release: bump version to 0.40.1

- Add earth-mobile-drawer-ui-plan documenting mobile drawer UX decisions
- Add goal-driven.md Claude Code command for autonomous task execution
- Add .codex/ config with OpenAI model definitions and goal-driven agent
- Add SKILL.md, openai.yaml, and prompt-template for Codex integration

.claude/commands/cleanup.md

@@ -0,0 +1,120 @@
+---
+description: 审查当前工作区未提交代码中的垃圾代码，并在不影响逻辑的前提下自动清理
+argument-hint: 可选：指定要检查的文件或目录（默认检查所有未提交修改）
+allowed-tools: ["Read", "Edit", "Bash", "Grep", "Glob"]
+---
+
+# /cleanup — 垃圾代码审查与清理
+
+分析当前工作区（git diff）中的未提交代码，找出并修复常见垃圾代码，**不得改变任何运行逻辑**。
+
+## 检查范围
+
+若 `$ARGUMENTS` 非空，则只检查指定文件/目录；否则检查所有未提交修改（`git diff HEAD`）。
+
+## 审查清单
+
+按优先级检查以下问题（只报告在本次 diff 中**新增或修改**的代码里存在的问题）：
+
+### 1. 重复逻辑 (Duplicate Logic)
+- 完全相同或高度相似的代码块在多处出现
+- 同一函数/方法被多个地方各自实现，已有公共版本未被复用
+- 相同的 DOM 查询、正则、模板字符串在同一文件重复
+
+### 2. Magic Numbers / Magic Strings
+- 裸数字直接参与计算（如偏移量、时间、尺寸、阈值），没有命名常量
+- 硬编码字符串（如 id 名、状态值、URL 片段）散落在逻辑中
+- 例外：`0`, `1`, `-1`, `100`, `""` 等语义明确的惯用值不算
+
+### 3. 命名问题
+- 含义不明的缩写变量（如 `or_`, `tmp2`, `x2`）
+- 命名与实际用途不符
+- 同一概念在不同地方用不同名字表达
+
+### 4. 死代码 / 无效代码
+- 注释掉的旧代码块（3行以上）
+- 声明后从未使用的变量/参数/导入
+- 永远不会执行的条件分支
+
+### 5. 代码风格问题
+- 尾部空白字符（trailing whitespace）
+- 同一文件内风格不一致（如混用单双引号、缩进不统一）
+- 空行使用不一致（连续多个空行等）
+
+### 6. 其他常见问题
+- 私有辅助函数应被 export 但没有，导致调用方重复实现
+- 类型/接口重复定义
+- 过于冗长的条件表达式可以简化（不改逻辑）
+
+## 执行步骤
+
+### Step 1 — 获取待检查文件列表
+
+```bash
+# 无参数时：获取所有未提交修改
+git diff HEAD --name-only
+
+# 有参数时：用 $ARGUMENTS 过滤
+```
+
+### Step 2 — 逐文件阅读并分析
+
+- 用 Read 工具读取完整文件（不只读 diff）
+- 对照审查清单，记录每个问题：文件名、行号、问题类型、建议修复方式
+
+### Step 3 — 报告问题清单
+
+在修改前，先以列表形式输出所有发现的问题：
+
+```
+发现 N 个问题：
+
+[文件] js/foo.js
+  · L34, L78: 重复逻辑 — 两处都实现了相同的 DOM 查询，可提取到 getPanel()
+  · L91: Magic number — 硬编码 14 作为偏移量，应命名为 TOOLTIP_OFFSET
+
+[文件] js/bar.js
+  · L12: 命名问题 — 变量 `or_` 语义不明，应命名为 outerR/outerG/outerB
+  ...
+```
+
+如果没有发现问题，直接输出"未发现垃圾代码，当前代码质量良好。"并停止。
+
+### Step 4 — 执行修复
+
+对每个问题，使用 Edit 工具进行**最小化修改**：
+
+- **重复逻辑**：提取为共享常量/函数，更新所有调用点
+- **Magic number**：在文件顶部或逻辑附近声明 `const NAME = value`，替换所有引用
+- **命名问题**：重命名变量，更新所有使用处
+- **死代码**：直接删除
+- **尾部空白/风格**：修正
+- **未 export 的函数**：添加 `export`，在调用方改为导入（不重复实现）
+
+**修复原则：**
+- 只改在审查清单中发现的问题，不做额外优化
+- 每次 Edit 只修改确实有问题的行，保持 diff 最小
+- 改完后用 `grep` 验证旧的坏代码已消失
+
+### Step 5 — 输出总结
+
+```
+清理完成：
+
+修复了 N 个问题：
+  ✓ earth.js — 提取重复 vertexShader 为 ATMOS_VERTEX_SHADER 常量
+  ✓ main.js — 提取 TOOLTIP_CURSOR_OFFSET = 14（4处引用）
+  ✓ controls.js — export updateLayerButtonState，移除 main.js 中的重复实现
+  ...
+
+未修改的问题（需人工确认）：
+  ! foo.js L45 — 注释代码块较长，建议手动确认是否可删除
+```
+
+## 约束
+
+- **禁止**改变函数签名、接口定义、导出 API（除非问题正是私有函数应被 export）
+- **禁止**添加新功能、新抽象、新参数
+- **禁止**修改注释内容（只删除注释掉的死代码）
+- **禁止**修改测试文件逻辑
+- 如果一个 Magic number 的语义不完全确定，**跳过**，在总结中标记为"需人工确认"

.claude/commands/goal-driven.md

@@ -0,0 +1,91 @@
+---
+description: 用 goal-driven 方法推动一个复杂任务持续执行，直到明确成功标准被满足
+argument-hint: 建议填写任务目标；若同时给出成功标准更好
+allowed-tools: ["Read", "Edit", "Bash", "Grep", "Glob"]
+---
+
+# /goal-driven — 目标驱动执行模式
+
+使用 `lidangzzz/goal-driven` 的核心思想来推进复杂任务：先固定目标与成功标准，再持续执行和反复验收，直到标准真正满足。
+
+适用场景：
+
+- 长周期实现任务
+- 高复杂度工程任务
+- 可被明确验收的研究、实现、迁移、验证类工作
+
+不适用场景：
+
+- 纯脑暴
+- 无法定义成功标准的模糊任务
+- 很小的一次性修改
+
+## 输入要求
+
+若 `$ARGUMENTS` 只包含目标，没有成功标准，先补全一版可执行的成功标准再开始。
+
+启动时先输出：
+
+```md
+Goal
+- ...
+
+Criteria for success
+- ...
+
+Plan
+1. ...
+2. ...
+3. ...
+
+Verification
+- ...
+```
+
+## 执行规则
+
+1. 先把任务固化为两个核心块：
+   - `Goal`
+   - `Criteria for success`
+
+2. 成功标准必须尽量客观，可验证，可落地。
+   优先写成：
+   - 需要交付什么
+   - 需要通过哪些测试或验证
+   - 如何判断结果真的完成
+
+3. 进入持续执行循环：
+   - 完成一个阶段
+   - 检查当前结果是否满足成功标准
+   - 若未满足，明确剩余差距并继续推进
+
+4. 任何“完成了”“差不多了”“已实现”之类的结论，都必须经过验证，不能直接接受。
+
+5. 如果验证失败：
+   - 明确指出哪条成功标准没满足
+   - 继续工作，不要把阶段性进展误判为完成
+
+6. 只有在以下情况之一才能停止：
+   - 成功标准已满足
+   - 用户明确要求停止
+
+## 执行风格
+
+- 重证据，轻口头判断
+- 重验收，轻自我感觉
+- 优先用测试、日志、产物、对比结果来证明完成
+- 对长期任务保持“未达标就继续”的节奏
+
+## 简版模板
+
+```md
+Goal: [[[[[在此填写最终目标]]]]]
+
+Criteria for success: [[[[[在此填写成功标准]]]]]
+
+循环执行：
+1. 推进任务
+2. 检查是否满足成功标准
+3. 若未满足，继续工作
+4. 直到满足标准或用户明确停止
+```

.claude/commands/release.md

@@ -0,0 +1,146 @@
+---
+description: 发版工作流：根据变更类型决定版本号，更新所有版本文件和 changelog，运行验证，commit 并 push
+argument-hint: 可选：feature | bugfix | 或直接描述本次发布内容
+allowed-tools: ["Read", "Edit", "Bash", "Glob", "Grep"]
+---
+
+# /release — Planet 发版工作流
+
+## 版本号规则
+
+| 变更类型 | 版本跳动 | 适用场景 |
+|---------|---------|---------|
+| `feature` | `+0.1.0` | 纯新功能，无 bugfix |
+| `improvement` | `+0.0.1` | UI 调整、小功能增强、bugfix 混合，或以 UI/体验改进为主的迭代 |
+| `bugfix` | `+0.0.1` | 纯 bug 修复，无新功能 |
+| `docs` / `maintenance` / `refactor` | 默认不发版，除非用户明确要求 |
+
+意图混合时以用户明确描述为准；bugfix + 小 feature 混合默认判定为 `improvement`（`+0.0.1`）。
+
+## 必须同步更新的文件
+
+使用 `git rev-parse --show-toplevel` 获取仓库根目录，以下路径均相对于根目录：
+
+- `VERSION`
+- `frontend/package.json`（`"version"` 字段）
+- `pyproject.toml`（`version =` 字段）
+- `uv.lock`（**不要手动编辑**，通过 `uv lock` 重新生成）
+- `docs/CHANGELOG.md`
+- `docs/version-history.md`
+
+## 执行步骤
+
+### Step 1 — 环境检查
+
+```bash
+git branch --show-current          # 确认在 dev 分支
+git status --short                 # 检查是否有无关的未暂存修改
+cat VERSION                        # 读取当前版本
+```
+
+若当前**不在 `dev` 分支**，停下来告知用户，不要继续。
+
+若存在无关的未暂存修改，列出并询问用户是否一并提交，或先 stash。
+
+### Step 2 — 确定发版类型与新版本号
+
+- 若 `$ARGUMENTS` 提供了明确类型（`feature` / `bugfix`），直接使用
+- 否则根据当前 `git diff HEAD` 和 `git log` 推断
+- 计算新版本号（例：`0.26.2` → bugfix → `0.26.3`）
+- **先输出发版计划供用户确认**：
+
+```
+发版计划：
+  类型：bugfix
+  版本：0.26.2 → 0.26.3
+  分支：dev
+  将更新：VERSION, frontend/package.json, pyproject.toml, uv.lock, CHANGELOG.md, version-history.md
+```
+
+### Step 3 — 更新版本号文件
+
+按顺序更新（每步用 Edit 工具，精确替换，不要重写整个文件）：
+
+1. `VERSION` — 直接替换全部内容为新版本号
+2. `frontend/package.json` — 替换 `"version": "x.x.x"` 行
+3. `pyproject.toml` — 替换 `version = "x.x.x"` 行
+4. 运行 `uv lock` 重新生成 `uv.lock`（在仓库根目录下执行）
+
+### Step 4 — 更新 CHANGELOG.md
+
+在文件顶部插入新条目，格式：
+
+```markdown
+## [x.x.x] — YYYY-MM-DD
+
+### ✨ Features / 🐛 Fixes / 🔧 Improvements
+- ...（只列高信号条目，最多 5 条）
+- ...
+
+---
+```
+
+日期使用 `date +%Y-%m-%d` 获取今天的日期。
+
+### Step 5 — 更新 docs/version-history.md
+
+- 更新文件头部的"当前开发版本"字段
+- 在时间线表格顶部插入新行：`| vx.x.x | YYYY-MM-DD | 一句话摘要 |`
+
+### Step 6 — 验证
+
+针对本次变更范围做最小验证：
+
+- Python 文件有修改：`python3 -m py_compile <changed_files>`
+- Frontend 文件有修改：运行项目标准检查（若无则跳过并说明）
+- 版本号一致性检查：用 grep 确认 VERSION、package.json、pyproject.toml 中的版本号完全一致
+
+```bash
+grep -h "version" VERSION frontend/package.json pyproject.toml
+```
+
+### Step 7 — 提交前预览
+
+展示将要提交的文件列表：
+
+```bash
+git diff --stat HEAD
+```
+
+再次确认所有必须文件都在变更列表中，**不包含**非预期文件（如调试文件、.env 等）。
+
+### Step 8 — Commit & Push（用户确认后）
+
+```bash
+git add VERSION frontend/package.json pyproject.toml uv.lock docs/CHANGELOG.md docs/version-history.md
+# 若有代码变更也一并 stage
+git add <code_files>
+
+git commit -m "release: bump version to x.x.x"
+git tag vx.x.x
+git push origin dev
+git push origin vx.x.x
+```
+
+commit message 固定格式：`release: bump version to x.x.x`
+
+### Step 9 — 完成确认
+
+输出摘要：
+
+```
+✓ 版本号已更新：0.26.2 → 0.26.3
+✓ CHANGELOG 已更新
+✓ version-history 已更新
+✓ uv.lock 已重新生成
+✓ 验证通过
+✓ commit: release: bump version to 0.26.3
+✓ tag: v0.26.3
+✓ 已 push 到 origin/dev
+```
+
+## 注意事项
+
+- `uv.lock` 只能通过 `uv lock` 生成，绝不手动编辑
+- 发版 commit 只包含版本文件 + 本次功能代码，不混入无关改动
+- 若环境中 `uv` 不可用，说明原因并跳过 lockfile 更新，提醒用户手动运行

.codex/config.toml

@@ -0,0 +1,3 @@
+approval_policy = "never"
+
+sandbox_mode = "danger-full-access"

.codex/skills/cleanup/SKILL.md

@@ -0,0 +1,124 @@
+---
+name: cleanup
+description: Use when the user asks to clean up, lint, or review uncommitted code for common code smells — duplicate logic, magic numbers, unclear naming, dead code, style inconsistencies. Fixes issues without changing any runtime behavior.
+---
+
+# Cleanup
+
+Review and fix code quality issues in the current working tree without altering any logic or behavior.
+
+## When To Use
+
+- The user asks to clean up, tidy, or lint uncommitted changes
+- The user wants a code smell review before releasing or committing
+- The user mentions magic numbers, duplicate logic, dead code, or naming issues
+
+Do not refactor architecture, add features, or change behavior.
+
+## Scope
+
+If the user specifies a file or directory, check only that. Otherwise check all uncommitted changes (`git diff HEAD`).
+
+Only report issues present in **newly added or modified** lines of this diff — do not audit unchanged code.
+
+## Checklist
+
+### 1. Duplicate Logic
+- Identical or near-identical code blocks appearing in multiple places
+- A function/helper that already exists but is re-implemented elsewhere instead of being reused
+- Repeated DOM queries, regex literals, or template strings within the same file
+
+### 2. Magic Numbers / Magic Strings
+- Bare numeric literals used in calculations (offsets, timeouts, sizes, thresholds) without a named constant
+- Hardcoded strings (IDs, status values, URL fragments) scattered through logic
+- Exceptions: `0`, `1`, `-1`, `100`, `""` and other idiomatically clear values are fine
+
+### 3. Naming Issues
+- Cryptic abbreviations (`or_`, `tmp2`, `x2`)
+- Names that do not match actual behavior
+- The same concept referred to by different names in different places
+
+### 4. Dead Code
+- Commented-out code blocks (3+ lines)
+- Variables, parameters, or imports declared but never used
+- Branches that can never execute
+
+### 5. Style Inconsistencies
+- Trailing whitespace
+- Mixed quote styles or indentation within the same file
+- Inconsistent blank-line usage (multiple consecutive blank lines, etc.)
+
+### 6. Other
+- Private helper functions that should be exported but are not, causing callers to duplicate the implementation
+- Overly verbose conditions that can be simplified without changing logic
+
+## Steps
+
+### Step 1 — Get the file list
+
+```bash
+git diff HEAD --name-only
+```
+
+Filter to the user-specified path if one was provided.
+
+### Step 2 — Read and analyze each file
+
+Read the full file (not just the diff) with the Read tool. For each file, record every issue found: filename, line number, category, and suggested fix.
+
+### Step 3 — Report findings before touching anything
+
+Print a structured list:
+
+```
+Found N issues:
+
+[file] js/foo.js
+  · L34, L78: Duplicate logic — same DOM query implemented twice; extract to getPanel()
+  · L91: Magic number — bare 14 used as pixel offset; name it TOOLTIP_OFFSET
+
+[file] js/bar.js
+  · L12: Naming — variable `or_` is unclear; rename to outerR, outerG, outerB
+  ...
+```
+
+If no issues are found, output "No code smells detected. Code quality looks good." and stop.
+
+### Step 4 — Fix each issue
+
+Use the Edit tool for **minimal, targeted changes**:
+
+- **Duplicate logic**: extract to a shared constant or function; update all call sites
+- **Magic number/string**: declare `const NAME = value` near the top of the relevant scope; replace all usages
+- **Naming**: rename the variable/function; update all references
+- **Dead code**: delete it
+- **Trailing whitespace / style**: fix in place
+- **Unexported helper**: add `export`; update callers to import instead of re-implementing
+
+Principles:
+- Only fix issues identified in the checklist — no extra improvements
+- Keep each Edit as small as possible
+- After fixing, verify the old bad pattern is gone with grep
+
+### Step 5 — Summary
+
+```
+Cleanup complete:
+
+Fixed N issues:
+  ✓ earth.js — extracted duplicate vertexShader into ATMOS_VERTEX_SHADER constant
+  ✓ main.js — extracted TOOLTIP_CURSOR_OFFSET = 14 (4 references updated)
+  ✓ controls.js — exported updateLayerButtonState; removed duplicate implementation in main.js
+  ...
+
+Skipped (needs manual review):
+  ! foo.js L45 — large commented-out block; confirm it is safe to delete
+```
+
+## Constraints
+
+- **Do not** change function signatures, exported interfaces, or public APIs (unless the issue is a missing export)
+- **Do not** add new features, abstractions, or parameters
+- **Do not** rewrite comments (only delete commented-out dead code)
+- **Do not** touch test file logic
+- If a magic number's intent is uncertain, skip it and flag it in the summary

.codex/skills/goal-driven/SKILL.md

@@ -0,0 +1,101 @@
+---
+name: goal-driven
+description: Run a goal-driven execution loop for very large, long-horizon, rigorously verifiable tasks. Use when the user explicitly wants the lidangzzz/goal-driven method, a master-agent plus worker-agent style workflow, or a persistent loop that keeps working until concrete success criteria are satisfied.
+---
+
+# Goal-Driven
+
+Use this skill when the user wants a strict goal-driven workflow for a hard task with:
+
+- one clear end goal
+- explicit success criteria
+- repeated verification against those criteria
+- continued execution until the criteria are actually met
+
+This skill is adapted from `lidangzzz/goal-driven`, but trimmed for local skill use to avoid bloating context.
+
+## When To Use
+
+Use it for tasks like:
+
+- compilers, interpreters, theorem-like proof work, deep refactors
+- long-running system design or implementation work
+- problems that are expensive and complex, but still objectively testable
+
+Do not use it for:
+
+- vague brainstorming without a success condition
+- short one-shot edits
+- tasks where "done" cannot be evaluated in a meaningful way
+
+## Core Model
+
+The workflow has two roles:
+
+1. Master role
+   Defines the goal, defines the success criteria, audits progress, and decides whether the work is actually complete.
+
+2. Worker role
+   Keeps advancing the task toward the goal. If a result is partial, stalled, or unverifiable, the worker continues.
+
+In Codex, only use actual subagents when the user explicitly asks for delegation or subagent work and the platform supports it. Otherwise emulate the same loop locally: keep working, checkpointing, and re-verifying until the criteria are satisfied.
+
+## Workflow
+
+1. Normalize the task into two blocks:
+   - `Goal`
+   - `Criteria for success`
+
+2. Make the criteria concrete and testable.
+   Good criteria usually include:
+   - required outputs
+   - required validations or tests
+   - edge cases or coverage thresholds
+   - what evidence proves completion
+
+3. Break the work into milestones that can each produce evidence.
+
+4. Execute the next milestone.
+   If subagents are explicitly allowed, the master may delegate bounded worker tasks.
+   If not, do the work locally but keep the master/worker mindset.
+
+5. Whenever work pauses, stalls, or appears complete, audit against the criteria directly.
+   Check artifacts, tests, logs, diffs, metrics, or other real evidence.
+
+6. If the criteria are not met, continue with a specific delta:
+   - what is still missing
+   - what evidence failed
+   - what the next worker pass must improve
+
+7. Stop only when the criteria are met, or when the user explicitly stops the process.
+
+## Operating Rules
+
+- Prefer objective checks over self-reported completion.
+- Do not confuse progress with completion.
+- If the worker says "done", verify it.
+- If verification fails, continue from the gap instead of restarting blindly.
+- Keep the goal stable unless the user changes it.
+- Tighten fuzzy criteria before sinking large amounts of effort.
+
+## Recommended Response Shape
+
+When starting a goal-driven task, structure the kickoff like this:
+
+```md
+Goal
+- ...
+
+Criteria for success
+- ...
+
+Current plan
+1. ...
+2. ...
+3. ...
+
+Verification
+- What evidence will prove completion
+```
+
+For a reusable prompt template, read [references/prompt-template.md](references/prompt-template.md).

.codex/skills/goal-driven/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: "Goal-Driven"
+  short_description: "Drive complex work until explicit success criteria are met."
+  default_prompt: "Use $goal-driven to turn this task into a concrete goal, explicit success criteria, and a verification-driven execution loop."
+
+policy:
+  allow_implicit_invocation: true

.codex/skills/goal-driven/references/prompt-template.md

@@ -0,0 +1,38 @@
+# Goal-Driven Prompt Template
+
+Use this when you want a reusable kickoff prompt for a master/worker execution loop.
+
+```md
+# Goal-Driven System
+
+Goal: [[[[[DEFINE THE FINAL GOAL HERE]]]]]
+
+Criteria for success: [[[[[DEFINE THE SUCCESS CRITERIA HERE]]]]]
+
+You are the master agent.
+
+Your job is to:
+1. Keep the goal and criteria fixed.
+2. Start worker execution toward the goal.
+3. Audit any claimed progress against the criteria.
+4. If the criteria are not met, continue the work with a precise next delta.
+5. Stop only when the criteria are satisfied or the user explicitly stops the process.
+
+Worker requirements:
+1. Break the task into subproblems.
+2. Keep producing concrete progress toward the goal.
+3. Report evidence, not just claims.
+4. Continue until the criteria are satisfied.
+
+Master audit loop:
+1. Check whether the worker is still making progress.
+2. If the worker stalls or claims completion, verify against the criteria.
+3. If verification fails, resume work from the remaining gap.
+4. Repeat until the criteria are met.
+```
+
+## Notes
+
+- Stronger criteria produce better results than stronger rhetoric.
+- Prefer measurable checks such as tests, parity checks, generated artifacts, benchmarks, or reviewable outputs.
+- If the environment does not support subagents, emulate the same loop locally.

