInlineLayout);
+  updateLayoutUI(container);
+
+  visiblePanels.forEach((panel) => {
+    const firstRect = firstRects.get(panel);
+    if (!firstRect) return;
+
+    const lastRect = panel.getBoundingClientRect();
+    const deltaX = firstRect.left - lastRect.left;
+    const deltaY = firstRect.top - lastRect.top;
+
+    if (Math.abs(deltaX) < 0.5 && Math.abs(deltaY) < 0.5) {
+      return;
+    }
+
+    panel.animate(
+      [
+        {
+          translate: `${deltaX}px ${deltaY}px`,
+        },
+        {
+          translate: "0 0",
+        },
+      ],
+      {
+        duration: PANEL_LAYOUT_ANIMATION_MS,
+        easing: "cubic-bezier(0.22, 1, 0.36, 1)",
+      },
+    );
+  });
+
+  window.setTimeout(() => {
+    panels.forEach((panel) => panel.classList.remove("is-layout-animating"));
+  }, PANEL_LAYOUT_ANIMATION_MS);
+
+  return expand;
+}
+
+function toggleLayoutExpanded(container) {
+  return animatePanelLayoutTransition(container, !layoutExpanded);
 }

frontend/public/earth/js/earth.js

@@ -1,7 +1,7 @@
 // earth.js - 3D Earth creation module
 
 import * as THREE from 'three';
-import { CONFIG, EARTH_CONFIG } from './constants.js';
+import { CONFIG, EARTH_CONFIG, EARTH_MATERIAL_CONFIG } from './constants.js';
 import { latLonToVector3 } from './utils.js';
 
 export let earth = null;
@@ -9,82 +9,108 @@ export let clouds = null;
 export let terrain = null;
 
 const textureLoader = new THREE.TextureLoader();
+let _earthMaterial = null;
 