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

@@ -1,78 +0,0 @@
----
-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

.codex/skills/release/SKILL.md

@@ -0,0 +1,157 @@
+---
+name: release
+description: Use when the user asks to release, bump version, update changelog/version files, or commit/push a repository release for the Planet repo. Determines version bump type from changes, updates all required version-bearing files, updates changelog and version-history, runs minimal validation, then commits, tags, and pushes.
+---
+
+# 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.
+
+## Required Files
+
+Use `git rev-parse --show-toplevel` to get the repo root. All paths are relative to it:
+
+- `VERSION`
+- `frontend/package.json` (`"version"` field)
+- `pyproject.toml` (`version =` field)
+- `uv.lock` (**never edit manually** — regenerate by running `uv lock`)
+- `docs/CHANGELOG.md`
+- `docs/version-history.md`
+
+## Workflow
+
+### Step 1 — Environment check
+
+```bash
+git branch --show-current   # must be on dev
+git status --short           # check for unrelated uncommitted changes
+cat VERSION                  # read current version
+```
+
+If not on `dev`, stop and tell the user. Do not proceed.
+
+If unrelated uncommitted changes exist, list them and ask the user whether to include them or stash first.
+
+### Step 2 — Determine release type and next version
+
+- If the user provided an explicit type (`feature` / `bugfix`), use it
+- Otherwise infer from `git diff HEAD` and recent `git log`
+- Compute the next version (e.g. `0.26.2` → bugfix → `0.26.3`)
+- **Show the release plan before making any changes:**
+
+```
+Release plan:
+  Type:    bugfix
+  Version: 0.26.2 → 0.26.3
+  Branch:  dev
+  Will update: VERSION, frontend/package.json, pyproject.toml, uv.lock, CHANGELOG.md, version-history.md
+```
+
+### Step 3 — Update version files
+
+Update in order (use Edit for precise replacement, never rewrite whole files):
+
+1. `VERSION` — replace entire content with new version string
+2. `frontend/package.json` — replace `"version": "x.x.x"` line
+3. `pyproject.toml` — replace `version = "x.x.x"` line
+4. Run `uv lock` at repo root to regenerate `uv.lock`
+
+### Step 4 — Update CHANGELOG.md
+
+Insert a new entry at the top of the file:
+
+```markdown
+## x.x.x
+
+Released: YYYY-MM-DD
+
+### Highlights
+
+- ...
+
+### Added / Fixed / Improved
+
+- ... (high-signal items only, max 5)
+
+---
+```
+
+Get today's date with `date +%Y-%m-%d`.
+
+### Step 5 — Update docs/version-history.md
+
+- Update the "current dev version" field in the file header
+- Insert a new row at the top of the timeline table: `| vx.x.x | YYYY-MM-DD | one-line summary |`
+
+### Step 6 — Validate
+
+Run the smallest relevant validation for the changes in scope:
+
+- Python files changed: `python3 -m py_compile <changed_files>`
+- Frontend files changed: run the project-standard check if available; otherwise skip and say so
+- Version consistency: confirm VERSION, package.json, pyproject.toml, and uv.lock all show the same version
+
+```bash
+grep -h "version" VERSION frontend/package.json pyproject.toml
+```
+
+### Step 7 — Pre-commit preview
+
+Show what will be committed:
+
+```bash
+git diff --stat HEAD
+```
+
+Confirm all required files are present and no unexpected files (debug files, `.env`, etc.) are included.
+
+### Step 8 — Commit, tag, and push
+
+```bash
+git add VERSION frontend/package.json pyproject.toml uv.lock docs/CHANGELOG.md docs/version-history.md
+# also stage any code changes included in this release
+git add <code_files>
+
+git commit -m "release: bump version to x.x.x"
+git tag vx.x.x
+git push origin dev
+git push origin vx.x.x
+```
+
+Commit message format is fixed: `release: bump version to x.x.x`
+
+### Step 9 — Completion summary
+
+```
+✓ Version bumped: 0.26.2 → 0.26.3
+✓ CHANGELOG updated
+✓ version-history updated
+✓ uv.lock regenerated
+✓ Validation passed
+✓ commit: release: bump version to 0.26.3
+✓ tag: v0.26.3
+✓ Pushed to origin/dev
+```
+
+## Notes
+
+- `uv.lock` must only be updated by running `uv lock`, never manually
+- The release commit should include only version files + the code for this release — no unrelated changes
+- If `uv` is unavailable in the environment, say so explicitly and remind the user to run it manually

README.md

@@ -227,6 +227,120 @@ bun run build
 
 启动服务后访问: `http://localhost:8000/docs`
 
+## WSL / Windows 局域网访问
+
+如果服务运行在 WSL 中，而你希望：
+
+- Windows 本机浏览器访问开发服务
+- 同一局域网内的手机或其他电脑访问开发服务
+
+推荐按下面顺序排查和配置。
+
+### 1. 在 WSL 中启动服务
+
+```bash
+./planet.sh start --allow-lan
+```
+
+这会让前端监听 `0.0.0.0:3000`，后端监听 `0.0.0.0:8000`。
+
+### 2. 先确认 WSL 内部服务正常
+
+在 WSL 中执行：
+
+```bash
+curl http://localhost:3000
+curl http://localhost:8000/health
+ss -ltnp | grep -E ':3000|:8000'
+```
+
+预期：
+
+- `3000` 返回前端 HTML
+- `8000/health` 返回健康检查 JSON
+- `ss` 中能看到 `0.0.0.0:3000` 和 `0.0.0.0:8000`
+
+如果这一步不通，先不要继续做 Windows 转发。
+
+### 3. 在 Windows 本机验证 localhost 直通
+
+在 Windows PowerShell 中执行：
+
+```powershell
+curl http://localhost:3000
+curl http://localhost:8000/health
+```
+
+在常见的 WSL2 开发环境下，Windows 通常可以直接通过 `localhost` 访问 WSL 中的服务。
+
+### 4. 如果需要让局域网设备访问，再做 Windows 端口转发
+
+注意：下面的命令必须在“以管理员身份运行”的 PowerShell 中执行。
+
+先把 Windows 对外网卡上的 `3000` / `8000` 转发到 Windows 本机 `127.0.0.1`：
+
+```powershell
+netsh interface portproxy delete v4tov4 listenaddress=0.0.0.0 listenport=3000
+netsh interface portproxy delete v4tov4 listenaddress=0.0.0.0 listenport=8000
+
+netsh interface portproxy add v4tov4 listenaddress=0.0.0.0 listenport=3000 connectaddress=127.0.0.1 connectport=3000
+netsh interface portproxy add v4tov4 listenaddress=0.0.0.0 listenport=8000 connectaddress=127.0.0.1 connectport=8000
+```
+
+再放行 Windows 防火墙：
+
+```powershell
+New-NetFirewallRule -DisplayName "WSL Planet 3000" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 3000
+New-NetFirewallRule -DisplayName "WSL Planet 8000" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 8000
+```
+
+检查转发规则是否生效：
+
+```powershell
+netsh interface portproxy show all
+```
+
+预期能看到：
+
+- `0.0.0.0:3000 -> 127.0.0.1:3000`
+- `0.0.0.0:8000 -> 127.0.0.1:8000`
+
+### 5. 查 Windows 局域网 IP，并让其他设备访问
+
+在 Windows PowerShell 中执行：
+
+```powershell
+ipconfig
+```
+
+找到当前联网网卡的 IPv4 地址，例如 `192.168.8.228`。
+
+局域网其他设备可访问：
+
+- `http://<Windows局域网IP>:3000/earth`
+- `http://<Windows局域网IP>:3000/admin`
+
+例如：
+
+- `http://192.168.8.228:3000/earth`
+
+### 6. 常见现象与判断
+
+- WSL 中 `curl localhost:3000` 能通，但 Windows 访问 `WSL 的局域网 IP:3000` 不通：这是正常现象之一，优先验证 Windows 的 `localhost:3000`
+- Windows `localhost:3000` 能通，但局域网设备访问 `Windows 局域网 IP:3000` 不通：通常缺少 `portproxy` 或防火墙放行
+- `whoami /groups` 中 `S-1-5-32-544` 显示 `deny only`：说明当前 PowerShell 不是提权管理员窗口
+
+### 7. 本项目一次性验证顺序
+
+建议固定按这个顺序验证：
+
+1. WSL 中执行 `curl http://localhost:3000`
+2. WSL 中执行 `curl http://localhost:8000/health`
+3. Windows 中执行 `curl http://localhost:3000`
+4. Windows 中执行 `curl http://localhost:8000/health`
+5. 管理员 PowerShell 配置 `portproxy` 和防火墙
+6. 用手机或其他电脑访问 `http://<Windows局域网IP>:3000/earth`
+
 ## 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。
@@ -328,11 +442,11 @@ AI_PROVIDER_SERVICE_TOKEN=change_me
 
 详细文档:
 
-- [docs/aiprovider.md](/home/ray/dev/linkong/planet/docs/aiprovider.md)
+- [docs/technical/agents-aiprovider.md](/home/ray/dev/linkong/planet/docs/technical/agents-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)
+- [docs/technical/frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/technical/frontend-layout-guidelines.md)
+- [docs/plans/frontend-ai-playground-development-plan.md](/home/ray/dev/linkong/planet/docs/plans/frontend-ai-playground-development-plan.md)
+- [docs/plans/agents-situational-awareness-foundation-plan.md](/home/ray/dev/linkong/planet/docs/plans/agents-situational-awareness-foundation-plan.md)
 
 ## 前端页面布局规范
 
@@ -346,7 +460,7 @@ AI_PROVIDER_SERVICE_TOKEN=change_me
 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)
+- [docs/technical/frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/technical/frontend-layout-guidelines.md)
 
 ## License
 

TODO.md

@@ -20,3 +20,18 @@
 - [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
 - [ ] 可选优化（非必做）：将 BGP incident/collector 标点改为 HTML marker（参考 worldmonitor 的 `htmlElementsData` 思路），实现近乎固定屏幕尺寸与更高密度可点击性
+- [ ] 保持 Earth 当前这批纯个人偏好设置继续走本地持久化：`旋转模式`、HUD 面板显示/隐藏、`地形透明度` 暂不升级到后端系统设置，避免把设备级偏好过早做成全局配置
+- [ ] 如果后续明确需要“账号级同步 Earth 偏好”，再单独设计 `Earth user preferences`：优先按用户维度而不是全局系统设置保存，并规划 `localStorage -> backend` 的平滑迁移策略
+- [ ] 为 Planet / Earth 补一个可用的日志查看系统：先明确前后端/AI Provider/采集任务的日志入口、最近日志聚合、筛选与 tail 能力，再决定是先做脚本级统一入口还是控制台内置日志面板
+- [ ] 把 Earth 态势新闻源从 [earth_news.py](/home/ray/dev/linkong/planet/backend/app/services/earth_news.py) 的硬编码列表抽成可配置目录，优先保持当前“实时聚合”链路不变，只先解决新闻源不可配置的问题
+- [ ] 为 Earth 态势新闻设计后续采集器化方案：明确新闻数据模型、去重策略、区域映射、过期清理和 Earth/AI 复用方式，再决定何时把新闻从实时抓取升级成正式 collector
+- [ ] 为 Earth 地球表面增加一层与基础纹理对齐的材质/纹理 overlay，并在同层叠加国界轮廓参考线；要求国界线与底图稳定对齐，且 hover 到国家轮廓时能高亮当前国家，便于校准地表和增强交互
+- [ ] 把 Earth 新闻接入通用巡航队列：按新闻发生地和时间排序生成巡航目标，巡航聚焦到新闻事件时显示对应新闻卡片，并保持实现边界为“通用巡航层 + 新闻业务适配层”，不要再把新闻逻辑直接耦合回 `main.js` 状态机
+- [ ] 为未知位置的算力中心建立分层坐标补全链路：优先 `精确坐标 > 站点/园区命中 > 城市 > 州/省 > 国家内主要算力城市 > 国家质心`，并把每次回退的 `confidence / reason / precision` 明确写进统一 GeoJSON
+- [ ] 为算力中心补一份可维护的本地位置注册表，例如 `canonical_name / aliases / operator / country / region / city / lat / lon / confidence / source_note`，避免把地点知识长期硬编码在 `visualization.py`
+- [ ] 增强 `epoch_ai_gpu` 和相关算力采集器的源页面解析：即使公开 API 不给坐标，也继续尝试从详情页、HTML、内嵌 JSON、schema.org、OpenGraph、脚本变量和 PDF/新闻稿链接里抽地点线索
+- [ ] 为未知位置算力中心增加外部富化策略评估：可选接入公开知识源或搜索兜底，只抓“站点名/园区名/城市名”级别线索，不直接抓经纬度结论，并把结果作为候选证据而不是真值
+- [ ] 为算力中心建立 `operator / cluster name / facility alias` 归一化层，先解决 `xAI / Colossus / Memphis`、`OpenAI / Stargate`、`CoreWeave`、`Lambda`、`Crusoe` 这类同一对象多种写法导致的地点匹配失败
+- [ ] 为估算位置增加更细的视觉和产品表达：除了问号角标，还要支持 tooltip/详情中的“估算依据”“精度级别”“最后核验时间”，并允许在设置中单独开关“仅看精确位置”
+- [ ] 为国家级估算点设计更合理的落点策略：优先落在“该国主要算力/数据中心城市候选集”而不是几何质心，必要时同国多节点做稳定散列分配，避免大量节点堆在荒漠或海上
+- [ ] 为未知位置算力中心建立人工校验工作流：支持导出待核验清单、记录人工确认结果，并把人工确认反哺到位置注册表，逐步减少问号点比例

VERSION

@@ -1 +1 @@
-0.27.0
+0.40.1

aiprovider/README.md

@@ -4,7 +4,7 @@
 
 完整使用说明见:
 
-- [docs/aiprovider.md](/home/ray/dev/linkong/planet/docs/aiprovider.md)
+- [docs/agents/aiprovider.md](/home/ray/dev/linkong/planet/docs/agents/aiprovider.md)
 
 当前支持:
 

backend/app/api/main.py

@@ -13,6 +13,7 @@ from app.api.v1 import (
     collected_data,
     visualization,
     bgp,
+    news,
     system_control,
     tv,
 )
@@ -35,3 +36,4 @@ 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"])
+api_router.include_router(news.router, prefix="/news", tags=["news"])

backend/app/api/v1/datasources.py

@@ -420,6 +420,7 @@ async def list_datasources(
         collector_list.append(
             {
                 "id": datasource.id,
+                "source": datasource.source,
                 "name": datasource.name,
                 "module": datasource.module,
                 "priority": datasource.priority,

backend/app/api/v1/news.py

@@ -0,0 +1,13 @@
+from fastapi import APIRouter, Query
+
+from app.services.earth_news import get_earth_news_payload
+
+router = APIRouter()
+
+
+@router.get("/earth-feed")
+async def get_earth_feed(
+    lat: float | None = Query(None, description="Current Earth view center latitude"),
+    lon: float | None = Query(None, description="Current Earth view center longitude"),
+):
+    return await get_earth_news_payload(lat=lat, lon=lon)

backend/app/api/v1/system_control.py

@@ -4,12 +4,15 @@ import os
 import subprocess
 import sys
 
-from fastapi import APIRouter, Depends, HTTPException, status
+from datetime import datetime
+
+from fastapi import APIRouter, Depends, HTTPException, Query, Request, status
 from pydantic import BaseModel
 
 from app.core.config import ROOT_DIR
 from app.core.security import get_current_user
 from app.models.user import User
+from app.services.persistent_logs import record_audit_log, record_system_log
 from app.services.system_control import (
     build_task_id,
     clear_active_task_id,
@@ -23,6 +26,15 @@ from app.services.system_control import (
     set_active_task_id,
     upsert_task_state,
 )
+from app.services.system_logs import (
+    DEFAULT_LOG_LINE_LIMIT,
+    MAX_LOG_LINE_LIMIT,
+    SUPPORTED_LOG_LEVELS,
+    append_buffer_log,
+    list_log_sources,
+    normalize_log_level,
+    read_log_snapshot,
+)
 
 router = APIRouter()
 
@@ -47,6 +59,59 @@ class RestartTaskLogsResponse(BaseModel):
     lines: list[str]
 
 
+class SystemLogSourceSummary(BaseModel):
+    source_id: str
+    name: str
+    kind: str
+    location: str
+    description: str
+    category: str
+    status: str
+
+
+class SystemLogSourcesResponse(BaseModel):
+    items: list[SystemLogSourceSummary]
+
+
+class SystemLogDailyMarker(BaseModel):
+    date_token: str
+    total: int
+    dominant_level: str
+
+
+class SystemLogSnapshotResponse(BaseModel):
+    source_id: str
+    name: str
+    kind: str
+    location: str
+    description: str
+    category: str
+    status: str
+    level: str
+    selected_levels: list[str] = []
+    search_query: str = ""
+    available_levels: list[str]
+    daily_markers: list[SystemLogDailyMarker] = []
+    line_limit: int
+    line_count: int
+    lines: list[str]
+
+
+class EarthClientLogEventCreate(BaseModel):
+    level: str = "error"
+    message: str
+    category: str | None = None
+    url: str | None = None
+    module: str | None = None
+    detail: str | None = None
+
+
+class EarthClientLogEventResponse(BaseModel):
+    accepted: bool
+    source_id: str
+    level: str
+
+
 def ensure_super_admin(current_user: User) -> None:
     if not require_super_admin(current_user.role):
         raise HTTPException(
@@ -55,9 +120,22 @@ def ensure_super_admin(current_user: User) -> None:
         )
 
 
+def validate_log_date(raw_value: str | None, field_name: str) -> str | None:
+    if raw_value in {None, ""}:
+        return None
+    try:
+        return datetime.strptime(raw_value, "%Y-%m-%d").date().isoformat()
+    except ValueError as exc:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=f"{field_name} must be in YYYY-MM-DD format",
+        ) from exc
+
+
 @router.post("/restart-tasks", response_model=RestartTaskResponse)
 async def create_restart_task(
     payload: RestartTaskCreate,
+    request: Request,
     current_user: User = Depends(get_current_user),
 ):
     ensure_super_admin(current_user)
@@ -133,11 +211,31 @@ async def create_restart_task(
             requested_by=requested_by,
         )
         clear_active_task_id(task_id)
+        await record_audit_log(
+            action="system.restart_task.requested",
+            actor_id=current_user.id,
+            actor_name=current_user.username,
+            target_type="restart_task",
+            target_id=task_id,
+            result="failed",
+            ip=request.client.host if request.client else None,
+            details={"action": payload.action, "message": task_state["message"]},
+        )
         raise HTTPException(
             status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
             detail=task_state["message"],
         ) from exc
 
+    await record_audit_log(
+        action="system.restart_task.requested",
+        actor_id=current_user.id,
+        actor_name=current_user.username,
+        target_type="restart_task",
+        target_id=task_id,
+        result="accepted",
+        ip=request.client.host if request.client else None,
+        details={"action": payload.action},
+    )
     return task_state
 
 
@@ -165,3 +263,92 @@ async def get_restart_task_logs(
     if task is None:
         raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Restart task not found")
     return {"task_id": task_id, "lines": get_task_logs(task_id)}
+
+
+@router.get("/logs/sources", response_model=SystemLogSourcesResponse)
+async def get_system_log_sources(
+    current_user: User = Depends(get_current_user),
+):
+    ensure_super_admin(current_user)
+    return {"items": list_log_sources()}
+
+
+@router.get("/logs/{source_id}", response_model=SystemLogSnapshotResponse)
+async def get_system_log_snapshot(
+    source_id: str,
+    limit: int = DEFAULT_LOG_LINE_LIMIT,
+    level: str = "all",
+    levels: str | None = Query(None, description="Comma-separated log levels"),
+    start_date: str | None = Query(None, description="Filter logs from this date (YYYY-MM-DD)"),
+    end_date: str | None = Query(None, description="Filter logs until this date (YYYY-MM-DD)"),
+    search: str | None = Query(None, description="Case-insensitive substring search"),
+    current_user: User = Depends(get_current_user),
+):
+    ensure_super_admin(current_user)
+
+    if limit < 1 or limit > MAX_LOG_LINE_LIMIT:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail=f"limit must be between 1 and {MAX_LOG_LINE_LIMIT}",
+        )
+    if str(level).strip().lower() not in SUPPORTED_LOG_LEVELS and normalize_log_level(level) == "all" and str(level).strip().lower() not in {"", "all"}:
+        raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail="Unsupported log level")
+    if levels:
+        for raw_level in str(levels).split(","):
+            normalized_level = str(raw_level).strip().lower()
+            if not normalized_level:
+                continue
+            if normalized_level not in SUPPORTED_LOG_LEVELS and normalize_log_level(normalized_level) == "all":
+                raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail="Unsupported log level")
+    normalized_start_date = validate_log_date(start_date, "start_date")
+    normalized_end_date = validate_log_date(end_date, "end_date")
+    if normalized_start_date and normalized_end_date and normalized_start_date > normalized_end_date:
+        raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail="start_date must be earlier than or equal to end_date")
+
+    snapshot = read_log_snapshot(
+        source_id,
+        limit,
+        level=level,
+        levels=levels,
+        start_date=normalized_start_date,
+        end_date=normalized_end_date,
+        search=search,
+    )
+    if snapshot is None:
+        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Log source not found")
+    return snapshot
+
+
+@router.post("/logs/earth-client", response_model=EarthClientLogEventResponse)
+async def ingest_earth_client_log(
+    payload: EarthClientLogEventCreate,
+    request: Request,
+):
+    normalized_level = normalize_log_level(payload.level)
+    append_buffer_log(
+        "earth-client",
+        level=normalized_level,
+        message=payload.message,
+        context={
+            "category": payload.category or "",
+            "url": payload.url or "",
+            "module": payload.module or "",
+            "detail": payload.detail or "",
+        },
+    )
+    await record_system_log(
+        source="earth-client",
+        service="earth",
+        module=payload.module or "earth-client",
+        event="earth.client.runtime_log",
+        level=normalized_level,
+        message=payload.message,
+        category=payload.category or "client-runtime",
+        context={
+            "url": payload.url or "",
+            "detail": payload.detail or "",
+            "module": payload.module or "",
+            "client_ip": request.client.host if request.client else "",
+        },
+    )
+    return {"accepted": True, "source_id": "earth-client", "level": normalized_level}

backend/app/api/v1/visualization.py

@@ -6,12 +6,14 @@ Returns GeoJSON format compatible with Three.js, CesiumJS, and Unreal Cesium.
 
 from datetime import UTC, datetime
 import math
-from fastapi import APIRouter, HTTPException, Depends, Query
+import httpx
+from fastapi import APIRouter, HTTPException, Depends, Query, Response
 from sqlalchemy.ext.asyncio import AsyncSession
 from sqlalchemy import select, func
 from typing import List, Dict, Any, Optional
 
 from app.core.collected_data_fields import get_record_field
+from app.core.countries import get_country_centroid
 from app.core.satellite_tle import build_tle_lines_from_elements
 from app.core.time import to_iso8601_utc
 from app.db.session import get_db
@@ -21,8 +23,14 @@ from app.models.collected_data import CollectedData
 from app.services.bgp_collectors import build_bgp_collector_coverage
 from app.services.cable_graph import build_graph_from_data, CableGraph, haversine_distance
 from app.services.collectors.bgp_common import RIPE_RIS_COLLECTOR_COORDS
+from app.services.persistent_logs import record_system_log
+from app.core.logging import get_logger
 
 router = APIRouter()
+logger = get_logger(__name__, service="api")
+TERRAIN_TILE_URL_TEMPLATE = (
+    "https://s3.amazonaws.com/elevation-tiles-prod/terrarium/{z}/{x}/{y}.png"
+)
 
 
 # ============== Converter Functions ==============
@@ -176,6 +184,12 @@ def convert_satellite_to_geojson(records: List[CollectedData]) -> Dict[str, Any]
                 mean_motion=metadata.get("mean_motion"),
             )
 
+        constellation_group = _normalize_satellite_constellation_group(
+            metadata.get("constellation_group"),
+            record.name,
+        )
+        footprint_policy = _get_satellite_footprint_policy(constellation_group)
+
         features.append(
             {
                 "type": "Feature",
@@ -185,6 +199,8 @@ def convert_satellite_to_geojson(records: List[CollectedData]) -> Dict[str, Any]
                     "id": record.id,
                     "norad_cat_id": norad_id,
                     "name": record.name,
+                    "constellation_group": constellation_group,
+                    "footprint_policy": footprint_policy,
                     "international_designator": metadata.get("international_designator"),
                     "epoch": metadata.get("epoch"),
                     "inclination": metadata.get("inclination"),
@@ -205,6 +221,31 @@ def convert_satellite_to_geojson(records: List[CollectedData]) -> Dict[str, Any]
     return {"type": "FeatureCollection", "features": features}
 
 
+def _normalize_satellite_constellation_group(
+    raw_group: Any,
+    name: Optional[str],
+) -> Optional[str]:
+    normalized_group = str(raw_group or "").strip().lower()
+    if normalized_group:
+        return normalized_group
+
+    normalized_name = str(name or "").strip().upper()
+    if normalized_name.startswith("STARLINK"):
+        return "starlink"
+    if normalized_name.startswith("IRIDIUM"):
+        return "iridium-next"
+
+    return None
+
+
+def _get_satellite_footprint_policy(constellation_group: Optional[str]) -> str:
+    if constellation_group == "starlink":
+        return "starlink_ground_footprint"
+    if constellation_group == "iridium-next":
+        return "iridium_coverage_ring"
+    return "none"
+
+
 def _current_collected_data_stmt(source: str):
     return (
         select(CollectedData)
@@ -359,6 +400,215 @@ def convert_gpu_cluster_to_geojson(records: List[CollectedData]) -> Dict[str, An
     return {"type": "FeatureCollection", "features": features}
 
 
+def _parse_float(value: Any) -> Optional[float]:
+    try:
+        if value in (None, ""):
+            return None
+        return float(value)
+    except (TypeError, ValueError):
+        return None
+
+
+COMPUTE_CENTER_COORDINATE_HINTS = (
+    ("el capitan", 37.6819, -121.7681),
+    ("livermore", 37.6819, -121.7681),
+    ("llnl", 37.6819, -121.7681),
+    ("lawrence livermore", 37.6819, -121.7681),
+    ("frontier", 35.9319, -84.3107),
+    ("oak ridge", 35.9319, -84.3107),
+    ("ornl", 35.9319, -84.3107),
+    ("aurora", 41.7130, -87.9820),
+    ("argonne", 41.7130, -87.9820),
+    ("anl", 41.7130, -87.9820),
+    ("fugaku", 34.6953, 135.1974),
+    ("kobe", 34.6953, 135.1974),
+    ("riken", 34.6953, 135.1974),
+    ("summit", 35.9319, -84.3107),
+    ("leonardo", 44.4949, 11.3426),
+    ("bologna", 44.4949, 11.3426),
+    ("alps", 46.0037, 8.9511),
+    ("lugano", 46.0037, 8.9511),
+    ("sunway taihulight", 31.4912, 120.3119),
+    ("wuxi", 31.4912, 120.3119),
+    ("tianhe-2", 23.1291, 113.2644),
+    ("tianhe-2a", 23.1291, 113.2644),
+    ("guangzhou", 23.1291, 113.2644),
+    ("colossus", 35.1495, -90.0490),
+    ("memphis", 35.1495, -90.0490),
+    ("xai", 35.1495, -90.0490),
+)
+
+
+def _normalize_hint_text(*parts: Any) -> str:
+    return " ".join(
+        str(part).strip().lower()
+        for part in parts
+        if part not in (None, "")
+    )
+
+
+def _resolve_compute_center_coordinates(
+    record: CollectedData,
+    metadata: Dict[str, Any],
+) -> Dict[str, Any]:
+    latitude = _parse_float(get_record_field(record, "latitude"))
+    longitude = _parse_float(get_record_field(record, "longitude"))
+    if latitude not in (None, 0.0) and longitude not in (None, 0.0):
+        return {
+            "latitude": latitude,
+            "longitude": longitude,
+            "location_precision": "precise",
+            "geography_mode": "source_coordinates",
+            "is_estimated": False,
+            "estimated_reason": None,
+        }
+
+    hint_text = _normalize_hint_text(
+        record.name,
+        get_record_field(record, "city"),
+        get_record_field(record, "country"),
+        metadata.get("site"),
+        metadata.get("organization"),
+        metadata.get("operator"),
+    )
+    for needle, resolved_latitude, resolved_longitude in COMPUTE_CENTER_COORDINATE_HINTS:
+        if needle in hint_text:
+            return {
+                "latitude": resolved_latitude,
+                "longitude": resolved_longitude,
+                "location_precision": "estimated_site",
+                "geography_mode": "site_hint",
+                "is_estimated": True,
+                "estimated_reason": f"Matched known site hint: {needle}",
+            }
+
+    centroid = get_country_centroid(get_record_field(record, "country"))
+    if centroid:
+        return {
+            "latitude": centroid.get("latitude"),
+            "longitude": centroid.get("longitude"),
+            "location_precision": "estimated_country",
+            "geography_mode": "country_centroid",
+            "is_estimated": True,
+            "estimated_reason": "Estimated from country centroid",
+        }
+
+    return {
+        "latitude": latitude,
+        "longitude": longitude,
+        "location_precision": "unknown",
+        "geography_mode": "unknown",
+        "is_estimated": True,
+        "estimated_reason": "No resolvable location hints",
+    }
+
+
+def _normalize_capacity_band(capacity_value: Optional[float], capacity_unit: str) -> str:
+    if capacity_value is None:
+        return "unknown"
+
+    unit = str(capacity_unit or "").strip().lower()
+    if unit in {"pflop/s", "pflops", "pflop"}:
+        normalized_tflops = capacity_value * 1000
+    elif unit in {"gflop/s", "gflops", "gflop"}:
+        normalized_tflops = capacity_value / 1000
+    else:
+        normalized_tflops = capacity_value
+
+    if normalized_tflops >= 1_000_000:
+        return "exascale"
+    if normalized_tflops >= 100_000:
+        return "ultra"
+    if normalized_tflops >= 10_000:
+        return "large"
+    if normalized_tflops > 0:
+        return "regional"
+    return "unknown"
+
+
+def convert_compute_centers_to_geojson(records: List[CollectedData]) -> Dict[str, Any]:
+    """Convert compute infrastructure records into a unified GeoJSON layer."""
+    features = []
+
+    for record in records:
+        metadata = record.extra_data or {}
+        coordinate_info = _resolve_compute_center_coordinates(record, metadata)
+        latitude = coordinate_info.get("latitude")
+        longitude = coordinate_info.get("longitude")
+        site_type = (
+            "supercomputer"
+            if record.source == "top500" or record.data_type == "supercomputer"
+            else "gpu_cluster"
+        )
+        if latitude in (None, 0.0) or longitude in (None, 0.0):
+            continue
+
+        if site_type == "supercomputer":
+            capacity_value = _parse_float(get_record_field(record, "rmax"))
+            capacity_unit = "GFlops"
+        else:
+            capacity_value = _parse_float(get_record_field(record, "value"))
+            capacity_unit = str(get_record_field(record, "unit") or "TFlop/s")
+
+        vendor = (
+            metadata.get("manufacturer")
+            or metadata.get("vendor")
+            or metadata.get("gpu_type")
+        )
+        operator = (
+            metadata.get("organization")
+            or metadata.get("operator")
+            or metadata.get("owner")
+        )
+        rank = metadata.get("rank")
+        if rank in (None, "") and site_type == "supercomputer":
+            rank = get_record_field(record, "rank")
+
+        updated_at = to_iso8601_utc(record.reference_date or record.collected_at)
+
+        features.append(
+            {
+                "type": "Feature",
+                "id": record.id,
+                "geometry": {
+                    "type": "Point",
+                    "coordinates": [longitude or 0, latitude or 0],
+                },
+                "properties": {
+                    "id": record.id,
+                    "source_id": record.source_id,
+                    "name": record.name,
+                    "site_type": site_type,
+                    "country": get_record_field(record, "country"),
+                    "city": get_record_field(record, "city"),
+                    "latitude": latitude,
+                    "longitude": longitude,
+                    "operator": operator,
+                    "vendor": vendor,
+                    "capacity_value": capacity_value,
+                    "capacity_unit": capacity_unit,
+                    "capacity_band": _normalize_capacity_band(capacity_value, capacity_unit),
+                    "rank": rank,
+                    "gpu_count": metadata.get("gpu_count"),
+                    "gpu_type": metadata.get("gpu_type"),
+                    "cores": get_record_field(record, "cores"),
+                    "power": get_record_field(record, "power"),
+                    "source": record.source,
+                    "updated_at": updated_at,
+                    "status": "observed",
+                    "location_precision": coordinate_info.get("location_precision"),
+                    "geography_mode": coordinate_info.get("geography_mode"),
+                    "is_estimated": coordinate_info.get("is_estimated", False),
+                    "estimated_reason": coordinate_info.get("estimated_reason"),
+                    "data_type": "compute_center",
+                    "metadata": metadata,
+                },
+            }
+        )
+
+    return {"type": "FeatureCollection", "features": features}
+
+
 def convert_bgp_anomalies_to_geojson(
     records: List[BGPAnomaly],
     geography_hints: Optional[Dict[str, Dict[str, Any]]] = None,
@@ -776,15 +1026,41 @@ async def get_cables_geojson(db: AsyncSession = Depends(get_db)):
     except HTTPException:
         raise
     except Exception as e:
+        logger.exception_event(
+            "Failed to build cables GeoJSON response",
+            event="visualization.cables.load_failed",
+            context={"error": str(e)},
+        )
+        await record_system_log(
+            source="backend",
+            service="api",
+            module=__name__,
+            event="visualization.cables.load_failed",
+            level="error",
+            message="Failed to build cables GeoJSON response",
+            category="visualization",
+            context={"error": str(e)},
+        )
         raise HTTPException(status_code=500, detail=f"Internal error: {str(e)}")
 
 
 @router.get("/geo/landing-points")
 async def get_landing_points_geojson(db: AsyncSession = Depends(get_db)):
     try:
-        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")
+        records_by_source = await _load_current_collected_data_by_sources(
+            db,
+            [
+                "arcgis_landing_points",
+                "arcgis_cable_landing_relation",
+                "arcgis_cables",
+            ],
+        )
+        records = records_by_source.get("arcgis_landing_points", [])
+        relation_records = records_by_source.get(
+            "arcgis_cable_landing_relation",
+            [],
+        )
+        cable_records = records_by_source.get("arcgis_cables", [])
 
         city_to_cable_ids_map, cable_id_to_name_map = _build_landing_point_cable_maps(
             relation_records,
@@ -801,9 +1077,68 @@ async def get_landing_points_geojson(db: AsyncSession = Depends(get_db)):
     except HTTPException:
         raise
     except Exception as e:
+        logger.exception_event(
+            "Failed to build landing points GeoJSON response",
+            event="visualization.landing_points.load_failed",
+            context={"error": str(e)},
+        )
+        await record_system_log(
+            source="backend",
+            service="api",
+            module=__name__,
+            event="visualization.landing_points.load_failed",
+            level="error",
+            message="Failed to build landing points GeoJSON response",
+            category="visualization",
+            context={"error": str(e)},
+        )
         raise HTTPException(status_code=500, detail=f"Internal error: {str(e)}")
 
 
+@router.get("/terrain/terrarium/{z}/{x}/{y}.png")
+async def get_terrarium_tile(z: int, x: int, y: int):
+    """Proxy Terrarium elevation tiles through the backend to avoid browser CORS issues."""
+    if z < 0 or x < 0 or y < 0:
+        raise HTTPException(status_code=400, detail="Invalid terrain tile coordinates")
+
+    url = TERRAIN_TILE_URL_TEMPLATE.format(z=z, x=x, y=y)
+
+    try:
+        async with httpx.AsyncClient(
+            timeout=20.0,
+            follow_redirects=True,
+        ) as client:
+            upstream = await client.get(url)
+            upstream.raise_for_status()
+    except httpx.HTTPStatusError as exc:
+        raise HTTPException(
+            status_code=exc.response.status_code,
+            detail=f"Terrain tile upstream error: {exc.response.status_code}",
+        ) from exc
+    except httpx.HTTPError as exc:
+        raise HTTPException(
+            status_code=502,
+            detail=f"Terrain tile fetch failed: {exc}",
+        ) from exc
+
+    cache_control = upstream.headers.get("cache-control") or "public, max-age=86400"
+    etag = upstream.headers.get("etag")
+    last_modified = upstream.headers.get("last-modified")
+    headers = {
+        "Cache-Control": cache_control,
+    }
+    if etag:
+        headers["ETag"] = etag
+    if last_modified:
+        headers["Last-Modified"] = last_modified
+
+    return Response(
+        content=upstream.content,
+        media_type=upstream.headers.get("content-type", "image/png"),
+        headers=headers,
+    )
+
+
 @router.get("/geo/all")
 async def get_all_geojson(db: AsyncSession = Depends(get_db)):
     records_by_source = await _load_current_collected_data_by_sources(
@@ -916,6 +1251,53 @@ async def get_gpu_clusters_geojson(
     }
 
 
+@router.get("/geo/compute-centers")
+async def get_compute_centers_geojson(
+    limit: int = Query(200, ge=1, le=1000),
+    db: AsyncSession = Depends(get_db),
+):
+    """获取统一算力中心 GeoJSON 数据"""
+    records_by_source = await _load_current_collected_data_by_sources(
+        db,
+        ["top500", "epoch_ai_gpu"],
+    )
+    records = _filter_known_records(
+        records_by_source.get("top500", []) + records_by_source.get("epoch_ai_gpu", []),
+    )
+    if limit is not None:
+        records = records[:limit]
+
+    if not records:
+        return {
+            "type": "FeatureCollection",
+            "features": [],
+            "count": 0,
+            "stats": {
+                "total": 0,
+                "supercomputers": 0,
+                "gpu_clusters": 0,
+            },
+        }
+
+    geojson = convert_compute_centers_to_geojson(records)
+    features = geojson.get("features", [])
+    return {
+        **geojson,
+        "count": len(features),
+        "stats": {
+            "total": len(features),
+            "supercomputers": sum(
+                1 for feature in features
+                if feature.get("properties", {}).get("site_type") == "supercomputer"
+            ),
+            "gpu_clusters": sum(
+                1 for feature in features
+                if feature.get("properties", {}).get("site_type") == "gpu_cluster"
+            ),
+        },
+    }
+
+
 @router.get("/geo/bgp-anomalies")
 async def get_bgp_anomalies_geojson(
     severity: Optional[str] = Query(None),

backend/app/api/v1/websocket.py

@@ -2,7 +2,6 @@
 
 import asyncio
 import json
-import logging
 from datetime import UTC, datetime
 from typing import Optional
 
@@ -10,10 +9,11 @@ from fastapi import APIRouter, WebSocket, WebSocketDisconnect, Query
 from jose import jwt, JWTError
 
 from app.core.config import settings
+from app.core.logging import get_logger
 from app.core.time import to_iso8601_utc
 from app.core.websocket.manager import manager
 
-logger = logging.getLogger(__name__)
+logger = get_logger(__name__, service="api")
 router = APIRouter()
 
 
@@ -22,11 +22,18 @@ async def authenticate_token(token: str) -> Optional[dict]:
     try:
         payload = jwt.decode(token, settings.SECRET_KEY, algorithms=[settings.ALGORITHM])
         if payload.get("type") != "access":
-            logger.warning(f"WebSocket auth failed: wrong token type")
+            logger.warning_event(
+                "WebSocket auth failed: wrong token type",
+                event="auth.websocket.invalid_token_type",
+            )
             return None
         return payload
     except JWTError as e:
-        logger.warning(f"WebSocket auth failed: {e}")
+        logger.warning_event(
+            "WebSocket auth failed",
+            event="auth.websocket.decode_failed",
+            context={"error": str(e)},
+        )
         return None
 
 
@@ -36,10 +43,17 @@ async def websocket_endpoint(
     token: str = Query(...),
 ):
     """WebSocket endpoint for real-time data"""
-    logger.info(f"WebSocket connection attempt with token: {token[:20]}...")
+    logger.info_event(
+        "WebSocket connection attempt",
+        event="auth.websocket.connection_attempt",
+        context={"token_preview": f"{token[:8]}..."},
+    )
     payload = await authenticate_token(token)
     if payload is None:
-        logger.warning("WebSocket authentication failed, closing connection")
+        logger.warning_event(
+            "WebSocket authentication failed, closing connection",
+            event="auth.websocket.connection_rejected",
+        )
         await websocket.close(code=4001)
         return
 

backend/app/core/cache.py

@@ -1,15 +1,15 @@
 """Redis caching service"""
 
 import json
-import logging
 from datetime import timedelta
 from typing import Optional, Any
 
 import redis
 
 from app.core.config import settings
+from app.core.logging import get_logger
 
-logger = logging.getLogger(__name__)
+logger = get_logger(__name__)
 
 
 # Lazy Redis client initialization
@@ -47,7 +47,7 @@ class CacheService:
                 return json.loads(value)
             return None
         except Exception as e:
-            logger.warning(f"Cache get error: {e}")
+            logger.warning_event("Cache get error", event="cache.get.failed", context={"error": str(e)})
             return None
 
     def set(
@@ -61,7 +61,7 @@ class CacheService:
             serialized = json.dumps(value, default=str)
             return self.client.setex(key, expire_seconds, serialized)
         except Exception as e:
-            logger.warning(f"Cache set error: {e}")
+            logger.warning_event("Cache set error", event="cache.set.failed", context={"error": str(e)})
             return False
 
     def delete(self, key: str) -> bool:
@@ -69,7 +69,7 @@ class CacheService:
         try:
             return self.client.delete(key) > 0
         except Exception as e:
-            logger.warning(f"Cache delete error: {e}")
+            logger.warning_event("Cache delete error", event="cache.delete.failed", context={"error": str(e)})
             return False
 
     def delete_pattern(self, pattern: str) -> int:
@@ -80,7 +80,7 @@ class CacheService:
                 return self.client.delete(*keys)
             return 0
         except Exception as e:
-            logger.warning(f"Cache delete_pattern error: {e}")
+            logger.warning_event("Cache delete_pattern error", event="cache.delete_pattern.failed", context={"error": str(e)})
             return 0
 
     def get_or_set(

backend/app/core/data_sources.py

@@ -30,6 +30,7 @@ COLLECTOR_URL_KEYS = {
     "iptoasn_prefix_geo": "iptoasn.combined_url",
     "opengeofeed_prefix_geo": "opengeofeed.public_csv_url",
     "nro_delegated_prefix_geo": "nro.delegated_stats_url",
+    "news_live_streams": "news_live_streams.channels_url",
 }
 