 export function createEarth(scene) {
   const geometry = new THREE.SphereGeometry(CONFIG.earthRadius, 128, 128);
-  
+
+  const C = EARTH_MATERIAL_CONFIG;
+
   const material = new THREE.MeshPhongMaterial({
-    color: 0xffffff,
-    specular: 0x111111,
-    shininess: 10,
-    emissive: 0x000000,
+    color: C.color,
+    specular: C.specular,
+    shininess: C.shininess,
+    emissive: C.emissive,
     transparent: true,
-    opacity: 0.8,
-    side: THREE.DoubleSide
+    opacity: C.opacity,
+    side: THREE.FrontSide,
+    depthWrite: true,
+    depthTest: true,
   });
-   
+  _earthMaterial = material;
+
   earth = new THREE.Mesh(geometry, material);
+  earth.renderOrder = 0;
   earth.rotation.x = EARTH_CONFIG.tiltRad;
   scene.add(earth);
-  
-  const textureUrls = [
-    './assets/8k_earth_daymap.jpg',
-    'https://raw.githubusercontent.com/mrdoob/three.js/dev/examples/textures/planets/earth_atmos_2048.jpg',
-    'https://threejs.org/examples/textures/planets/earth_atmos_2048.jpg',
-    'https://assets.codepen.io/982762/earth_texture_2048.jpg'
-  ];
-  
-  let textureLoaded = false;
-  
-  textureLoader.load(
-    textureUrls[0],
-    function(texture) {
-      console.log('高分辨率地球纹理加载成功');
-      textureLoaded = true;
-      
-      texture.wrapS = THREE.RepeatWrapping;
-      texture.wrapT = THREE.ClampToEdgeWrapping;
-      texture.anisotropy = 16;
-      texture.minFilter = THREE.LinearMipmapLinearFilter;
-      texture.magFilter = THREE.LinearFilter;
-      
-      material.map = texture;
-      material.needsUpdate = true;
-      
-      document.getElementById('loading').style.display = 'none';
-    },
-    function(xhr) {
-      console.log('纹理加载中: ' + (xhr.loaded / xhr.total * 100) + '%');
-    },
-    function(err) {
-      console.log('第一个纹理加载失败，尝试第二个...');
-      
-      textureLoader.load(
-        textureUrls[1],
-        function(texture) {
-          console.log('第二个纹理加载成功');
-          textureLoaded = true;
-          
-          texture.wrapS = THREE.RepeatWrapping;
-          texture.wrapT = THREE.ClampToEdgeWrapping;
-          texture.anisotropy = 16;
-          texture.minFilter = THREE.LinearMipmapLinearFilter;
-          texture.magFilter = THREE.LinearFilter;
-          
-          material.map = texture;
-          material.needsUpdate = true;
-          
-          document.getElementById('loading').style.display = 'none';
-        },
-        null,
-        function(err) {
-          console.log('所有纹理加载失败');
-          document.getElementById('loading').style.display = 'none';
-        }
-      );
-    }
+
+  // Depth-mask occluder — invisible sphere slightly inside the earth,
+  // writes to the depth buffer so far-side cables/satellites are occluded.
+  const occluderGeometry = new THREE.SphereGeometry(
+    CONFIG.earthRadius * C.occluderRadiusFactor,
+    C.occluderSegments,
+    C.occluderSegments,
   );
-  
+  const occluderMaterial = new THREE.MeshBasicMaterial({
+    colorWrite: false,
+    side: THREE.FrontSide,
+  });
+  const occluder = new THREE.Mesh(occluderGeometry, occluderMaterial);
+  occluder.renderOrder = -1;
+  earth.add(occluder);
+
+  // Shared Fresnel vertex shader for both atmosphere layers
+  const ATMOS_VERTEX_SHADER = `
+    varying vec3 vNormal;
+    void main() {
+      vNormal = normalize(normalMatrix * normal);
+      gl_Position = projectionMatrix * modelViewMatrix * vec4(position, 1.0);
+    }
+  `;
+
+  // Fresnel atmosphere — inner rim
+  const [ir, ig, ib] = C.atmosInnerColor;
+  const atmosInnerGeo = new THREE.SphereGeometry(
+    CONFIG.earthRadius * C.atmosInnerRadiusFactor,
+    C.atmosInnerSegments,
+    C.atmosInnerSegments,
+  );
+  const atmosInnerMat = new THREE.ShaderMaterial({
+    vertexShader: ATMOS_VERTEX_SHADER,
+    fragmentShader: `
+      varying vec3 vNormal;
+      void main() {
+        float rim = 1.0 - abs(dot(vNormal, vec3(0.0, 0.0, 1.0)));
+        float intensity = pow(rim, ${C.atmosInnerRimPower.toFixed(1)});
+        gl_FragColor = vec4(${ir.toFixed(2)}, ${ig.toFixed(2)}, ${ib.toFixed(2)}, intensity * ${C.atmosInnerIntensity.toFixed(2)});
+      }
+    `,
+    blending: THREE.AdditiveBlending,
+    side: THREE.BackSide,
+    transparent: true,
+    depthWrite: false,
+  });
+  const atmosInner = new THREE.Mesh(atmosInnerGeo, atmosInnerMat);
+  atmosInner.renderOrder = 1;
+  earth.add(atmosInner);
+
+  // Fresnel atmosphere — outer corona
+  const [outerR, outerG, outerB] = C.atmosOuterColor;
+  const atmosOuterGeo = new THREE.SphereGeometry(
+    CONFIG.earthRadius * C.atmosOuterRadiusFactor,
+    C.atmosOuterSegments,
+    C.atmosOuterSegments,
+  );
+  const atmosOuterMat = new THREE.ShaderMaterial({
+    vertexShader: ATMOS_VERTEX_SHADER,
+    fragmentShader: `
+      varying vec3 vNormal;
+      void main() {
+        float rim = 1.0 - abs(dot(vNormal, vec3(0.0, 0.0, 1.0)));
+        float intensity = pow(rim, ${C.atmosOuterRimPower.toFixed(1)});
+        gl_FragColor = vec4(${outerR.toFixed(2)}, ${outerG.toFixed(2)}, ${outerB.toFixed(2)}, intensity * ${C.atmosOuterIntensity.toFixed(2)});
+      }
+    `,
+    blending: THREE.AdditiveBlending,
+    side: THREE.BackSide,
+    transparent: true,
+    depthWrite: false,
+  });
+  const atmosOuter = new THREE.Mesh(atmosOuterGeo, atmosOuterMat);
+  atmosOuter.renderOrder = 1;
+  earth.add(atmosOuter);
+
+  // Texture is loaded separately via loadEarthTexture() for staged loading
   return earth;
 }
 
@@ -238,3 +264,40 @@ export function getEarth() {
 export function getClouds() {
   return clouds;
 }
+
+export function clearEarthTexture() {
+  if (!_earthMaterial) return;
+  _earthMaterial.map = null;
+  _earthMaterial.needsUpdate = true;
+}
+
+export function loadEarthTexture() {
+  return new Promise((resolve) => {
+    if (!_earthMaterial) { resolve(); return; }
+
+    const urls = EARTH_MATERIAL_CONFIG.textureUrls;
+    const tryLoad = (index) => {
+      if (index >= urls.length) {
+        console.warn('所有地球纹理加载失败');
+        resolve();
+        return;
+      }
+      textureLoader.load(
+        urls[index],
+        (texture) => {
+          texture.wrapS = THREE.RepeatWrapping;
+          texture.wrapT = THREE.ClampToEdgeWrapping;
+          texture.anisotropy = 16;
+          texture.minFilter = THREE.LinearMipmapLinearFilter;
+          texture.magFilter = THREE.LinearFilter;
+          _earthMaterial.map = texture;
+          _earthMaterial.needsUpdate = true;
+          resolve();
+        },
+        null,
+        () => tryLoad(index + 1),
+      );
+    };
+    tryLoad(0);
+  });
+}