 

backend/app/core/data_sources.yaml

@@ -86,3 +86,11 @@ opengeofeed:
 nro:
   # NRO delegated stats 下载地址
   delegated_stats_url: "https://ftp.ripe.net/pub/stats/ripencc/nro-stats/latest/nro-delegated-stats"
+
+news_live_streams:
+  # IPTV-org 频道元数据 JSON
+  channels_url: "https://iptv-org.github.io/api/channels.json"
+  # IPTV-org 频道播放流 JSON
+  streams_url: "https://iptv-org.github.io/api/streams.json"
+  # IPTV-org 台标 JSON
+  logos_url: "https://iptv-org.github.io/api/logos.json"

backend/app/core/logging.py

@@ -0,0 +1,161 @@
+from __future__ import annotations
+
+import json
+import logging
+import os
+import re
+
+from collections.abc import Mapping, Sequence
+from typing import Any
+
+from app.core.request_context import get_request_id
+
+DEFAULT_SERVICE = "backend"
+DEFAULT_EVENT = "app.log"
+DEFAULT_LOG_LEVEL = os.getenv("PLANET_LOG_LEVEL", "INFO").upper()
+REDACTED = "[REDACTED]"
+SENSITIVE_FIELD_NAMES = {
+    "access_token",
+    "api_key",
+    "authorization",
+    "cookie",
+    "password",
+    "refresh_token",
+    "secret",
+    "token",
+}
+SENSITIVE_TEXT_PATTERNS = (
+    re.compile(r"(?i)(authorization\s*[:=]\s*)(.+)"),
+    re.compile(r"(?i)(bearer\s+)([A-Za-z0-9._\-]+)"),
+    re.compile(r"(?i)(token\s*[:=]\s*)(.+)"),
+    re.compile(r"(?i)(password\s*[:=]\s*)(.+)"),
+    re.compile(r"(?i)(cookie\s*[:=]\s*)(.+)"),
+)
+
+
+def sanitize_log_value(value: Any) -> Any:
+    if isinstance(value, Mapping):
+        return {
+            str(key): (REDACTED if str(key).lower() in SENSITIVE_FIELD_NAMES else sanitize_log_value(item))
+            for key, item in value.items()
+        }
+    if isinstance(value, Sequence) and not isinstance(value, (str, bytes, bytearray)):
+        return [sanitize_log_value(item) for item in value]
+    if isinstance(value, str):
+        sanitized = value
+        for pattern in SENSITIVE_TEXT_PATTERNS:
+            sanitized = pattern.sub(lambda match: f"{match.group(1)}{REDACTED}", sanitized)
+        return sanitized
+    return value
+
+
+def _normalize_context(context: Any) -> dict[str, Any]:
+    if context is None:
+        return {}
+    if isinstance(context, Mapping):
+        sanitized = sanitize_log_value(context)
+        return {str(key): value for key, value in sanitized.items()}
+    return {"value": sanitize_log_value(context)}
+
+
+class PlanetContextFilter(logging.Filter):
+    def filter(self, record: logging.LogRecord) -> bool:
+        record.request_id = getattr(record, "request_id", None) or get_request_id() or "-"
+        record.service = getattr(record, "service", None) or DEFAULT_SERVICE
+        record.event = getattr(record, "event", None) or DEFAULT_EVENT
+        record.context = _normalize_context(getattr(record, "context", None))
+        record.message = sanitize_log_value(record.getMessage())
+        return True
+
+
+class PlanetFormatter(logging.Formatter):
+    def format(self, record: logging.LogRecord) -> str:
+        timestamp = self.formatTime(record, self.datefmt)
+        level = record.levelname
+        service = getattr(record, "service", DEFAULT_SERVICE)
+        module_name = record.name
+        event = getattr(record, "event", DEFAULT_EVENT)
+        request_id = getattr(record, "request_id", "-")
+        message = sanitize_log_value(record.getMessage())
+        context = _normalize_context(getattr(record, "context", None))
+        context_suffix = ""
+        if context:
+            context_suffix = f" context={json.dumps(context, ensure_ascii=False, sort_keys=True)}"
+        rendered = (
+            f"{timestamp} {level} service={service} module={module_name} "
+            f"event={event} request_id={request_id} message={message}{context_suffix}"
+        )
+        if record.exc_info:
+            rendered = f"{rendered}\n{self.formatException(record.exc_info)}"
+        return rendered
+
+
+class PlanetLoggerAdapter(logging.LoggerAdapter):
+    def process(self, msg: Any, kwargs: dict[str, Any]) -> tuple[Any, dict[str, Any]]:
+        extra = dict(self.extra)
+        extra.update(kwargs.get("extra", {}))
+        if "context" in extra:
+            extra["context"] = _normalize_context(extra.get("context"))
+        kwargs["extra"] = extra
+        return sanitize_log_value(msg), kwargs
+
+    def log_event(
+        self,
+        level: int,
+        message: str,
+        *,
+        event: str,
+        context: Mapping[str, Any] | None = None,
+        **extra: Any,
+    ) -> None:
+        self.log(level, message, extra={"event": event, "context": context or {}, **extra})
+
+    def debug_event(self, message: str, *, event: str, context: Mapping[str, Any] | None = None, **extra: Any) -> None:
+        self.log_event(logging.DEBUG, message, event=event, context=context, **extra)
+
+    def info_event(self, message: str, *, event: str, context: Mapping[str, Any] | None = None, **extra: Any) -> None:
+        self.log_event(logging.INFO, message, event=event, context=context, **extra)
+
+    def warning_event(self, message: str, *, event: str, context: Mapping[str, Any] | None = None, **extra: Any) -> None:
+        self.log_event(logging.WARNING, message, event=event, context=context, **extra)
+
+    def error_event(self, message: str, *, event: str, context: Mapping[str, Any] | None = None, **extra: Any) -> None:
+        self.log_event(logging.ERROR, message, event=event, context=context, **extra)
+
+    def exception_event(
+        self,
+        message: str,
+        *,
+        event: str,
+        context: Mapping[str, Any] | None = None,
+        **extra: Any,
+    ) -> None:
+        self.error(message, exc_info=True, extra={"event": event, "context": context or {}, **extra})
+
+
+def get_logger(name: str, *, service: str = DEFAULT_SERVICE) -> PlanetLoggerAdapter:
+    return PlanetLoggerAdapter(logging.getLogger(name), {"service": service})
+
+
+def configure_logging(level: str | None = None) -> None:
+    root_logger = logging.getLogger()
+    if getattr(configure_logging, "_configured", False):
+        if level:
+            root_logger.setLevel(level.upper())
+        return
+
+    handler = logging.StreamHandler()
+    handler.setFormatter(PlanetFormatter(datefmt="%Y-%m-%d %H:%M:%S"))
+    handler.addFilter(PlanetContextFilter())
+
+    root_logger.handlers.clear()
+    root_logger.addHandler(handler)
+    root_logger.setLevel((level or DEFAULT_LOG_LEVEL).upper())
+
+    for logger_name in ("uvicorn", "uvicorn.error", "uvicorn.access"):
+        target_logger = logging.getLogger(logger_name)
+        target_logger.handlers.clear()
+        target_logger.propagate = True
+
+    logging.captureWarnings(True)
+    configure_logging._configured = True

backend/app/core/request_context.py

@@ -0,0 +1,14 @@
+from __future__ import annotations
+
+from contextvars import ContextVar
+
+
+request_id_context: ContextVar[str | None] = ContextVar("request_id", default=None)
+
+
+def set_request_id(request_id: str | None) -> None:
+    request_id_context.set(request_id)
+
+
+def get_request_id() -> str | None:
+    return request_id_context.get()

backend/app/db/session.py

@@ -5,10 +5,22 @@ from sqlalchemy.ext.asyncio import AsyncSession, create_async_engine, async_sess
 from sqlalchemy.orm import declarative_base
 
 from app.core.config import settings
+from app.core.logging import get_logger
+
+logger = get_logger(__name__)
+
+DB_POOL_CONFIG = {
+    "pool_pre_ping": True,
+    "pool_recycle": 1800,
+    "pool_size": 10,
+    "max_overflow": 20,
+    "pool_timeout": 30,
+}
 
 engine = create_async_engine(
     settings.DATABASE_URL,
     echo=settings.DEBUG if hasattr(settings, "DEBUG") else False,
+    **DB_POOL_CONFIG,
 )
 
 async_session_factory = async_sessionmaker(engine, class_=AsyncSession, expire_on_commit=False)
@@ -97,6 +109,19 @@ async def init_db():
     import app.models.system_setting  # noqa: F401
     import app.models.playground_session  # noqa: F401
     import app.models.playground_message  # noqa: F401
+    import app.models.system_log  # noqa: F401
+
+    logger.warning_event(
+        "Database pool settings active",
+        event="database.pool.initialized",
+        context={
+            "pool_pre_ping": DB_POOL_CONFIG["pool_pre_ping"],
+            "pool_recycle": DB_POOL_CONFIG["pool_recycle"],
+            "pool_size": DB_POOL_CONFIG["pool_size"],
+            "max_overflow": DB_POOL_CONFIG["max_overflow"],
+            "pool_timeout": DB_POOL_CONFIG["pool_timeout"],
+        },
+    )
 
     async with engine.begin() as conn:
         await conn.run_sync(Base.metadata.create_all)

backend/app/main.py

@@ -1,4 +1,5 @@
 from contextlib import asynccontextmanager
+from uuid import uuid4
 
 from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
@@ -7,6 +8,8 @@ from starlette.middleware.base import BaseHTTPMiddleware
 from app.api.main import api_router
 from app.api.v1 import websocket
 from app.core.config import settings
+from app.core.logging import configure_logging
+from app.core.request_context import set_request_id
 from app.core.websocket.broadcaster import broadcaster
 from app.db.session import init_db
 from app.services.scheduler import (
@@ -17,6 +20,9 @@ from app.services.scheduler import (
 )
 
 
+configure_logging()
+
+
 class WebSocketCORSMiddleware(BaseHTTPMiddleware):
     async def dispatch(self, request, call_next):
         if request.url.path.startswith("/ws") and request.method == "GET":
@@ -28,6 +34,18 @@ class WebSocketCORSMiddleware(BaseHTTPMiddleware):
         return await call_next(request)
 
 
+class RequestContextMiddleware(BaseHTTPMiddleware):
+    async def dispatch(self, request, call_next):
+        request_id = request.headers.get("X-Request-ID") or uuid4().hex
+        set_request_id(request_id)
+        try:
+            response = await call_next(request)
+        finally:
+            set_request_id(None)
+        response.headers["X-Request-ID"] = request_id
+        return response
+
+
 @asynccontextmanager
 async def lifespan(app: FastAPI):
     await init_db()
@@ -58,6 +76,7 @@ app.add_middleware(
     allow_headers=["*"],
 )
 
+app.add_middleware(RequestContextMiddleware)
 app.add_middleware(WebSocketCORSMiddleware)
 
 app.include_router(api_router, prefix="/api/v1")

backend/app/models/__init__.py

@@ -11,6 +11,7 @@ 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
+from app.models.system_log import SystemLog, AuditLog
 
 __all__ = [
     "User",
@@ -26,4 +27,6 @@ __all__ = [
     "BGPAnomaly",
     "BGPIncident",
     "BGPObservation",
+    "SystemLog",
+    "AuditLog",
 ]

backend/app/models/system_log.py

@@ -0,0 +1,40 @@
+from sqlalchemy import JSON, Column, DateTime, Integer, String, Text
+from sqlalchemy.sql import func
+
+from app.db.session import Base
+
+
+class SystemLog(Base):
+    __tablename__ = "system_logs"
+
+    id = Column(Integer, primary_key=True, autoincrement=True)
+    occurred_at = Column(DateTime(timezone=True), server_default=func.now(), index=True)
+    source = Column(String(50), nullable=False, index=True)
+    service = Column(String(50), nullable=True)
+    module = Column(String(120), nullable=True)
+    event = Column(String(160), nullable=True, index=True)
+    level = Column(String(20), nullable=False, index=True)
+    message = Column(Text, nullable=False)
+    request_id = Column(String(64), nullable=True, index=True)
+    trace_id = Column(String(64), nullable=True)
+    user_id = Column(Integer, nullable=True, index=True)
+    category = Column(String(80), nullable=True, index=True)
+    context = Column(JSON, nullable=False, default=dict)
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+
+
+class AuditLog(Base):
+    __tablename__ = "audit_logs"
+
+    id = Column(Integer, primary_key=True, autoincrement=True)
+    occurred_at = Column(DateTime(timezone=True), server_default=func.now(), index=True)
+    actor_id = Column(Integer, nullable=True, index=True)
+    actor_name = Column(String(255), nullable=True)
+    action = Column(String(120), nullable=False, index=True)
+    target_type = Column(String(80), nullable=True)
+    target_id = Column(String(120), nullable=True)
+    result = Column(String(40), nullable=True, index=True)
+    request_id = Column(String(64), nullable=True, index=True)
+    ip = Column(String(64), nullable=True)
+    details = Column(JSON, nullable=False, default=dict)
+    created_at = Column(DateTime(timezone=True), server_default=func.now())

backend/app/services/collectors/celestrak.py

@@ -46,6 +46,9 @@ class CelesTrakTLECollector(BaseCollector):
                     if response.status_code == 200:
                         data = response.json()
                         if isinstance(data, list):
+                            for item in data:
+                                if isinstance(item, dict):
+                                    item["_celestrak_group"] = group
                             all_satellites.extend(data)
                             print(f"CelesTrak: Fetched {len(data)} satellites from group '{group}'")
                 except Exception as e:
@@ -78,6 +81,7 @@ class CelesTrakTLECollector(BaseCollector):
                     "name": item.get("OBJECT_NAME", "Unknown"),
                     "reference_date": item.get("EPOCH", ""),
                     "metadata": {
+                        "constellation_group": item.get("_celestrak_group"),
                         "norad_cat_id": item.get("NORAD_CAT_ID"),
                         "international_designator": item.get("OBJECT_ID"),
                         "epoch": item.get("EPOCH"),

backend/app/services/collectors/news_live_streams.py

@@ -1,10 +1,16 @@
 from __future__ import annotations
 
+import asyncio
+import base64
 from datetime import UTC, datetime
 from typing import Any
+from urllib.parse import urlparse
 
 import httpx
+from sqlalchemy import select
 
+from app.core.data_sources import get_data_sources_config
+from app.models.datasource_config import DataSourceConfig
 from app.services.collectors.base import BaseCollector
 
 
@@ -18,52 +24,537 @@ class NewsLiveStreamsCollector(BaseCollector):
     data_type = "news_live_stream"
     fail_on_empty = False
 
+    DEFAULT_TIMEOUT = 45.0
+    DEFAULT_HEADERS = {
+        "User-Agent": "Planet-Intelligence-System/1.0 (Python/collector)",
+        "Accept": "application/json",
+    }
+    RESPONSE_CANDIDATE_KEYS = ("sources", "streams", "channels", "items", "results", "data")
+    DEFAULT_ADAPTER = "iptv_org"
+    DEFAULT_IPTV_ORG_STREAMS_URL = "https://iptv-org.github.io/api/streams.json"
+    DEFAULT_IPTV_ORG_LOGOS_URL = "https://iptv-org.github.io/api/logos.json"
+    DEFAULT_IPTV_ORG_NEWS_CATEGORIES = ("news", "business", "weather")
+    DEFAULT_IPTV_ORG_EXCLUDE_CATEGORIES = ("music", "sports", "kids", "entertainment")
+    DEFAULT_IPTV_ORG_MAX_SOURCES = 120
+
     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(
+        datasource_config = await self._load_datasource_config()
+        effective_config = self._get_effective_config(datasource_config)
+        adapter = str(effective_config.get("adapter") or "").strip().lower()
+        if adapter == "iptv_org":
+            return await self._fetch_iptv_org(request_url, effective_config)
+
+        request_headers = self._build_request_headers(datasource_config)
+        request_config = self._get_request_config(datasource_config)
+        request_params = self._build_request_params(datasource_config)
+        request_json = self._build_request_json_body(datasource_config)
+        request_data = self._build_request_form_body(datasource_config)
+        timeout = self._get_timeout(datasource_config)
+
+        async with httpx.AsyncClient(timeout=timeout, follow_redirects=True) as client:
+            response = await client.request(
+                request_config["method"],
                 request_url,
-                headers={
-                    "User-Agent": "Planet-Intelligence-System/1.0 (Python/collector)",
-                    "Accept": "application/json",
-                },
+                headers=request_headers,
+                params=request_params or None,
+                json=request_json,
+                data=request_data,
             )
             response.raise_for_status()
-            return self.parse_response(response.json())
+            return self.parse_response(
+                response.json(),
+                response_path=request_config["response_path"],
+            )
 
-    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
+    async def _load_datasource_config(self) -> DataSourceConfig | None:
+        if not self._db_session:
+            return None
+
+        result = await self._db_session.execute(
+            select(DataSourceConfig)
+            .where(DataSourceConfig.name == self.name)
+            .where(DataSourceConfig.is_active.is_(True))
+            .limit(1)
+        )
+        return result.scalar_one_or_none()
+
+    def _get_effective_config(self, datasource_config: DataSourceConfig | None) -> dict[str, Any]:
+        payload = dict(datasource_config.config or {}) if datasource_config else {}
+        if payload:
+            return payload
+
+        yaml_config = get_data_sources_config()
+        return {
+            "adapter": self.DEFAULT_ADAPTER,
+            "streams_url": yaml_config.get_yaml_value("news_live_streams.streams_url")
+            or self.DEFAULT_IPTV_ORG_STREAMS_URL,
+            "logos_url": yaml_config.get_yaml_value("news_live_streams.logos_url")
+            or self.DEFAULT_IPTV_ORG_LOGOS_URL,
+            "news_categories": list(self.DEFAULT_IPTV_ORG_NEWS_CATEGORIES),
+            "exclude_categories": list(self.DEFAULT_IPTV_ORG_EXCLUDE_CATEGORIES),
+            "max_sources": self.DEFAULT_IPTV_ORG_MAX_SOURCES,
+        }
+
+    def _get_request_config(self, datasource_config: DataSourceConfig | None) -> dict[str, Any]:
+        payload = self._get_effective_config(datasource_config)
+        raw_method = payload.get("method") or payload.get("request_method") or "GET"
+        method = str(raw_method).strip().upper() or "GET"
+        if method not in {"GET", "POST"}:
+            method = "GET"
+
+        response_path = payload.get("response_path") or payload.get("payload_path") or payload.get("items_path")
+        if isinstance(response_path, str):
+            response_path = response_path.strip()
         else:
-            candidates = []
+            response_path = None
+
+        return {
+            "method": method,
+            "response_path": response_path or None,
+        }
+
+    def _get_timeout(self, datasource_config: DataSourceConfig | None) -> float:
+        payload = self._get_effective_config(datasource_config)
+        try:
+            return float(payload.get("timeout", self.DEFAULT_TIMEOUT))
+        except (TypeError, ValueError):
+            return self.DEFAULT_TIMEOUT
+
+    def _build_request_headers(self, datasource_config: DataSourceConfig | None) -> dict[str, str]:
+        headers = dict(self.DEFAULT_HEADERS)
+        if datasource_config:
+            headers.update(self._normalize_headers(datasource_config.headers))
+            headers.update(self._build_auth_headers(datasource_config))
+        return headers
+
+    def _build_request_params(self, datasource_config: DataSourceConfig | None) -> dict[str, Any]:
+        params: dict[str, Any] = {}
+        if not datasource_config:
+            return params
+
+        payload = datasource_config.config or {}
+        candidate = payload.get("params") or payload.get("query_params")
+        if isinstance(candidate, dict):
+            params.update(candidate)
+
+        if datasource_config.auth_type == "api_key":
+            auth_config = datasource_config.auth_config or {}
+            if str(auth_config.get("in") or auth_config.get("location") or "header").lower() == "query":
+                api_key = auth_config.get("api_key")
+                key_name = auth_config.get("key_name") or auth_config.get("param_name") or "api_key"
+                if api_key and key_name:
+                    params[str(key_name)] = api_key
+
+        return params
+
+    def _build_request_json_body(self, datasource_config: DataSourceConfig | None) -> Any:
+        if not datasource_config:
+            return None
+
+        payload = datasource_config.config or {}
+        body = payload.get("json_body")
+        if body is None and str(payload.get("body_type") or "").lower() in {"json", ""}:
+            candidate = payload.get("body")
+            if isinstance(candidate, (dict, list)):
+                body = candidate
+        return body
+
+    def _build_request_form_body(self, datasource_config: DataSourceConfig | None) -> Any:
+        if not datasource_config:
+            return None
+
+        payload = datasource_config.config or {}
+        form_body = payload.get("form_body")
+        if form_body is not None:
+            return form_body
+
+        if str(payload.get("body_type") or "").lower() == "form":
+            candidate = payload.get("body")
+            if isinstance(candidate, dict):
+                return candidate
+        return None
+
+    def _normalize_headers(self, headers: Any) -> dict[str, str]:
+        if not isinstance(headers, dict):
+            return {}
+        normalized: dict[str, str] = {}
+        for key, value in headers.items():
+            header_name = str(key).strip()
+            if not header_name or value is None:
+                continue
+            normalized[header_name] = str(value)
+        return normalized
+
+    def _build_auth_headers(self, datasource_config: DataSourceConfig | None) -> dict[str, str]:
+        if not datasource_config:
+            return {}
+
+        auth_type = str(datasource_config.auth_type or "none").lower()
+        auth_config = datasource_config.auth_config or {}
+        if auth_type == "bearer" and auth_config.get("token"):
+            return {"Authorization": f"Bearer {auth_config['token']}"}
+
+        if auth_type == "api_key" and auth_config.get("api_key"):
+            location = str(auth_config.get("in") or auth_config.get("location") or "header").lower()
+            if location == "query":
+                return {}
+            key_name = auth_config.get("key_name") or "X-API-Key"
+            return {str(key_name): str(auth_config["api_key"])}
+
+        if auth_type == "basic":
+            username = str(auth_config.get("username") or "")
+            password = str(auth_config.get("password") or "")
+            encoded = base64.b64encode(f"{username}:{password}".encode()).decode()
+            return {"Authorization": f"Basic {encoded}"}
+
+        return {}
+
+    def _extract_candidates(self, response: Any, response_path: str | None) -> list[Any]:
+        if response_path:
+            extracted = self._extract_from_path(response, response_path)
+            if isinstance(extracted, list):
+                return extracted
+            if isinstance(extracted, dict):
+                for key in self.RESPONSE_CANDIDATE_KEYS:
+                    nested = extracted.get(key)
+                    if isinstance(nested, list):
+                        return nested
+                return [extracted]
+
+        if isinstance(response, dict):
+            for key in self.RESPONSE_CANDIDATE_KEYS:
+                nested = response.get(key)
+                if isinstance(nested, list):
+                    return nested
+            return []
+
+        if isinstance(response, list):
+            return response
+        return []
+
+    def _extract_from_path(self, payload: Any, path: str) -> Any:
+        current = payload
+        for segment in (part.strip() for part in path.split(".") if part.strip()):
+            if isinstance(current, dict):
+                current = current.get(segment)
+                continue
+            if isinstance(current, list):
+                try:
+                    current = current[int(segment)]
+                except (TypeError, ValueError, IndexError):
+                    return None
+                continue
+            return None
+        return current
+
+    def _infer_source_type(self, item: dict[str, Any]) -> str:
+        explicit = str(item.get("source_type") or item.get("type") or "").strip().lower()
+        if explicit in {"iframe", "hls", "video", "external", "youtube"}:
+            return explicit
+
+        youtube_video_id = self._clean_text(
+            item.get("youtube_video_id")
+            or item.get("video_id")
+            or item.get("youtubeVideoId")
+        )
+        youtube_channel = self._clean_text(item.get("youtube_channel") or item.get("channel_handle"))
+        embed_url = self._clean_url(item.get("embed_url") or item.get("embed") or item.get("page_url"))
+        stream_url = self._clean_url(item.get("stream_url") or item.get("stream") or item.get("playback_url") or item.get("hls_url"))
+        homepage_url = self._clean_url(item.get("homepage_url") or item.get("source_url") or item.get("website"))
+
+        if youtube_video_id or youtube_channel:
+            return "youtube"
+        if stream_url.endswith(".m3u8"):
+            return "hls"
+        if stream_url:
+            return "video"
+        if embed_url:
+            parsed = urlparse(embed_url)
+            if "youtube.com" in (parsed.netloc or "") or "youtu.be" in (parsed.netloc or ""):
+                return "youtube"
+            return "iframe"
+        if homepage_url:
+            return "external"
+        return "iframe"
+
+    def _parse_enabled(self, item: dict[str, Any]) -> bool:
+        if "is_enabled" in item:
+            return self._to_bool(item.get("is_enabled"), default=True)
+        if "enabled" in item:
+            return self._to_bool(item.get("enabled"), default=True)
+        if "active" in item:
+            return self._to_bool(item.get("active"), default=True)
+        if "status" in item:
+            status = str(item.get("status") or "").strip().lower()
+            if status in {"disabled", "inactive", "offline"}:
+                return False
+            if status in {"enabled", "active", "online", "live"}:
+                return True
+        return True
+
+    def _to_bool(self, 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", "enabled", "active", "online", "live"}:
+                return True
+            if lowered in {"0", "false", "no", "off", "disabled", "inactive", "offline"}:
+                return False
+        return bool(value)
+
+    def _clean_text(self, value: Any) -> str:
+        if value is None:
+            return ""
+        return str(value).strip()
+
+    def _clean_url(self, value: Any) -> str:
+        text = self._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
+
+    async def _fetch_iptv_org(self, channels_url: str, collector_config: dict[str, Any]) -> list[dict[str, Any]]:
+        streams_url = self._clean_url(collector_config.get("streams_url")) or self.DEFAULT_IPTV_ORG_STREAMS_URL
+        logos_url = self._clean_url(collector_config.get("logos_url")) or self.DEFAULT_IPTV_ORG_LOGOS_URL
+        news_categories = {
+            self._clean_text(value).lower()
+            for value in (collector_config.get("news_categories") or self.DEFAULT_IPTV_ORG_NEWS_CATEGORIES)
+            if self._clean_text(value)
+        }
+        exclude_categories = {
+            self._clean_text(value).lower()
+            for value in (collector_config.get("exclude_categories") or self.DEFAULT_IPTV_ORG_EXCLUDE_CATEGORIES)
+            if self._clean_text(value)
+        }
+        try:
+            max_sources = int(collector_config.get("max_sources", self.DEFAULT_IPTV_ORG_MAX_SOURCES))
+        except (TypeError, ValueError):
+            max_sources = self.DEFAULT_IPTV_ORG_MAX_SOURCES
+
+        timeout = self.DEFAULT_TIMEOUT
+        try:
+            timeout = float(collector_config.get("timeout", self.DEFAULT_TIMEOUT))
+        except (TypeError, ValueError):
+            timeout = self.DEFAULT_TIMEOUT
+
+        async with httpx.AsyncClient(timeout=timeout, follow_redirects=True) as client:
+            channels_payload, streams_payload, logos_payload = await self._gather_iptv_org_payloads(
+                client,
+                channels_url,
+                streams_url,
+                logos_url,
+            )
+
+        channels = channels_payload if isinstance(channels_payload, list) else []
+        streams = streams_payload if isinstance(streams_payload, list) else []
+        logos = logos_payload if isinstance(logos_payload, list) else []
+
+        logo_by_channel = {
+            self._clean_text(item.get("channel")): self._clean_url(item.get("url"))
+            for item in logos
+            if isinstance(item, dict) and self._clean_text(item.get("channel")) and self._clean_url(item.get("url"))
+        }
+
+        streams_by_channel: dict[str, list[dict[str, Any]]] = {}
+        for stream in streams:
+            if not isinstance(stream, dict):
+                continue
+            channel_id = self._clean_text(stream.get("channel"))
+            if not channel_id:
+                continue
+            streams_by_channel.setdefault(channel_id, []).append(stream)
+
+        normalized: list[dict[str, Any]] = []
+        for channel in channels:
+            if not isinstance(channel, dict):
+                continue
+
+            categories = [
+                self._clean_text(value).lower()
+                for value in (channel.get("categories") or [])
+                if self._clean_text(value)
+            ]
+            if news_categories and not any(category in news_categories for category in categories):
+                continue
+            if exclude_categories and any(category in exclude_categories for category in categories):
+                continue
+            if channel.get("is_nsfw") is True:
+                continue
+            if channel.get("closed"):
+                continue
+
+            channel_id = self._clean_text(channel.get("id"))
+            if not channel_id:
+                continue
+
+            stream = self._pick_iptv_org_stream(streams_by_channel.get(channel_id) or [])
+            if not stream:
+                continue
+
+            stream_url = self._clean_url(stream.get("url"))
+            if not stream_url:
+                continue
+
+            name = self._clean_text(channel.get("name")) or channel_id
+            notes_parts = [
+                f"Imported from IPTV-org catalog ({channel_id})",
+                f"Categories: {', '.join(categories)}" if categories else "",
+                f"Quality: {self._clean_text(stream.get('quality'))}" if self._clean_text(stream.get("quality")) else "",
+            ]
+            metadata = {
+                "provider": self._clean_text(channel.get("network")) or "IPTV-org",
+                "region": self._clean_text(channel.get("country")) or "Global",
+                "language": "und",
+                "source_type": "hls" if stream_url.endswith(".m3u8") else "video",
+                "embed_url": "",
+                "stream_url": stream_url,
+                "homepage_url": self._clean_url(channel.get("website")),
+                "poster_url": logo_by_channel.get(channel_id, ""),
+                "youtube_video_id": "",
+                "youtube_channel": "",
+                "sort_order": 400 + len(normalized),
+                "notes": "; ".join(part for part in notes_parts if part),
+                "is_enabled": True,
+                "collector_adapter": "iptv_org",
+                "channel_id": channel_id,
+                "categories": categories,
+                "quality": self._clean_text(stream.get("quality")),
+                "stream_label": self._clean_text(stream.get("label") or stream.get("title")),
+                "stream_referrer": self._clean_text(stream.get("referrer")),
+                "stream_user_agent": self._clean_text(stream.get("user_agent")),
+            }
+
+            normalized.append(
+                {
+                    "source_id": channel_id,
+                    "name": name,
+                    "description": metadata["notes"],
+                    "metadata": metadata,
+                    "reference_date": datetime.now(UTC).isoformat(),
+                }
+            )
+            if len(normalized) >= max_sources:
+                break
+
+        return normalized
+
+    async def _gather_iptv_org_payloads(
+        self,
+        client: httpx.AsyncClient,
+        channels_url: str,
+        streams_url: str,
+        logos_url: str,
+    ) -> tuple[Any, Any, Any]:
+        headers = dict(self.DEFAULT_HEADERS)
+        channels_payload, streams_payload, logos_payload = await asyncio.gather(
+            client.get(channels_url, headers=headers),
+            client.get(streams_url, headers=headers),
+            client.get(logos_url, headers=headers),
+        )
+        channels_payload.raise_for_status()
+        streams_payload.raise_for_status()
+        logos_payload.raise_for_status()
+        return channels_payload.json(), streams_payload.json(), logos_payload.json()
+
+    def _pick_iptv_org_stream(self, streams: list[dict[str, Any]]) -> dict[str, Any] | None:
+        if not streams:
+            return None
+
+        def score(stream: dict[str, Any]) -> tuple[int, int]:
+            url = self._clean_url(stream.get("url"))
+            quality = self._clean_text(stream.get("quality")).lower()
+            quality_score = 0
+            if quality.endswith("p"):
+                try:
+                    quality_score = int(quality[:-1])
+                except ValueError:
+                    quality_score = 0
+            stream_score = 1000 if url.endswith(".m3u8") else 0
+            return stream_score, quality_score
+
+        sorted_streams = sorted(streams, key=score, reverse=True)
+        return sorted_streams[0]
+
+    def parse_response(self, response: Any, *, response_path: str | None = None) -> list[dict[str, Any]]:
+        candidates = self._extract_candidates(response, response_path)
 
         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()
+            stream_id = (
+                item.get("id")
+                or item.get("source_id")
+                or item.get("slug")
+                or item.get("channel_id")
+                or item.get("code")
+                or f"news-live-{index + 1}"
+            )
+            name = self._clean_text(
+                item.get("name")
+                or item.get("title")
+                or item.get("channel")
+                or item.get("display_name")
+                or f"News Live {index + 1}"
+            )
             if not name:
                 continue
 
+            source_type = self._infer_source_type(item)
+            stream_url = self._clean_url(
+                item.get("stream_url")
+                or item.get("stream")
+                or item.get("playback_url")
+                or item.get("hls_url")
+                or item.get("m3u8_url")
+            )
+            embed_url = self._clean_url(
+                item.get("embed_url")
+                or item.get("embed")
+                or item.get("page_url")
+                or (item.get("url") if source_type == "iframe" else "")
+            )
+            homepage_url = self._clean_url(
+                item.get("homepage_url")
+                or item.get("source_url")
+                or item.get("website")
+                or item.get("url")
+            )
             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 "",
+                "provider": self._clean_text(item.get("provider") or item.get("publisher") or item.get("network")) or "Collector",
+                "region": self._clean_text(item.get("region") or item.get("country") or item.get("market")) or "Global",
+                "language": self._clean_text(item.get("language") or item.get("lang") or item.get("locale")) or "und",
+                "source_type": source_type,
+                "embed_url": embed_url,
+                "stream_url": stream_url,
+                "homepage_url": homepage_url,
+                "poster_url": self._clean_url(item.get("poster_url") or item.get("thumbnail_url") or item.get("logo_url")),
+                "youtube_video_id": self._clean_text(
+                    item.get("youtube_video_id")
+                    or item.get("video_id")
+                    or item.get("youtubeVideoId")
+                ),
+                "youtube_channel": self._clean_text(
+                    item.get("youtube_channel")
+                    or item.get("channel_handle")
+                    or item.get("youtubeChannel")
+                ),
                 "sort_order": item.get("sort_order", 200 + index),
-                "notes": item.get("notes") or item.get("description") or "",
-                "is_enabled": item.get("is_enabled", True),
+                "notes": self._clean_text(item.get("notes") or item.get("description") or item.get("summary")),
+                "is_enabled": self._parse_enabled(item),
             }
 
             normalized.append(
@@ -72,7 +563,7 @@ class NewsLiveStreamsCollector(BaseCollector):
                     "name": name,
                     "description": metadata["notes"],
                     "metadata": metadata,
-                    "reference_date": item.get("reference_date", datetime.now(UTC).isoformat()),
+                    "reference_date": item.get("reference_date") or datetime.now(UTC).isoformat(),
                 }
             )
 

backend/app/services/earth_news.py

@@ -0,0 +1,540 @@
+from __future__ import annotations
+
+import asyncio
+from dataclasses import dataclass
+from datetime import UTC, datetime
+from email.utils import parsedate_to_datetime
+import hashlib
+import html
+import re
+from typing import Any
+from urllib.parse import quote
+import xml.etree.ElementTree as ET
+
+import httpx
+from bs4 import BeautifulSoup
+
+
+USER_AGENT = "PlanetEarthNewsBoard/1.0 (+https://planet.local)"
+REQUEST_TIMEOUT = 12.0
+MAX_ITEMS_PER_SOURCE = 6
+MAX_ITEMS_TOTAL = 12
+STALE_CACHE_MAX_AGE_SECONDS = 60 * 45
+
+
+@dataclass(frozen=True)
+class RegionProfile:
+    key: str
+    label: str
+    query: str
+    accent: str
+
+
+@dataclass(frozen=True)
+class RegionAnchor:
+    region: str
+    label: str
+    latitude: float
+    longitude: float
+
+
+@dataclass(frozen=True)
+class NewsFeedSource:
+    id: str
+    name: str
+    region: str
+    feed_url: str
+    homepage_url: str
+    source_type: str = "rss"
+    priority: int = 100
+
+
+@dataclass
+class ParsedNewsItem:
+    id: str
+    title: str
+    summary: str
+    url: str
+    source: str
+    feed_name: str
+    feed_region: str
+    homepage_url: str
+    published_at: datetime | None
+
+
+@dataclass
+class CachedRegionFeed:
+    region: str
+    fetched_at: datetime
+    items: list[ParsedNewsItem]
+    sources: list[NewsFeedSource]
+
+
+REGION_PROFILES: dict[str, RegionProfile] = {
+    "americas": RegionProfile(
+        key="americas",
+        label="美洲焦点",
+        query='Americas geopolitics OR Latin America OR "United States" OR Canada',
+        accent="#79d3ff",
+    ),
+    "europe": RegionProfile(
+        key="europe",
+        label="欧洲焦点",
+        query='Europe geopolitics OR EU OR NATO OR "Eastern Europe"',
+        accent="#8fd4ff",
+    ),
+    "middle-east-africa": RegionProfile(
+        key="middle-east-africa",
+        label="中东与非洲焦点",
+        query='"Middle East" OR Africa geopolitics OR Red Sea OR Gulf',
+        accent="#ffb56a",
+    ),
+    "asia-pacific": RegionProfile(
+        key="asia-pacific",
+        label="亚太焦点",
+        query='"Asia Pacific" OR Indo-Pacific OR China OR Japan OR Korea OR ASEAN',
+        accent="#78f2cf",
+    ),
+    "global": RegionProfile(
+        key="global",
+        label="全球焦点",
+        query='"world news" OR geopolitics OR "global affairs"',
+        accent="#d6e6ff",
+    ),
+}
+
+REGION_ANCHORS: dict[str, RegionAnchor] = {
+    "americas": RegionAnchor(
+        region="americas",
+        label="美洲",
+        latitude=37.0902,
+        longitude=-95.7129,
+    ),
+    "europe": RegionAnchor(
+        region="europe",
+        label="欧洲",
+        latitude=50.1109,
+        longitude=8.6821,
+    ),
+    "middle-east-africa": RegionAnchor(
+        region="middle-east-africa",
+        label="中东与非洲",
+        latitude=25.2048,
+        longitude=55.2708,
+    ),
+    "asia-pacific": RegionAnchor(
+        region="asia-pacific",
+        label="亚太",
+        latitude=1.3521,
+        longitude=103.8198,
+    ),
+    "global": RegionAnchor(
+        region="global",
+        label="全球",
+        latitude=20.0,
+        longitude=0.0,
+    ),
+}
+
+
+def _google_news_feed(query: str, *, hl: str, gl: str, ceid: str) -> str:
+    return (
+        "https://news.google.com/rss/search?q="
+        + quote(query, safe="")
+        + f"&hl={hl}&gl={gl}&ceid={ceid}"
+    )
+
+
+NEWS_FEED_SOURCES: tuple[NewsFeedSource, ...] = (
+    NewsFeedSource(
+        id="bbc-world",
+        name="BBC World",
+        region="global",
+        feed_url="https://feeds.bbci.co.uk/news/world/rss.xml",
+        homepage_url="https://www.bbc.com/news/world",
+        priority=10,
+    ),
+    NewsFeedSource(
+        id="dw-top",
+        name="DW Top Stories",
+        region="europe",
+        feed_url="https://rss.dw.com/rdf/rss-en-top",
+        homepage_url="https://www.dw.com/en/top-stories/s-9097",
+        priority=20,
+    ),
+    NewsFeedSource(
+        id="global-scan",
+        name="Global Monitor / World",
+        region="global",
+        feed_url=_google_news_feed(
+            REGION_PROFILES["global"].query,
+            hl="en-US",
+            gl="US",
+            ceid="US:en",
+        ),
+        homepage_url="https://news.google.com/",
+        source_type="aggregated",
+        priority=30,
+    ),
+    NewsFeedSource(
+        id="google-americas",
+        name="Global Monitor / Americas",
+        region="americas",
+        feed_url=_google_news_feed(
+            REGION_PROFILES["americas"].query,
+            hl="en-US",
+            gl="US",
+            ceid="US:en",
+        ),
+        homepage_url="https://news.google.com/",
+        source_type="aggregated",
+        priority=40,
+    ),
+    NewsFeedSource(
+        id="google-europe",
+        name="Global Monitor / Europe",
+        region="europe",
+        feed_url=_google_news_feed(
+            REGION_PROFILES["europe"].query,
+            hl="en-GB",
+            gl="GB",
+            ceid="GB:en",
+        ),
+        homepage_url="https://news.google.com/",
+        source_type="aggregated",
+        priority=40,
+    ),
+    NewsFeedSource(
+        id="google-mea",
+        name="Global Monitor / MEA",
+        region="middle-east-africa",
+        feed_url=_google_news_feed(
+            REGION_PROFILES["middle-east-africa"].query,
+            hl="en-US",
+            gl="US",
+            ceid="US:en",
+        ),
+        homepage_url="https://news.google.com/",
+        source_type="aggregated",
+        priority=40,
+    ),
+    NewsFeedSource(
+        id="google-apac",
+        name="Global Monitor / APAC",
+        region="asia-pacific",
+        feed_url=_google_news_feed(
+            REGION_PROFILES["asia-pacific"].query,
+            hl="en-SG",
+            gl="SG",
+            ceid="SG:en",
+        ),
+        homepage_url="https://news.google.com/",
+        source_type="aggregated",
+        priority=40,
+    ),
+)
+
+
+_REGION_CACHE: dict[str, CachedRegionFeed] = {}
+
+
+def determine_focus_region(lat: float | None, lon: float | None) -> str:
+    if lat is None or lon is None:
+        return "global"
+    if -170 <= lon <= -30:
+        return "americas"
+    if -30 < lon <= 45:
+        return "europe" if lat >= 30 else "middle-east-africa"
+    if 45 < lon <= 150:
+        return "middle-east-africa" if lat < 10 else "asia-pacific"
+    return "asia-pacific"
+
+
+def get_region_profile(region: str) -> RegionProfile:
+    return REGION_PROFILES.get(region, REGION_PROFILES["global"])
+
+
+def get_region_anchor(region: str) -> RegionAnchor:
+    return REGION_ANCHORS.get(region, REGION_ANCHORS["global"])
+
+
+def get_sources_for_region(region: str) -> list[NewsFeedSource]:
+    return sorted(
+        [source for source in NEWS_FEED_SOURCES if source.region in {"global", region}],
+        key=lambda source: (source.priority, source.name),
+    )
+
+
+def _strip_html(value: str) -> str:
+    if not value:
+        return ""
+    soup = BeautifulSoup(value, "html.parser")
+    return re.sub(r"\s+", " ", soup.get_text(" ", strip=True)).strip()
+
+
+def _truncate(value: str, limit: int = 180) -> str:
+    text = value.strip()
+    if len(text) <= limit:
+        return text
+    return text[: limit - 1].rstrip() + "…"
+
+
+def _normalize_source_name(raw: str, fallback: str) -> str:
+    text = html.unescape((raw or "").strip())
+    if " - " in text:
+        return text.split(" - ")[-1].strip() or fallback
+    return text or fallback
+
+
+def _parse_datetime(raw: str | None) -> datetime | None:
+    if not raw:
+        return None
+    text = raw.strip()
+    if not text:
+        return None
+
+    for parser in (
+        lambda value: parsedate_to_datetime(value),
+        lambda value: datetime.fromisoformat(value.replace("Z", "+00:00")),
+    ):
+        try:
+            parsed = parser(text)
+            if parsed.tzinfo is None:
+                parsed = parsed.replace(tzinfo=UTC)
+            return parsed.astimezone(UTC)
+        except Exception:
+            continue
+    return None
+
+
+def _extract_item_text(element: ET.Element, *names: str) -> str:
+    for name in names:
+        node = element.find(name)
+        if node is not None and node.text:
+            return node.text.strip()
+    return ""
+
+
+def _parse_feed_entries(xml_text: str, source: NewsFeedSource) -> list[ParsedNewsItem]:
+    root = ET.fromstring(xml_text)
+    items: list[ParsedNewsItem] = []
+
+    rss_items = root.findall("./channel/item")
+    atom_entries = root.findall("{http://www.w3.org/2005/Atom}entry")
+    nodes = rss_items or atom_entries
+
+    for node in nodes[:MAX_ITEMS_PER_SOURCE]:
+        if node.tag.endswith("entry"):
+            title = _extract_item_text(node, "{http://www.w3.org/2005/Atom}title")
+            summary = _extract_item_text(
+                node,
+                "{http://www.w3.org/2005/Atom}summary",
+                "{http://www.w3.org/2005/Atom}content",
+            )
+            link_node = node.find("{http://www.w3.org/2005/Atom}link")
+            link = link_node.get("href", "").strip() if link_node is not None else ""
+            published = _extract_item_text(
+                node,
+                "{http://www.w3.org/2005/Atom}updated",
+                "{http://www.w3.org/2005/Atom}published",
+            )
+        else:
+            title = _extract_item_text(node, "title")
+            summary = _extract_item_text(node, "description", "content")
+            link = _extract_item_text(node, "link")
+            published = _extract_item_text(node, "pubDate", "published", "updated")
+
+        clean_title = html.unescape(title).strip()
+        clean_summary = _truncate(_strip_html(summary), 180)
+        if not clean_title or not link:
+            continue
+
+        item_source = _normalize_source_name(clean_title, source.name)
+        display_title = clean_title
+        if source.source_type == "aggregated" and " - " in clean_title:
+            parts = clean_title.rsplit(" - ", 1)
+            display_title = parts[0].strip()
+            item_source = _normalize_source_name(parts[1], source.name)
+
+        items.append(
+            ParsedNewsItem(
+                id=f"{source.id}:{hashlib.sha1(link.encode('utf-8')).hexdigest()[:12]}",
+                title=display_title,
+                summary=clean_summary,
+                url=link,
+                source=item_source,
+                feed_name=source.name,
+                feed_region=source.region,
+                homepage_url=source.homepage_url,
+                published_at=_parse_datetime(published),
+            )
+        )
+
+    return items
+
+
+def _serialize_sources(sources: list[NewsFeedSource]) -> list[dict[str, Any]]:
+    return [
+        {
+            "id": source.id,
+            "name": source.name,
+            "region": source.region,
+            "homepage_url": source.homepage_url,
+        }
+        for source in sources
+    ]
+
+
+def _serialize_item(item: ParsedNewsItem, *, active_region: str) -> dict[str, Any]:
+    published_at = item.published_at
+    anchor = get_region_anchor(item.feed_region)
+    return {
+        "id": item.id,
+        "title": item.title,
+        "summary": item.summary,
+        "url": item.url,
+        "source": item.source,
+        "feed_name": item.feed_name,
+        "region": item.feed_region,
+        "homepage_url": item.homepage_url,
+        "published_at": published_at.isoformat().replace("+00:00", "Z") if published_at else None,
+        "latitude": anchor.latitude,
+        "longitude": anchor.longitude,
+        "location_label": anchor.label,
+        "location_inferred": True,
+        "is_focus_match": item.feed_region == active_region,
+    }
+
+
+def _build_payload(
+    *,
+    lat: float | None,
+    lon: float | None,
+    active_region: str,
+    items: list[ParsedNewsItem],
+    sources: list[NewsFeedSource],
+    errors: list[str],
+    stale: bool,
+    generated_at: datetime | None = None,
+) -> dict[str, Any]:
+    profile = get_region_profile(active_region)
+    timestamp = generated_at or datetime.now(UTC)
+    return {
+        "generated_at": timestamp.isoformat().replace("+00:00", "Z"),
+        "focus": {
+            "lat": lat,
+            "lon": lon,
+            "region": active_region,
+            "label": profile.label,
+            "accent": profile.accent,
+        },
+        "sources": _serialize_sources(sources),
+        "items": [_serialize_item(item, active_region=active_region) for item in items],
+        "errors": errors,
+        "stale": stale,
+    }
+
+
+def _rank_and_trim_items(items: list[ParsedNewsItem], *, active_region: str) -> list[ParsedNewsItem]:
+    deduped: dict[str, ParsedNewsItem] = {}
+    for item in items:
+        key = item.url.strip() or item.title.strip().lower()
+        if key not in deduped:
+            deduped[key] = item
+
+    return sorted(
+        deduped.values(),
+        key=lambda item: (
+            item.feed_region != active_region,
+            item.published_at is None,
+            -(item.published_at.timestamp() if item.published_at else 0),
+            item.feed_name,
+        ),
+    )[:MAX_ITEMS_TOTAL]
+
+
+def _get_cached_region_feed(region: str) -> CachedRegionFeed | None:
+    cached = _REGION_CACHE.get(region)
+    if not cached:
+        return None
+    age_seconds = (datetime.now(UTC) - cached.fetched_at).total_seconds()
+    if age_seconds > STALE_CACHE_MAX_AGE_SECONDS:
+        return None
+    return cached
+
+
+def _store_region_cache(region: str, *, items: list[ParsedNewsItem], sources: list[NewsFeedSource]) -> None:
+    _REGION_CACHE[region] = CachedRegionFeed(
+        region=region,
+        fetched_at=datetime.now(UTC),
+        items=list(items),
+        sources=list(sources),
+    )
+
+
+async def _fetch_source(
+    client: httpx.AsyncClient,
+    source: NewsFeedSource,
+) -> tuple[NewsFeedSource, list[ParsedNewsItem], str | None]:
+    try:
+        response = await client.get(source.feed_url)
+        response.raise_for_status()
+        return source, _parse_feed_entries(response.text, source), None
+    except Exception as exc:
+        return source, [], str(exc)
+
+
+async def get_earth_news_payload(lat: float | None = None, lon: float | None = None) -> dict[str, Any]:
+    active_region = determine_focus_region(lat, lon)
+    sources = get_sources_for_region(active_region)
+    errors: list[str] = []
+
+    async with httpx.AsyncClient(
+        timeout=REQUEST_TIMEOUT,
+        follow_redirects=True,
+        headers={"User-Agent": USER_AGENT},
+    ) as client:
+        results = await asyncio.gather(*(_fetch_source(client, source) for source in sources))
+
+    fetched_items: list[ParsedNewsItem] = []
+    for source, items, error in results:
+        if error:
+            errors.append(f"{source.name}: {error}")
+            continue
+        fetched_items.extend(items)
+
+    ranked_items = _rank_and_trim_items(fetched_items, active_region=active_region)
+    if ranked_items:
+        _store_region_cache(active_region, items=ranked_items, sources=sources)
+        return _build_payload(
+            lat=lat,
+            lon=lon,
+            active_region=active_region,
+            items=ranked_items,
+            sources=sources,
+            errors=errors,
+            stale=False,
+        )
+
+    cached = _get_cached_region_feed(active_region)
+    if cached:
+        return _build_payload(
+            lat=lat,
+            lon=lon,
+            active_region=active_region,
+            items=cached.items,
+            sources=cached.sources,
+            errors=errors,
+            stale=True,
+            generated_at=cached.fetched_at,
+        )
+
+    return _build_payload(
+        lat=lat,
+        lon=lon,
+        active_region=active_region,
+        items=[],
+        sources=sources,
+        errors=errors,
+        stale=False,
+    )

backend/app/services/persistent_logs.py

@@ -0,0 +1,86 @@
+from __future__ import annotations
+
+from typing import Any
+
+from app.core.logging import get_logger, sanitize_log_value
+from app.core.request_context import get_request_id
+from app.db.session import async_session_factory
+from app.models.system_log import AuditLog, SystemLog
+
+logger = get_logger(__name__)
+
+
+async def record_system_log(
+    *,
+    source: str,
+    level: str,
+    message: str,
+    service: str | None = None,
+    module: str | None = None,
+    event: str | None = None,
+    request_id: str | None = None,
+    trace_id: str | None = None,
+    user_id: int | None = None,
+    category: str | None = None,
+    context: dict[str, Any] | None = None,
+) -> None:
+    try:
+        async with async_session_factory() as session:
+            session.add(
+                SystemLog(
+                    source=source,
+                    service=service,
+                    module=module,
+                    event=event,
+                    level=level.lower(),
+                    message=str(sanitize_log_value(message)),
+                    request_id=request_id or get_request_id(),
+                    trace_id=trace_id,
+                    user_id=user_id,
+                    category=category,
+                    context=sanitize_log_value(context or {}),
+                )
+            )
+            await session.commit()
+    except Exception:
+        logger.exception_event(
+            "Failed to persist system log",
+            event="system_log.persist.failed",
+            context={"event_name": event, "source": source},
+        )
+
+
+async def record_audit_log(
+    *,
+    action: str,
+    actor_id: int | None = None,
+    actor_name: str | None = None,
+    target_type: str | None = None,
+    target_id: str | None = None,
+    result: str | None = None,
+    request_id: str | None = None,
+    ip: str | None = None,
+    details: dict[str, Any] | None = None,
+) -> None:
+    try:
+        async with async_session_factory() as session:
+            session.add(
+                AuditLog(
+                    actor_id=actor_id,
+                    actor_name=actor_name,
+                    action=action,
+                    target_type=target_type,
+                    target_id=target_id,
+                    result=result,
+                    request_id=request_id or get_request_id(),
+                    ip=ip,
+                    details=sanitize_log_value(details or {}),
+                )
+            )
+            await session.commit()
+    except Exception:
+        logger.exception_event(
+            "Failed to persist audit log",
+            event="audit_log.persist.failed",
+            context={"action": action},
+        )

backend/app/services/scheduler.py

@@ -1,7 +1,6 @@
 """Task Scheduler for running collection jobs."""
 
 import asyncio
-import logging
 from datetime import UTC, datetime, timedelta
 from typing import Any, Dict, Optional
 
@@ -9,13 +8,14 @@ from apscheduler.schedulers.asyncio import AsyncIOScheduler
 from apscheduler.triggers.interval import IntervalTrigger
 from sqlalchemy import select
 
+from app.core.logging import get_logger
 from app.db.session import async_session_factory
 from app.core.time import to_iso8601_utc
 from app.models.datasource import DataSource
 from app.models.task import CollectionTask
 from app.services.collectors.registry import collector_registry
 
-logger = logging.getLogger(__name__)
+logger = get_logger(__name__)
 
 scheduler = AsyncIOScheduler()
 RUNNING_TASK_GUARD_TIMEOUT_MINUTES = 90
@@ -54,7 +54,11 @@ async def _update_next_run_at(datasource: DataSource, session) -> None:
 async def _apply_datasource_schedule(datasource: DataSource, session) -> None:
     collector = collector_registry.get(datasource.source)
     if not collector:
-        logger.warning("Collector not found for datasource %s", datasource.source)
+        logger.warning_event(
+            "Collector not found for datasource",
+            event="collector.schedule.collector_missing",
+            context={"collector_name": datasource.source},
+        )
         return
 
     collector_registry.set_active(datasource.source, datasource.is_active)
@@ -72,13 +76,17 @@ async def _apply_datasource_schedule(datasource: DataSource, session) -> None:
             replace_existing=True,
             kwargs={"collector_name": datasource.source},
         )
-        logger.info(
-            "Scheduled collector: %s (every %sm)",
-            datasource.source,
-            datasource.frequency_minutes,
+        logger.info_event(
+            "Scheduled collector",
+            event="collector.schedule.updated",
+            context={"collector_name": datasource.source, "frequency_minutes": datasource.frequency_minutes},
         )
     else:
-        logger.info("Collector disabled: %s", datasource.source)
+        logger.info_event(
+            "Collector disabled",
+            event="collector.schedule.disabled",
+            context={"collector_name": datasource.source},
+        )
 
     await _update_next_run_at(datasource, session)
 
@@ -87,18 +95,30 @@ async def run_collector_task(collector_name: str):
     """Run a single collector task."""
     collector = collector_registry.get(collector_name)
     if not collector:
-        logger.error("Collector not found: %s", collector_name)
+        logger.error_event(
+            "Collector not found",
+            event="collector.run.collector_missing",
+            context={"collector_name": collector_name},
+        )
         return
 
     async with async_session_factory() as db:
         result = await db.execute(select(DataSource).where(DataSource.source == collector_name))
         datasource = result.scalar_one_or_none()
         if not datasource:
-            logger.error("Datasource not found for collector: %s", collector_name)
+            logger.error_event(
+                "Datasource not found for collector",
+                event="collector.run.datasource_missing",
+                context={"collector_name": collector_name},
+            )
             return
 
         if not datasource.is_active:
-            logger.info("Skipping disabled collector: %s", collector_name)
+            logger.info_event(
+                "Skipping disabled collector",
+                event="collector.run.skipped_disabled",
+                context={"collector_name": collector_name},
+            )
             return
 
         running_result = await db.execute(
@@ -122,10 +142,10 @@ async def run_collector_task(collector_name: str):
                 and (now - started_at) > timedelta(minutes=RUNNING_TASK_GUARD_TIMEOUT_MINUTES)
             )
             if not is_stale:
-                logger.warning(
-                    "Skipping collector %s trigger because task %s is already running",
-                    collector_name,
-                    existing_running.id,
+                logger.warning_event(
+                    "Skipping collector trigger because task is already running",
+                    event="collector.run.skipped_already_running",
+                    context={"collector_name": collector_name, "task_id": existing_running.id},
                 )
                 return
 
@@ -143,31 +163,47 @@ async def run_collector_task(collector_name: str):
                 else stale_reason
             )
             await db.commit()
-            logger.warning(
-                "Marked stale running task %s as failed before rerun of %s",
-                existing_running.id,
-                collector_name,
+            logger.warning_event(
+                "Marked stale running task as failed before rerun",
+                event="collector.run.stale_task_failed",
+                context={"collector_name": collector_name, "task_id": existing_running.id},
             )
 
         try:
             collector._datasource_id = datasource.id
-            logger.info("Running collector: %s (datasource_id=%s)", collector_name, datasource.id)
+            logger.info_event(
+                "Running collector",
+                event="collector.run.started",
+                context={"collector_name": collector_name, "datasource_id": datasource.id},
+            )
             task_result = await collector.run(db)
             datasource.last_run_at = datetime.now(UTC)
             datasource.last_status = task_result.get("status")
             await _update_next_run_at(datasource, db)
-            logger.info("Collector %s completed: %s", collector_name, task_result)
+            logger.info_event(
+                "Collector completed",
+                event="collector.run.completed",
+                context={"collector_name": collector_name, "datasource_id": datasource.id, "result": 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)
+            logger.warning_event(
+                "Collector cancelled by operator",
+                event="collector.run.cancelled",
+                context={"collector_name": collector_name, "datasource_id": datasource.id},
+            )
             raise
         except Exception as exc:
             datasource.last_run_at = datetime.now(UTC)
             datasource.last_status = "failed"
             await db.commit()
-            logger.exception("Collector %s failed: %s", collector_name, exc)
+            logger.exception_event(
+                "Collector failed",
+                event="collector.run.failed",
+                context={"collector_name": collector_name, "datasource_id": datasource.id, "error": str(exc)},
+            )
 
 
 async def cleanup_stale_running_tasks(max_age_hours: int = 2) -> int:
@@ -194,7 +230,11 @@ async def cleanup_stale_running_tasks(max_age_hours: int = 2) -> int:
 
         if stale_tasks:
             await db.commit()
-            logger.warning("Cleaned up %s stale running collection task(s)", len(stale_tasks))
+            logger.warning_event(
+                "Cleaned up stale running collection tasks",
+                event="collector.cleanup.stale_tasks_cleaned",
+                context={"count": len(stale_tasks)},
+            )
 
         return len(stale_tasks)
 
@@ -203,14 +243,14 @@ def start_scheduler() -> None:
     """Start the scheduler."""
     if not scheduler.running:
         scheduler.start()
-        logger.info("Scheduler started")
+        logger.info_event("Scheduler started", event="scheduler.started")
 
 
 def stop_scheduler() -> None:
     """Stop the scheduler."""
     if scheduler.running:
         scheduler.shutdown(wait=False)
-        logger.info("Scheduler stopped")
+        logger.info_event("Scheduler stopped", event="scheduler.stopped")
 
 
 async def sync_scheduler_with_datasources() -> None:
@@ -271,12 +311,20 @@ def run_collector_now(collector_name: str) -> bool:
     """Run a collector immediately (not scheduled)."""
     collector = collector_registry.get(collector_name)
     if not collector:
-        logger.error("Collector not found: %s", collector_name)
+        logger.error_event(
+            "Collector not found",
+            event="collector.trigger.collector_missing",
+            context={"collector_name": 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)
+        logger.warning_event(
+            "Collector is already running in-memory; skipping duplicate trigger",
+            event="collector.trigger.skipped_already_running",
+            context={"collector_name": collector_name},
+        )
         return False
 
     try:
@@ -289,10 +337,18 @@ def run_collector_now(collector_name: str) -> bool:
                 RUNNING_COLLECTOR_TASKS.pop(collector_name, None)
 
         task.add_done_callback(_cleanup_task)
-        logger.info("Triggered collector: %s", collector_name)
+        logger.info_event(
+            "Triggered collector",
+            event="collector.trigger.started",
+            context={"collector_name": collector_name},
+        )
         return True
     except Exception as exc:
-        logger.error("Failed to trigger collector %s: %s", collector_name, exc)
+        logger.error_event(
+            "Failed to trigger collector",
+            event="collector.trigger.failed",
+            context={"collector_name": collector_name, "error": str(exc)},
+        )
         return False
 
 

backend/app/services/system_logs.py

@@ -0,0 +1,532 @@
+from __future__ import annotations
+
+import json
+import re
+import shutil
+import subprocess
+
+from collections import Counter, deque
+from dataclasses import dataclass
+from datetime import UTC, datetime
+from pathlib import Path
+from typing import Any
+
+from app.core.security import redis_client
+
+DEFAULT_LOG_LINE_LIMIT = 200
+MAX_LOG_LINE_LIMIT = 1000
+BUFFER_LOG_LIMIT = 1000
+BUFFER_LOG_TTL_SECONDS = 7 * 24 * 60 * 60
+LOG_BUFFER_KEY_PREFIX = "planet:system_logs"
+
+LOG_LEVEL_ERROR = "error"
+LOG_LEVEL_WARNING = "warning"
+LOG_LEVEL_INFO = "info"
+LOG_LEVEL_DEBUG = "debug"
+LOG_LEVEL_ALL = "all"
+
+SUPPORTED_LOG_LEVELS = {
+    LOG_LEVEL_ALL,
+    LOG_LEVEL_ERROR,
+    LOG_LEVEL_WARNING,
+    LOG_LEVEL_INFO,
+    LOG_LEVEL_DEBUG,
+}
+
+LOG_LEVEL_ALIASES = {
+    "warn": LOG_LEVEL_WARNING,
+    "warning": LOG_LEVEL_WARNING,
+    "err": LOG_LEVEL_ERROR,
+    "error": LOG_LEVEL_ERROR,
+    "info": LOG_LEVEL_INFO,
+    "information": LOG_LEVEL_INFO,
+    "debug": LOG_LEVEL_DEBUG,
+    "trace": LOG_LEVEL_DEBUG,
+    "critical": LOG_LEVEL_ERROR,
+    "fatal": LOG_LEVEL_ERROR,
+}
+
+TIMESTAMP_FORMATS = (
+    "%Y-%m-%d %H:%M:%S.%f",
+    "%Y-%m-%d %H:%M:%S",
+    "%Y-%m-%dT%H:%M:%S.%f",
+    "%Y-%m-%dT%H:%M:%S",
+)
+
+LEVEL_PATTERNS = (
+    ("CRITICAL", LOG_LEVEL_ERROR),
+    ("FATAL", LOG_LEVEL_ERROR),
+    ("ERROR", LOG_LEVEL_ERROR),
+    ("WARNING", LOG_LEVEL_WARNING),
+    ("WARN", LOG_LEVEL_WARNING),
+    ("INFO", LOG_LEVEL_INFO),
+    ("DEBUG", LOG_LEVEL_DEBUG),
+    ("TRACE", LOG_LEVEL_DEBUG),
+)
+
+LEADING_LEVEL_PATTERN = re.compile(
+    r"^\s*(?:\[[^\]]+\]\s*)?(CRITICAL|FATAL|ERROR|WARNING|WARN|INFO|DEBUG|TRACE)\b[:\s-]*",
+    re.IGNORECASE,
+)
+EMBEDDED_LEVEL_PATTERN = re.compile(
+    r"\b(CRITICAL|FATAL|ERROR|WARNING|WARN|INFO|DEBUG|TRACE)\b",
+    re.IGNORECASE,
+)
+CONTROL_CHAR_PATTERN = re.compile(r"[\x00-\x08\x0b\x0c\x0e-\x1f\x7f]")
+
+
+@dataclass(frozen=True)
+class LogSource:
+    source_id: str
+    name: str
+    kind: str
+    location: str
+    description: str
+    category: str
+    status: str = "ok"
+    buffer_key: str | None = None
+    container_name: str | None = None
+
+
+@dataclass
+class StructuredLogEntry:
+    timestamp: datetime | None
+    level: str | None
+    display_line: str
+    raw_line: str
+    search_text: str
+
+
+@dataclass
+class DailyLogMarker:
+    date_token: str
+    total: int
+    dominant_level: str
+
+
+LOG_SOURCES: dict[str, LogSource] = {
+    "backend": LogSource(
+        source_id="backend",
+        name="后端服务",
+        kind="file",
+        location="/tmp/planet_backend.log",
+        description="FastAPI 后端、调度器和采集任务共享日志。",
+        category="service",
+    ),
+    "frontend": LogSource(
+        source_id="frontend",
+        name="前端开发服务",
+        kind="file",
+        location="/tmp/planet_frontend.log",
+        description="控制台与 Earth 前端开发服务输出。",
+        category="service",
+    ),
+    "ai-provider": LogSource(
+        source_id="ai-provider",
+        name="AI Provider",
+        kind="docker",
+        location="docker://planet_aiprovider",
+        description="AI Provider 容器实时输出日志。",
+        category="service",
+        container_name="planet_aiprovider",
+    ),
+    "earth-client": LogSource(
+        source_id="earth-client",
+        name="Earth 浏览器端",
+        kind="buffer",
+        location="redis://planet:system_logs:earth-client",
+        description="Earth 浏览器端上报的运行时错误与关键业务日志。",
+        category="client",
+        buffer_key=f"{LOG_BUFFER_KEY_PREFIX}:earth-client",
+    ),
+}
+
+
+def normalize_log_level(level: str | None) -> str:
+    if level is None:
+        return LOG_LEVEL_ALL
+    normalized = str(level).strip().lower()
+    if normalized in {"", LOG_LEVEL_ALL}:
+        return LOG_LEVEL_ALL
+    return LOG_LEVEL_ALIASES.get(normalized, LOG_LEVEL_ALL)
+
+
+def normalize_log_levels(level: str | None = None, levels: str | None = None) -> tuple[str, ...]:
+    normalized_levels: list[str] = []
+    if levels:
+        for item in str(levels).split(","):
+            normalized = normalize_log_level(item)
+            if normalized != LOG_LEVEL_ALL and normalized not in normalized_levels:
+                normalized_levels.append(normalized)
+    normalized_level = normalize_log_level(level)
+    if normalized_level != LOG_LEVEL_ALL and normalized_level not in normalized_levels:
+        normalized_levels.append(normalized_level)
+    return tuple(normalized_levels)
+
+
+def get_source_status(source: LogSource) -> str:
+    if source.kind == "file":
+        path = Path(source.location)
+        if not path.exists():
+            return "missing"
+        return "ok" if path.stat().st_size > 0 else "empty"
+    if source.kind == "docker":
+        return "ok" if shutil.which("docker") else "docker_unavailable"
+    if source.kind == "buffer":
+        if not source.buffer_key:
+            return "source_unavailable"
+        try:
+            return "ok" if redis_client.llen(source.buffer_key) > 0 else "empty"
+        except Exception:
+            return "source_unavailable"
+    return "source_unavailable"
+
+
+def list_log_sources() -> list[dict[str, str]]:
+    items: list[dict[str, str]] = []
+    for source in LOG_SOURCES.values():
+        items.append(
+            {
+                "source_id": source.source_id,
+                "name": source.name,
+                "kind": source.kind,
+                "location": source.location,
+                "description": source.description,
+                "category": source.category,
+                "status": get_source_status(source),
+            }
+        )
+    return items
+
+
+def get_buffer_log_key(source_id: str) -> str:
+    return f"{LOG_BUFFER_KEY_PREFIX}:{source_id}"
+
+
+def append_buffer_log(
+    source_id: str,
+    *,
+    level: str,
+    message: str,
+    context: dict[str, Any] | None = None,
+) -> None:
+    payload = {
+        "timestamp": datetime.now(tz=UTC).isoformat(),
+        "level": normalize_log_level(level),
+        "message": message,
+        "context": context or {},
+    }
+    buffer_key = get_buffer_log_key(source_id)
+    redis_client.rpush(buffer_key, json.dumps(payload, ensure_ascii=False))
+    redis_client.ltrim(buffer_key, -BUFFER_LOG_LIMIT, -1)
+    redis_client.expire(buffer_key, BUFFER_LOG_TTL_SECONDS)
+
+
+def parse_timestamp(raw_value: str | None) -> datetime | None:
+    if not raw_value:
+        return None
+    candidate = str(raw_value).strip()
+    if not candidate:
+        return None
+    candidate = candidate.replace("Z", "+00:00")
+    try:
+        parsed = datetime.fromisoformat(candidate)
+        return parsed if parsed.tzinfo else parsed.replace(tzinfo=UTC)
+    except ValueError:
+        pass
+    for fmt in TIMESTAMP_FORMATS:
+        try:
+            return datetime.strptime(candidate, fmt).replace(tzinfo=UTC)
+        except ValueError:
+            continue
+    return None
+
+
+def parse_prefixed_timestamp(line: str) -> tuple[datetime | None, str]:
+    stripped = line.strip()
+    if not stripped:
+        return None, ""
+    for prefix_length in (35, 32, 29, 26, 23, 19):
+        if len(stripped) < prefix_length:
+            continue
+        prefix = stripped[:prefix_length]
+        timestamp = parse_timestamp(prefix)
+        if timestamp is not None:
+            return timestamp, stripped[prefix_length:].lstrip()
+    first_token = stripped.split(maxsplit=1)[0]
+    timestamp = parse_timestamp(first_token)
+    if timestamp is not None:
+        remainder = stripped[len(first_token):].lstrip()
+        return timestamp, remainder
+    return None, stripped
+
+
+def infer_log_level_from_text(text: str, *, allow_embedded: bool = True) -> str | None:
+    leading_match = LEADING_LEVEL_PATTERN.match(text)
+    if leading_match:
+        return normalize_log_level(leading_match.group(1))
+
+    if allow_embedded:
+        embedded_match = EMBEDDED_LEVEL_PATTERN.search(text)
+        if embedded_match:
+            return normalize_log_level(embedded_match.group(1))
+        upper_text = text.upper()
+        for pattern, normalized in LEVEL_PATTERNS:
+            if f"{pattern}:" in upper_text or f"{pattern} " in upper_text:
+                return normalized
+    return None
+
+
+def build_display_line(timestamp: datetime | None, level: str | None, message: str) -> str:
+    message_part = message.strip() if message else ""
+    parts = []
+    if timestamp is not None:
+        parts.append(timestamp.astimezone(UTC).strftime("%Y-%m-%d %H:%M:%S"))
+    if level:
+        parts.append(level.upper())
+    if message_part:
+        parts.append(message_part)
+    return " ".join(parts).strip()
+
+
+def sanitize_text_log_line(line: str) -> str:
+    return CONTROL_CHAR_PATTERN.sub("", line)
+
+
+def parse_text_log_entry(line: str) -> StructuredLogEntry:
+    sanitized_line = sanitize_text_log_line(line).rstrip("\n")
+    timestamp, remainder = parse_prefixed_timestamp(sanitized_line)
+    level = infer_log_level_from_text(remainder or sanitized_line, allow_embedded=False)
+    display_line = sanitized_line
+    return StructuredLogEntry(
+        timestamp=timestamp,
+        level=level,
+        display_line=display_line,
+        raw_line=display_line,
+        search_text=display_line.lower(),
+    )
+
+
+def build_buffer_entry(payload: dict[str, Any]) -> StructuredLogEntry:
+    timestamp = parse_timestamp(str(payload.get("timestamp", "")).strip())
+    level = normalize_log_level(payload.get("level"))
+    if level == LOG_LEVEL_ALL:
+        level = None
+    message = str(payload.get("message", "")).strip()
+    context = payload.get("context")
+    context_map = context if isinstance(context, dict) else {}
+    context_fragments = []
+    for key in ("category", "module", "url", "detail"):
+        value = str(context_map.get(key, "")).strip()
+        if value:
+            context_fragments.append(f"{key}={value}")
+    message_with_context = " | ".join([message, *context_fragments]) if context_fragments else message
+    display_line = build_display_line(timestamp, level, message_with_context)
+    search_text = " ".join(
+        [
+            message,
+            json.dumps(context_map, ensure_ascii=False, sort_keys=True),
+            display_line,
+        ]
+    ).lower()
+    return StructuredLogEntry(
+        timestamp=timestamp,
+        level=level,
+        display_line=display_line,
+        raw_line=json.dumps(payload, ensure_ascii=False, sort_keys=True),
+        search_text=search_text,
+    )
+
+
+def read_file_entries(source: LogSource, scan_limit: int) -> list[StructuredLogEntry]:
+    path = Path(source.location)
+    if not path.exists():
+        return []
+    with path.open("r", encoding="utf-8", errors="replace") as handle:
+        recent_lines = deque(handle, maxlen=scan_limit)
+    return [
+        parse_text_log_entry(line)
+        for line in recent_lines
+        if sanitize_text_log_line(line).strip()
+    ]
+
+
+def read_docker_entries(source: LogSource, scan_limit: int) -> list[StructuredLogEntry]:
+    if not shutil.which("docker") or not source.container_name:
+        return []
+    try:
+        completed = subprocess.run(
+            [
+                "docker",
+                "logs",
+                "--timestamps",
+                "--tail",
+                str(scan_limit),
+                source.container_name,
+            ],
+            capture_output=True,
+            text=True,
+            check=False,
+        )
+    except OSError:
+        return []
+    if completed.returncode != 0:
+        return []
+    return [
+        parse_text_log_entry(line)
+        for line in completed.stdout.splitlines()
+        if line.strip()
+    ]
+
+
+def read_buffer_entries(source: LogSource, scan_limit: int) -> list[StructuredLogEntry]:
+    if not source.buffer_key:
+        return []
+    try:
+        raw_items = redis_client.lrange(source.buffer_key, -scan_limit, -1)
+    except Exception:
+        return []
+    entries: list[StructuredLogEntry] = []
+    for raw_item in raw_items:
+        try:
+            payload = json.loads(raw_item)
+        except json.JSONDecodeError:
+            entries.append(parse_text_log_entry(str(raw_item)))
+            continue
+        if isinstance(payload, dict):
+            entries.append(build_buffer_entry(payload))
+        else:
+            entries.append(parse_text_log_entry(str(raw_item)))
+    return entries
+
+
+def read_source_entries(source: LogSource, scan_limit: int) -> list[StructuredLogEntry]:
+    if source.kind == "file":
+        return read_file_entries(source, scan_limit)
+    if source.kind == "docker":
+        return read_docker_entries(source, scan_limit)
+    if source.kind == "buffer":
+        return read_buffer_entries(source, scan_limit)
+    return []
+
+
+def matches_levels(entry: StructuredLogEntry, selected_levels: tuple[str, ...]) -> bool:
+    if not selected_levels:
+        return True
+    return entry.level in selected_levels
+
+
+def matches_date_range(
+    entry: StructuredLogEntry,
+    start_date: str | None,
+    end_date: str | None,
+) -> bool:
+    if not start_date and not end_date:
+        return True
+    if entry.timestamp is None:
+        return False
+    date_token = entry.timestamp.astimezone(UTC).date().isoformat()
+    if start_date and date_token < start_date:
+        return False
+    if end_date and date_token > end_date:
+        return False
+    return True
+
+
+def matches_search(entry: StructuredLogEntry, search: str | None) -> bool:
+    if search is None:
+        return True
+    query = search.strip().lower()
+    if not query:
+        return True
+    return query in entry.search_text
+
+
+def build_daily_log_markers(entries: list[StructuredLogEntry]) -> list[dict[str, Any]]:
+    grouped: dict[str, list[StructuredLogEntry]] = {}
+    for entry in entries:
+        if entry.timestamp is None:
+            continue
+        date_token = entry.timestamp.astimezone(UTC).date().isoformat()
+        grouped.setdefault(date_token, []).append(entry)
+
+    markers: list[DailyLogMarker] = []
+    for date_token, group in sorted(grouped.items()):
+        level_counts = Counter(
+            entry.level
+            for entry in group
+            if entry.level in SUPPORTED_LOG_LEVELS and entry.level != LOG_LEVEL_ALL
+        )
+        dominant_level = LOG_LEVEL_INFO
+        if level_counts:
+            dominant_level = sorted(
+                level_counts.items(),
+                key=lambda item: (
+                    -item[1],
+                    ("error", "warning", "info", "debug").index(item[0]),
+                ),
+            )[0][0]
+        markers.append(
+            DailyLogMarker(
+                date_token=date_token,
+                total=len(group),
+                dominant_level=dominant_level,
+            )
+        )
+    return [marker.__dict__ for marker in markers]
+
+
+def read_log_snapshot(
+    source_id: str,
+    limit: int,
+    *,
+    level: str = LOG_LEVEL_ALL,
+    levels: str | None = None,
+    start_date: str | None = None,
+    end_date: str | None = None,
+    search: str | None = None,
+) -> dict[str, Any] | None:
+    source = LOG_SOURCES.get(source_id)
+    if source is None:
+        return None
+
+    selected_levels = normalize_log_levels(level, levels)
+    search_query = (search or "").strip()
+    scan_limit = max(min(MAX_LOG_LINE_LIMIT * 5, 5000), limit * 5, BUFFER_LOG_LIMIT if source.kind == "buffer" else 1000)
+    all_entries = read_source_entries(source, scan_limit)
+    marker_entries = [
+        entry
+        for entry in all_entries
+        if matches_levels(entry, selected_levels) and matches_search(entry, search_query)
+    ]
+    filtered_entries = [
+        entry
+        for entry in marker_entries
+        if matches_date_range(entry, start_date, end_date)
+    ]
+    visible_entries = filtered_entries[-limit:]
+
+    compatibility_level = selected_levels[0] if len(selected_levels) == 1 else LOG_LEVEL_ALL
+    return {
+        "source_id": source.source_id,
+        "name": source.name,
+        "kind": source.kind,
+        "location": source.location,
+        "description": source.description,
+        "category": source.category,
+        "status": get_source_status(source),
+        "level": compatibility_level,
+        "selected_levels": list(selected_levels),
+        "search_query": search_query,
+        "available_levels": [
+            LOG_LEVEL_ALL,
+            LOG_LEVEL_ERROR,
+            LOG_LEVEL_WARNING,
+            LOG_LEVEL_INFO,
+            LOG_LEVEL_DEBUG,
+        ],
+        "daily_markers": build_daily_log_markers(marker_entries),
+        "line_limit": limit,
+        "line_count": len(visible_entries),
+        "lines": [entry.display_line for entry in visible_entries],
+    }

backend/app/services/tv_streams.py

@@ -17,7 +17,7 @@ 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,
+    "default_source_id": DEFAULT_TV_SOURCE_ID,
     "auto_fallback": True,
     "sources": [
         {
@@ -362,7 +362,7 @@ def _build_collected_tv_source(record: CollectedData, index: int) -> dict[str, A
             "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)),
+            "updated_at": to_iso8601_utc(record.collected_at or record.reference_date or datetime.now(UTC)),
         },
         index=index,
     )

backend/tests/test_api.py

@@ -35,6 +35,7 @@ async def test_health_check():
     data = response.json()
     assert data["status"] == "healthy"
     assert "version" in data
+    assert response.headers["x-request-id"]
 
 
 @pytest.mark.asyncio
@@ -161,6 +162,345 @@ async def test_alerts_endpoint_with_auth(auth_headers):
         app.dependency_overrides.clear()
 
 
+@pytest.mark.asyncio
+async def test_system_log_sources_requires_super_admin(auth_headers):
+    def override_get_current_user():
+        return User(
+            id=1,
+            username="testuser",
+            email="test@example.com",
+            password_hash="hashed",
+            role="admin",
+            is_active=True,
+        )
+
+    app.dependency_overrides = {
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+    }
+    transport = ASGITransport(app=app)
+    try:
+        async with AsyncClient(transport=transport, base_url="http://test") as client:
+            response = await client.get("/api/v1/system/logs/sources", headers=auth_headers)
+        assert response.status_code == 403
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_system_log_sources_with_super_admin(auth_headers):
+    def override_get_current_user():
+        return User(
+            id=1,
+            username="root",
+            email="root@example.com",
+            password_hash="hashed",
+            role="super_admin",
+            is_active=True,
+        )
+
+    app.dependency_overrides = {
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+    }
+    transport = ASGITransport(app=app)
+    try:
+        with patch(
+            "app.api.v1.system_control.list_log_sources",
+            return_value=[
+                {
+                    "source_id": "backend",
+                    "name": "后端服务",
+                    "kind": "file",
+                    "location": "/tmp/planet_backend.log",
+                    "description": "FastAPI 后端、调度器和采集任务共享日志。",
+                    "category": "service",
+                    "status": "ok",
+                }
+            ],
+        ):
+            async with AsyncClient(transport=transport, base_url="http://test") as client:
+                response = await client.get("/api/v1/system/logs/sources", headers=auth_headers)
+        assert response.status_code == 200
+        data = response.json()
+        assert data["items"][0]["source_id"] == "backend"
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_system_log_snapshot_with_super_admin(auth_headers):
+    def override_get_current_user():
+        return User(
+            id=1,
+            username="root",
+            email="root@example.com",
+            password_hash="hashed",
+            role="super_admin",
+            is_active=True,
+        )
+
+    app.dependency_overrides = {
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+    }
+    transport = ASGITransport(app=app)
+    try:
+        with patch(
+            "app.api.v1.system_control.read_log_snapshot",
+            return_value={
+                "source_id": "backend",
+                "name": "后端服务",
+                "kind": "file",
+                "location": "/tmp/planet_backend.log",
+                "description": "FastAPI 后端、调度器和采集任务共享日志。",
+                "category": "service",
+                "status": "ok",
+                "level": "all",
+                "selected_levels": [],
+                "search_query": "",
+                "available_levels": ["all", "error", "warning", "info", "debug"],
+                "daily_markers": [],
+                "line_limit": 50,
+                "line_count": 2,
+                "lines": ["line 1", "line 2"],
+            },
+        ):
+            async with AsyncClient(transport=transport, base_url="http://test") as client:
+                response = await client.get("/api/v1/system/logs/backend?limit=50", headers=auth_headers)
+        assert response.status_code == 200
+        data = response.json()
+        assert data["source_id"] == "backend"
+        assert data["line_count"] == 2
+        assert data["lines"] == ["line 1", "line 2"]
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_system_log_snapshot_supports_level_filter(auth_headers):
+    def override_get_current_user():
+        return User(
+            id=1,
+            username="root",
+            email="root@example.com",
+            password_hash="hashed",
+            role="super_admin",
+            is_active=True,
+        )
+
+    app.dependency_overrides = {
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+    }
+    transport = ASGITransport(app=app)
+    try:
+        with patch(
+            "app.api.v1.system_control.read_log_snapshot",
+            return_value={
+                "source_id": "backend",
+                "name": "后端服务",
+                "kind": "file",
+                "location": "/tmp/planet_backend.log",
+                "description": "FastAPI 后端、调度器和采集任务共享日志。",
+                "category": "service",
+                "status": "ok",
+                "level": "error",
+                "selected_levels": ["error"],
+                "search_query": "",
+                "available_levels": ["all", "error", "warning", "info", "debug"],
+                "daily_markers": [],
+                "line_limit": 50,
+                "line_count": 1,
+                "lines": ["ERROR: failed"],
+            },
+        ):
+            async with AsyncClient(transport=transport, base_url="http://test") as client:
+                response = await client.get("/api/v1/system/logs/backend?limit=50&level=error", headers=auth_headers)
+        assert response.status_code == 200
+        data = response.json()
+        assert data["level"] == "error"
+        assert data["lines"] == ["ERROR: failed"]
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_system_log_snapshot_supports_date_range_filter(auth_headers):
+    def override_get_current_user():
+        return User(
+            id=1,
+            username="root",
+            email="root@example.com",
+            password_hash="hashed",
+            role="super_admin",
+            is_active=True,
+        )
+
+    app.dependency_overrides = {
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+    }
+    transport = ASGITransport(app=app)
+    try:
+        with patch(
+            "app.api.v1.system_control.read_log_snapshot",
+            return_value={
+                "source_id": "backend",
+                "name": "后端服务",
+                "kind": "file",
+                "location": "/tmp/planet_backend.log",
+                "description": "FastAPI 后端、调度器和采集任务共享日志。",
+                "category": "service",
+                "status": "ok",
+                "level": "all",
+                "selected_levels": [],
+                "search_query": "",
+                "available_levels": ["all", "error", "warning", "info", "debug"],
+                "daily_markers": [],
+                "line_limit": 50,
+                "line_count": 1,
+                "lines": ["2026-04-23 INFO: service started"],
+            },
+        ) as mock_read_log_snapshot:
+            async with AsyncClient(transport=transport, base_url="http://test") as client:
+                response = await client.get(
+                    "/api/v1/system/logs/backend?limit=50&start_date=2026-04-20&end_date=2026-04-23",
+                    headers=auth_headers,
+                )
+        assert response.status_code == 200
+        mock_read_log_snapshot.assert_called_once_with(
+            "backend",
+            50,
+            level="all",
+            levels=None,
+            start_date="2026-04-20",
+            end_date="2026-04-23",
+            search=None,
+        )
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_system_log_snapshot_supports_levels_and_search_filter(auth_headers):
+    def override_get_current_user():
+        return User(
+            id=1,
+            username="root",
+            email="root@example.com",
+            password_hash="hashed",
+            role="super_admin",
+            is_active=True,
+        )
+
+    app.dependency_overrides = {
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+    }
+    transport = ASGITransport(app=app)
+    try:
+        with patch(
+            "app.api.v1.system_control.read_log_snapshot",
+            return_value={
+                "source_id": "backend",
+                "name": "后端服务",
+                "kind": "file",
+                "location": "/tmp/planet_backend.log",
+                "description": "FastAPI 后端、调度器和采集任务共享日志。",
+                "category": "service",
+                "status": "ok",
+                "level": "all",
+                "selected_levels": ["error", "warning"],
+                "search_query": "timeout",
+                "available_levels": ["all", "error", "warning", "info", "debug"],
+                "daily_markers": [],
+                "line_limit": 50,
+                "line_count": 1,
+                "lines": ["2026-04-23 10:00:00 ERROR timeout"],
+            },
+        ) as mock_read_log_snapshot:
+            async with AsyncClient(transport=transport, base_url="http://test") as client:
+                response = await client.get(
+                    "/api/v1/system/logs/backend?limit=50&levels=error,warning&search=timeout",
+                    headers=auth_headers,
+                )
+        assert response.status_code == 200
+        mock_read_log_snapshot.assert_called_once_with(
+            "backend",
+            50,
+            level="all",
+            levels="error,warning",
+            start_date=None,
+            end_date=None,
+            search="timeout",
+        )
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_system_log_snapshot_rejects_invalid_date_range(auth_headers):
+    def override_get_current_user():
+        return User(
+            id=1,
+            username="root",
+            email="root@example.com",
+            password_hash="hashed",
+            role="super_admin",
+            is_active=True,
+        )
+
+    app.dependency_overrides = {
+        __import__("app.core.security", fromlist=["get_current_user"]).get_current_user: override_get_current_user,
+    }
+    transport = ASGITransport(app=app)
+    try:
+        async with AsyncClient(transport=transport, base_url="http://test") as client:
+            response = await client.get(
+                "/api/v1/system/logs/backend?start_date=2026-04-31",
+                headers=auth_headers,
+            )
+        assert response.status_code == 400
+        assert "start_date must be in YYYY-MM-DD format" in response.json()["detail"]
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_ingest_earth_client_log_accepts_public_events():
+    transport = ASGITransport(app=app)
+    try:
+        with patch("app.api.v1.system_control.append_buffer_log") as mock_append_buffer_log:
+            with patch("app.api.v1.system_control.record_system_log", new_callable=AsyncMock) as mock_record_system_log:
+                async with AsyncClient(transport=transport, base_url="http://test") as client:
+                    response = await client.post(
+                        "/api/v1/system/logs/earth-client",
+                        json={
+                            "level": "error",
+                            "message": "登陆点加载失败: 登陆点接口返回 HTTP 500",
+                            "category": "startup-load",
+                            "module": "layer-startup",
+                        },
+                    )
+        assert response.status_code == 200
+        data = response.json()
+        assert data["accepted"] is True
+        assert data["source_id"] == "earth-client"
+        mock_append_buffer_log.assert_called_once()
+        mock_record_system_log.assert_awaited_once()
+        persisted_kwargs = mock_record_system_log.await_args.kwargs
+        assert persisted_kwargs["source"] == "earth-client"
+        assert persisted_kwargs["event"] == "earth.client.runtime_log"
+        assert persisted_kwargs["category"] == "startup-load"
+        assert persisted_kwargs["level"] == "error"
+    finally:
+        app.dependency_overrides.clear()
+
+
+@pytest.mark.asyncio
+async def test_request_id_header_is_echoed_when_provided():
+    transport = ASGITransport(app=app)
+    async with AsyncClient(transport=transport, base_url="http://test") as client:
+        response = await client.get("/health", headers={"X-Request-ID": "planet-test-request"})
+    assert response.status_code == 200
+    assert response.headers["x-request-id"] == "planet-test-request"
+
+
 @pytest.mark.asyncio
 async def test_invalid_token():
     """Test that invalid token is rejected"""
@@ -263,6 +603,8 @@ async def test_ai_situational_analysis_returns_503_when_disabled(auth_headers):
         assert "content_blocks" in data
         assert "text_blocks" in data
         assert "thinking_blocks" in data
+    finally:
+        app.dependency_overrides.clear()
 
 
 @pytest.mark.asyncio
@@ -382,8 +724,6 @@ async def test_save_playground_session_with_auth(auth_headers):
         assert data["state"]["objective"] == "测试目标"
     finally:
         app.dependency_overrides.clear()
-    finally:
-        app.dependency_overrides.clear()
 
 
 @pytest.mark.asyncio

backend/tests/test_earth_news.py

@@ -0,0 +1,49 @@
+from datetime import UTC, datetime
+
+from app.services.earth_news import ParsedNewsItem, _serialize_item
+
+
+def test_serialize_item_includes_region_anchor_for_cruise():
+    item = ParsedNewsItem(
+        id="google-apac:test",
+        title="Example APAC story",
+        summary="Example summary",
+        url="https://example.com/story",
+        source="Example Source",
+        feed_name="Global Monitor / APAC",
+        feed_region="asia-pacific",
+        homepage_url="https://example.com",
+        published_at=datetime(2026, 4, 23, 2, 30, tzinfo=UTC),
+    )
+
+    payload = _serialize_item(item, active_region="asia-pacific")
+
+    assert payload["latitude"] == 1.3521
+    assert payload["longitude"] == 103.8198
+    assert payload["location_label"] == "亚太"
+    assert payload["location_inferred"] is True
+    assert payload["is_focus_match"] is True
+    assert payload["published_at"] == "2026-04-23T02:30:00Z"
+
+
+def test_serialize_item_falls_back_to_global_anchor():
+    item = ParsedNewsItem(
+        id="custom:test",
+        title="Fallback story",
+        summary="Fallback summary",
+        url="https://example.com/fallback",
+        source="Fallback Source",
+        feed_name="Fallback Feed",
+        feed_region="unknown-region",
+        homepage_url="https://example.com",
+        published_at=None,
+    )
+
+    payload = _serialize_item(item, active_region="americas")
+
+    assert payload["latitude"] == 20.0
+    assert payload["longitude"] == 0.0
+    assert payload["location_label"] == "全球"
+    assert payload["location_inferred"] is True
+    assert payload["is_focus_match"] is False
+    assert payload["published_at"] is None

backend/tests/test_logging.py

@@ -0,0 +1,78 @@
+from __future__ import annotations
+
+import logging
+
+from io import StringIO
+
+from app.core.logging import PlanetContextFilter, PlanetFormatter, get_logger
+from app.core.request_context import set_request_id
+
+
+def _capture_output(callback):
+    stream = StringIO()
+    handler = logging.StreamHandler(stream)
+    handler.setFormatter(PlanetFormatter(datefmt="%Y-%m-%d %H:%M:%S"))
+    handler.addFilter(PlanetContextFilter())
+
+    adapter = get_logger("tests.logging")
+    target_logger = adapter.logger
+    original_handlers = list(target_logger.handlers)
+    original_level = target_logger.level
+    original_propagate = target_logger.propagate
+
+    target_logger.handlers = [handler]
+    target_logger.setLevel(logging.INFO)
+    target_logger.propagate = False
+
+    try:
+        callback(adapter)
+    finally:
+        handler.flush()
+        target_logger.handlers = original_handlers
+        target_logger.setLevel(original_level)
+        target_logger.propagate = original_propagate
+
+    return stream.getvalue()
+
+
+def test_structured_logger_injects_request_id_and_event():
+    set_request_id("req-test-123")
+    try:
+        output = _capture_output(
+            lambda logger: logger.info_event(
+                "collector started",
+                event="collector.run.started",
+                context={"collector_name": "bgp_news"},
+            )
+        )
+    finally:
+        set_request_id(None)
+
+    assert "request_id=req-test-123" in output
+    assert "event=collector.run.started" in output
+    assert "service=backend" in output
+    assert '"collector_name": "bgp_news"' in output
+
+
+def test_structured_logger_redacts_sensitive_text_and_context():
+    set_request_id("req-test-redact")
+    try:
+        output = _capture_output(
+            lambda logger: logger.error_event(
+                "Authorization: Bearer super-secret-token",
+                event="auth.token.failed",
+                context={
+                    "token": "plain-secret",
+                    "nested": {"password": "hunter2"},
+                    "safe": "visible",
+                },
+            )
+        )
+    finally:
+        set_request_id(None)
+
+    assert "super-secret-token" not in output
+    assert "plain-secret" not in output
+    assert "hunter2" not in output
+    assert "[REDACTED]" in output
+    assert '"safe": "visible"' in output

backend/tests/test_system_logs.py

@@ -0,0 +1,217 @@
+from __future__ import annotations
+
+import json
+
+from pathlib import Path
+
+from app.services import system_logs
+
+
+class FakeRedis:
+    def __init__(self) -> None:
+        self.store: dict[str, list[str]] = {}
+
+    def rpush(self, key: str, value: str) -> None:
+        self.store.setdefault(key, []).append(value)
+
+    def ltrim(self, key: str, start: int, end: int) -> None:
+        items = self.store.get(key, [])
+        normalized_end = None if end == -1 else end + 1
+        self.store[key] = items[start:normalized_end]
+
+    def expire(self, key: str, seconds: int) -> None:
+        return None
+
+    def lrange(self, key: str, start: int, end: int) -> list[str]:
+        items = self.store.get(key, [])
+        normalized_end = None if end == -1 else end + 1
+        return items[start:normalized_end]
+
+    def llen(self, key: str) -> int:
+        return len(self.store.get(key, []))
+
+
+def test_read_log_snapshot_uses_structured_buffer_timestamp_level_and_search(monkeypatch):
+    fake_redis = FakeRedis()
+    monkeypatch.setattr(system_logs, "redis_client", fake_redis)
+    monkeypatch.setattr(
+        system_logs,
+        "LOG_SOURCES",
+        {
+            "earth-client": system_logs.LogSource(
+                source_id="earth-client",
+                name="Earth 浏览器端",
+                kind="buffer",
+                location="redis://planet:system_logs:earth-client",
+                description="Earth 浏览器端上报日志",
+                category="client",
+                buffer_key=system_logs.get_buffer_log_key("earth-client"),
+            )
+        },
+    )
+
+    fake_redis.rpush(
+        system_logs.get_buffer_log_key("earth-client"),
+        json.dumps(
+            {
+                "timestamp": "2026-04-22T10:15:30Z",
+                "level": "warning",
+                "message": "news feed degraded",
+                "context": {"module": "news", "detail": "timeout"},
+            },
+            ensure_ascii=False,
+        ),
+    )
+    fake_redis.rpush(
+        system_logs.get_buffer_log_key("earth-client"),
+        json.dumps(
+            {
+                "timestamp": "2026-04-23T06:01:00Z",
+                "level": "error",
+                "message": "landing points failed",
+                "context": {"module": "layer-startup", "detail": "http 500"},
+            },
+            ensure_ascii=False,
+        ),
+    )
+
+    snapshot = system_logs.read_log_snapshot(
+        "earth-client",
+        50,
+        levels="error,warning",
+        start_date="2026-04-23",
+        end_date="2026-04-23",
+        search="landing",
+    )
+
+    assert snapshot is not None
+    assert snapshot["selected_levels"] == ["error", "warning"]
+    assert snapshot["search_query"] == "landing"
+    assert snapshot["line_count"] == 1
+    assert snapshot["lines"][0].startswith("2026-04-23 06:01:00 ERROR landing points failed")
+    assert snapshot["daily_markers"] == [
+        {"date_token": "2026-04-23", "total": 1, "dominant_level": "error"}
+    ]
+
+
+def test_read_log_snapshot_parses_file_timestamp_and_builds_markers(tmp_path: Path, monkeypatch):
+    log_path = tmp_path / "backend.log"
+    log_path.write_text(
+        "\n".join(
+            [
+                "2026-04-22 08:00:00 INFO service booted",
+                "2026-04-23 09:15:00 WARNING disk pressure detected",
+                "2026-04-23 09:16:00 ERROR sync failed",
+                "2026-04-24 10:00:00 DEBUG collector trace",
+            ]
+        ),
+        encoding="utf-8",
+    )
+
+    monkeypatch.setattr(
+        system_logs,
+        "LOG_SOURCES",
+        {
+            "backend": system_logs.LogSource(
+                source_id="backend",
+                name="后端服务",
+                kind="file",
+                location=str(log_path),
+                description="测试文件日志",
+                category="service",
+            )
+        },
+    )
+
+    snapshot = system_logs.read_log_snapshot(
+        "backend",
+        50,
+        levels="warning,error",
+        search="failed",
+    )
+
+    assert snapshot is not None
+    assert snapshot["line_count"] == 1
+    assert snapshot["lines"] == ["2026-04-23 09:16:00 ERROR sync failed"]
+    assert snapshot["daily_markers"] == [
+        {"date_token": "2026-04-23", "total": 1, "dominant_level": "error"}
+    ]
+    assert snapshot["status"] == "ok"
+
+
+def test_append_buffer_log_persists_normalized_level(monkeypatch):
+    fake_redis = FakeRedis()
+    monkeypatch.setattr(system_logs, "redis_client", fake_redis)
+
+    system_logs.append_buffer_log(
+        "earth-client",
+        level="warn",
+        message="feed delayed",
+        context={"module": "news"},
+    )
+
+    stored_items = fake_redis.lrange(system_logs.get_buffer_log_key("earth-client"), 0, -1)
+    payload = json.loads(stored_items[0])
+    assert payload["level"] == "warning"
+    assert payload["message"] == "feed delayed"
+
+
+def test_infer_log_level_prefers_leading_prefix_over_query_string():
+    line = 'INFO: 127.0.0.1 - "GET /api/v1/system/logs/backend?limit=200&level=error&levels=error HTTP/1.1" 200 OK'
+
+    entry = system_logs.parse_text_log_entry(line)
+
+    assert entry.level == "info"
+
+
+def test_parse_text_log_entry_does_not_promote_exception_context_to_error():
+    line = "websockets.exceptions.ConnectionClosedError: sent 1011 (internal error) keepalive ping timeout"
+
+    entry = system_logs.parse_text_log_entry(line)
+
+    assert entry.level is None
+
+
+def test_parse_text_log_entry_still_detects_explicit_error_prefix():
+    line = "ERROR: [Errno 98] Address already in use"
+
+    entry = system_logs.parse_text_log_entry(line)
+
+    assert entry.level == "error"
+
+
+def test_read_log_snapshot_strips_nul_bytes_from_file_lines(tmp_path: Path, monkeypatch):
+    log_path = tmp_path / "backend.log"
+    log_path.write_bytes(
+        (
+            b"INFO: service booted\n"
+            b"ERROR: bind failed\n"
+            + b"\x00" * 32
+            + b"2026-04-23 23:41:32 INFO service=backend message=request served\n"
+        )
+    )
+
+    monkeypatch.setattr(
+        system_logs,
+        "LOG_SOURCES",
+        {
+            "backend": system_logs.LogSource(
+                source_id="backend",
+                name="后端服务",
+                kind="file",
+                location=str(log_path),
+                description="测试文件日志",
+                category="service",
+            )
+        },
+    )
+
+    snapshot = system_logs.read_log_snapshot("backend", 50)
+
+    assert snapshot is not None
+    assert snapshot["line_count"] == 3
+    assert snapshot["lines"] == [
+        "INFO: service booted",
+        "ERROR: bind failed",
+        "2026-04-23 23:41:32 INFO service=backend message=request served",
+    ]

backend/tests/test_visualization_compute_centers.py

@@ -0,0 +1,217 @@
+from datetime import datetime, timezone
+
+import pytest
+from httpx import ASGITransport, AsyncClient
+
+from app.api.v1.visualization import convert_compute_centers_to_geojson
+from app.db.session import get_db
+from app.main import app
+from app.models.collected_data import CollectedData
+
+
+def _build_record(
+    *,
+    record_id: int,
+    source: str,
+    data_type: str,
+    name: str,
+    country: str,
+    city: str,
+    latitude: float,
+    longitude: float,
+    metadata: dict,
+):
+    return CollectedData(
+        id=record_id,
+        source=source,
+        data_type=data_type,
+        source_id=f"{source}-{record_id}",
+        name=name,
+        extra_data={
+            "country": country,
+            "city": city,
+            "latitude": latitude,
+            "longitude": longitude,
+            **metadata,
+        },
+        collected_at=datetime(2026, 4, 22, tzinfo=timezone.utc),
+        reference_date=datetime(2026, 4, 21, tzinfo=timezone.utc),
+        is_current=True,
+    )
+
+
+def test_convert_compute_centers_to_geojson_unifies_sources():
+    top500_record = _build_record(
+        record_id=1,
+        source="top500",
+        data_type="supercomputer",
+        name="Frontier",
+        country="United States",
+        city="Oak Ridge",
+        latitude=35.93,
+        longitude=-84.31,
+        metadata={
+            "rank": 1,
+            "manufacturer": "HPE",
+            "organization": "ORNL",
+            "rmax": 1102000.0,
+            "cores": 8730112,
+            "power": 21510.0,
+        },
+    )
+    gpu_record = _build_record(
+        record_id=2,
+        source="epoch_ai_gpu",
+        data_type="gpu_cluster",
+        name="Colossus",
+        country="United States",
+        city="Memphis",
+        latitude=35.15,
+        longitude=-90.05,
+        metadata={
+            "organization": "xAI",
+            "gpu_type": "H100",
+            "gpu_count": 100000,
+            "value": "20000",
+            "unit": "TFlop/s",
+        },
+    )
+
+    payload = convert_compute_centers_to_geojson([top500_record, gpu_record])
+
+    assert payload["type"] == "FeatureCollection"
+    assert len(payload["features"]) == 2
+
+    supercomputer_feature = payload["features"][0]
+    assert supercomputer_feature["properties"]["site_type"] == "supercomputer"
+    assert supercomputer_feature["properties"]["capacity_unit"] == "GFlops"
+    assert supercomputer_feature["properties"]["capacity_band"] == "exascale"
+    assert supercomputer_feature["properties"]["operator"] == "ORNL"
+    assert supercomputer_feature["properties"]["location_precision"] == "precise"
+    assert supercomputer_feature["properties"]["is_estimated"] is False
+
+    gpu_feature = payload["features"][1]
+    assert gpu_feature["properties"]["site_type"] == "gpu_cluster"
+    assert gpu_feature["properties"]["vendor"] == "H100"
+    assert gpu_feature["properties"]["gpu_count"] == 100000
+    assert gpu_feature["properties"]["capacity_band"] == "large"
+    assert gpu_feature["properties"]["location_precision"] == "precise"
+
+
+def test_convert_compute_centers_to_geojson_uses_coordinate_hints():
+    hinted_record = _build_record(
+        record_id=3,
+        source="top500",
+        data_type="supercomputer",
+        name="Frontier",
+        country="United States",
+        city="",
+        latitude=0.0,
+        longitude=0.0,
+        metadata={
+            "organization": "Oak Ridge National Laboratory",
+            "rmax": 1102000.0,
+        },
+    )
+
+    payload = convert_compute_centers_to_geojson([hinted_record])
+
+    assert len(payload["features"]) == 1
+    coords = payload["features"][0]["geometry"]["coordinates"]
+    assert coords[0] == pytest.approx(-84.3107)
+    assert coords[1] == pytest.approx(35.9319)
+    assert payload["features"][0]["properties"]["is_estimated"] is True
+    assert payload["features"][0]["properties"]["location_precision"] == "estimated_site"
+
+
+def test_convert_compute_centers_to_geojson_falls_back_to_country_centroid():
+    centroid_record = _build_record(
+        record_id=4,
+        source="epoch_ai_gpu",
+        data_type="gpu_cluster",
+        name="Unknown Cluster",
+        country="United States",
+        city="",
+        latitude=0.0,
+        longitude=0.0,
+        metadata={
+            "organization": "Unknown Operator",
+            "value": "10000",
+            "unit": "TFlop/s",
+        },
+    )
+
+    payload = convert_compute_centers_to_geojson([centroid_record])
+
+    assert len(payload["features"]) == 1
+    props = payload["features"][0]["properties"]
+    coords = payload["features"][0]["geometry"]["coordinates"]
+    assert coords[0] == pytest.approx(-98.5795)
+    assert coords[1] == pytest.approx(39.8283)
+    assert props["is_estimated"] is True
+    assert props["location_precision"] == "estimated_country"
+    assert props["geography_mode"] == "country_centroid"
+
+
+@pytest.mark.asyncio
+async def test_compute_centers_geojson_endpoint_returns_stats():
+    records = [
+        _build_record(
+            record_id=1,
+            source="top500",
+            data_type="supercomputer",
+            name="Frontier",
+            country="United States",
+            city="Oak Ridge",
+            latitude=35.93,
+            longitude=-84.31,
+            metadata={"rank": 1, "rmax": 1102000.0},
+        ),
+        _build_record(
+            record_id=2,
+            source="epoch_ai_gpu",
+            data_type="gpu_cluster",
+            name="Colossus",
+            country="United States",
+            city="Memphis",
+            latitude=35.15,
+            longitude=-90.05,
+            metadata={"value": "20000", "unit": "TFlop/s"},
+        ),
+    ]
+
+    class _ScalarResult:
+        def __init__(self, rows):
+            self._rows = rows
+
+        def scalars(self):
+            class _Scalars:
+                def __init__(self, rows):
+                    self._rows = rows
+
+                def all(self):
+                    return self._rows
+
+            return _Scalars(self._rows)
+
+    class _FakeSession:
+        async def execute(self, _query):
+            return _ScalarResult(records)
+
+    async def override_get_db():
+        yield _FakeSession()
+
+    app.dependency_overrides[get_db] = override_get_db
+    transport = ASGITransport(app=app)
+    try:
+        async with AsyncClient(transport=transport, base_url="http://test") as client:
+            response = await client.get("/api/v1/visualization/geo/compute-centers")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["count"] == 2
+        assert data["stats"]["supercomputers"] == 1
+        assert data["stats"]["gpu_clusters"] == 1
+        assert data["features"][0]["properties"]["data_type"] == "compute_center"
+    finally:
+        app.dependency_overrides.clear()

docs/CHANGELOG.md

@@ -5,8 +5,505 @@ All notable changes to `planet` are documented here.
 This project follows the repository versioning rule:
 
 - `feature` -> `+0.1.0`
+- `improvement` -> `+0.0.1`（bugfix + 小功能混合）
 - `bugfix` -> `+0.0.1`
 
+## [0.39.0] — 2026-04-24
+
+## [0.40.1] — 2026-04-24
+
+### 🔧 Improvements
+- 卫星选中标记（lockedring / lockeddot / 光晕）颜色统一跟随图例轨道倾角分类配色
+- 修复 Starlink footprint 在特定视角下遮蔽卫星点的渲染顺序问题（Group renderOrder 影响子 Mesh 排序）
+- footprint 材质改为 `depthTest: false` + 相机朝向 limbFade，替代 polygonOffset 深度竞争方案
+- 修复选中海缆时误触发附近卫星高亮（该行为属于 BGP 事件点逻辑，不应用于海缆）
+
+---
+
+## [0.40.0] — 2026-04-24
+
+### ✨ Highlights
+- Earth 卫星 footprint 正式按星座能力分层：Starlink 保留专用地表覆盖，Iridium 改为独立外圈覆盖表达，其它非 Starlink 星座不再误用同一套 footprint
+- Earth 卫星详情卡补齐覆盖能力与当前显示说明，用户现在可以直接看见每颗卫星为什么显示 footprint、为何回退为自身发光
+
+### 🔧 Improvements
+- 后端可视化接口新增并透传 `constellation_group` 与 `footprint_policy`，前端据此执行 capability-gated footprint renderer
+- 新增 Iridium 独立 coverage ring adapter，并继续保留 Starlink 专用 footprint 调校与昼夜可读性增强
+- 新增 Earth 卫星 footprint 策略技术文档，明确 GNSS、generic LEO、GEO 与 Iridium 的显示边界
+
+### 🐛 Fixes
+- 修复前后端对 Iridium footprint policy 命名不一致，导致策略分发语义含混的问题
+- 清理 Starlink footprint 渲染中的未使用常量与过时命名，减少后续继续调校时的歧义
+
+---
+
+## [0.39.0] — 2026-04-24
+
+### ✨ Highlights
+- 后端正式落下统一结构化日志地基：请求上下文、事件名、脱敏与持久化链路开始收口为可扩展的企业级日志体系
+- 系统日志页重构为真正的日志工作台：顶部筛选更紧凑，终端日志区成为主视觉，移动端 Earth 新闻/态势细节交互继续补稳
+
+### 🔧 Improvements
+- 新增 `backend/app/core/logging.py`，统一 `request_id`、`service`、`event` 注入与敏感字段脱敏，并接入后端主入口、调度器、缓存、数据库和可视化链路
+- 系统日志页筛选区重排为更紧凑的两层结构，信息摘要并入终端工具栏 tooltip，日志终端区留出更稳定的按钮避让空间
+- Earth 移动端态势抽屉补齐宽度约束与图例换行规则，新闻详情抽屉在巡航切换时可同步更新标题和摘要
+
+### 🐛 Fixes
+- 修复 `/tmp/planet_backend.log` 中混入空字节时，日志摘要条行数与实际可见日志不一致的问题
+- 修复移动端“态势”tab 在内容渲染后被图例文本撑宽、超出一屏的问题
+- 修复移动端新闻详情抽屉在巡航切换下一条新闻时标题更新但 summary 不同步的问题
+
+---
+
+## [0.38.0] — 2026-04-23
+
+### ✨ Highlights
+- Earth 新闻正式接入通用巡航层：新闻和 BGP 统一进入可配置巡航模块，桌面端与移动端都能在巡航聚焦时展示对应新闻卡片
+- 系统日志页升级为结构化过滤链路：按真实时间戳、结构化级别和字符串检索统一筛选，不再依赖前端或后端从日志文本里猜结果
+
+### 🔧 Improvements
+- 新闻巡航补齐业务适配层：按发生地与时间生成巡航目标，桌面端与移动端统一标题 + summary 卡片风格，并增加连线与打字机摘要展示
+- 日志页筛选体验重排，统一服务源、级别、行数、时间和检索布局，日历标记改为由后端返回的结构化每日聚合结果驱动
+- 后端补充 `system_logs` 结构化解析与多级别精确过滤能力，Earth 浏览器端日志缓冲与系统日志 API 现在走同一套筛选语义
+
+### 🐛 Fixes
+- 修复新闻巡航模块开启后难以关闭、桌面/移动端设置状态互相污染的问题
+- 修复新闻巡航卡片缺少摘要、移动端详情样式不统一、新闻巡航缺少连线的问题
+- 修复日志级别筛选会被访问日志 query string 中的 `level=error` 等参数污染，从而把 `INFO` 行误判为 `ERROR` 的问题
+
+---
+
+## [0.37.2] — 2026-04-23
+
+### ✨ Highlights
+- Earth 图层系统新增经纬线开关，桌面图层面板与移动端抽屉都可直接控制
+
+### 🔧 Improvements
+- 经纬线正式接入 Earth layer registry，复用现有图层切换、移动端图层卡片与设置持久化流
+
+### 🐛 Fixes
+- 修复经纬线只能默认常驻、无法作为独立图层开关控制的问题
+
+---
+
+## [0.37.1] — 2026-04-23
+
+### ✨ Highlights
+- `planet.sh` 后端重启链路修复 `uvicorn --reload` 残留 worker 场景，`restart` 现在能真正替换旧实例
+
+### 🔧 Improvements
+- 收口后端清理逻辑，统一按 `uvicorn` 进程、端口占用进程和进程组执行清理，减少 reload 场景漏杀分支
+
+### 🐛 Fixes
+- 修复部分机器执行 `./planet.sh restart --allow-lan` 后后端仍停留旧实例，导致 `/api/v1/visualization/geo/compute-centers` 返回 `404` 的问题
+
+---
+
+## [0.35.1] — 2026-04-22
+## [0.37.0] — 2026-04-23
+
+### ✨ Highlights
+- Earth 连线系统正式从巡航里解耦成通用 callout connector：桌面端和移动端统一支持对象级锚点、四边切换与临界区边缘滑动
+- BGP 巡航展示继续收口为稳定的“先定位卡片、再连真实锚点、再展示卡片”链路，移动端 popup 与桌面 info panel 的路线规则统一
+
+### 🔧 Improvements
+- connector 配置从 `CRUISE_CONFIG` 拆到独立 `CONNECTOR_CONFIG`，默认类名、动画名和实例命名也全部去 cruise 语义
+- 移动端 popup 增加更稳定的 dock/obstacle 处理，拖动卡片时连线起终点会持续按几何关系自适应刷新
+- Earth 多个图层与控制逻辑继续收口，补充算力中心/BGP 风格对齐、layer panel 与相关交互细节调整
+
+### 🐛 Fixes
+- 修复巡航模式下终点只像“视觉锚点”而不是真实绑定对象的问题，卡片拖动后终点现在会跟随
+- 修复移动端与桌面端多类连线路线异常：压线、反向、临界区折返、起点遮挡事件点等问题
+- 修复对象矩形临界区内连线仍强制中点到中点导致路线像“先钻进 source 内部”再出去的问题
+
+---
+
+## [0.35.1] — 2026-04-22
+## [0.36.0] — 2026-04-22
+
+### ✨ Highlights
+- Earth 新增统一“算力中心”图层：接入超算与 GPU 集群，支持搜索、统计、图例、详情卡与独立图层开关
+- 算力中心支持精确位置与估算位置两种状态，估算点会以问号角标区分，避免数据不全时整批节点在地图上消失
+
+### 🔧 Improvements
+- Earth 详情卡拖拽与地球拖拽交互继续收口，减少拖动卡片和旋转地球时的选中文本与 pointer 竞争
+- `planet.sh` 改为通过独立脚本计算 AI Provider 依赖指纹，降低与根仓库依赖版本文件的无关耦合
+- README 补充 WSL / Windows 局域网访问排查与转发配置说明，便于开发环境联调
+
+### 🐛 Fixes
+- 修复 Earth 算力中心图层在无原始坐标时无法显示的问题，支持站点提示和国家级估算回退
+- 修复信息卡拖拽事件可能被卡片级 stopPropagation 吞掉，导致拖拽流中断的问题
+
+---
+
+## [0.35.1] — 2026-04-22
+
+### ✨ Highlights
+- Earth 统计展示改为统一 `data-earth-stat` 绑定机制，桌面 HUD 和移动端抽屉复用同一套状态更新入口
+
+### 🔧 Improvements
+- 收口海缆、登陆点、卫星、BGP 事件与 BGP 状态的统计写入逻辑，减少后续继续补桌面/移动双写分支的成本
+
+### 🐛 Fixes
+- 修复移动端态势抽屉中的海缆、登陆点与 BGP 统计在图层切换后可能停留旧值的问题
+
+---
+
+## [0.35.0] — 2026-04-22
+
+### ✨ Highlights
+- Earth 移动端底部抽屉系统全面上线：响应式布局自动切换、Tab 导航、手势上拉/下滑开合、惯性速度判定
+- 移动端点击可交互物件（海缆、登陆点、卫星、BGP）后弹出智能定位悬浮卡片，可拖动，点击跳转详情
+
+### 🔧 Improvements
+- 抽屉把手区域缩小至 36px（collapsed 时仅露出把手，不遮挡地球操作区）
+- 抽屉定期弹跳动画提示用户可上拉，5 秒间隔，打开后自动停止
+- 通知胶囊位置调整，不再覆盖品牌 logo
+- 移动端单指旋转、双指捏合缩放地球，触控事件冲突修复（pointer-events 级联）
+
+### 🐛 Fixes
+- 修复移动端抽屉 shell 因 layout 高度（240px+）遮挡地球触控区域，pointer-events 改为按层级精确控制
+- 修复悬浮卡片因 setPointerCapture 在 iOS Safari 抑制合成 click 事件导致无法点击的问题
+
+---
+
+## [0.34.0] — 2026-04-22
+
+### ✨ Highlights
+- Earth 搜索面板正式接入，支持搜索海缆、登陆点、卫星、BGP 事件与观测站，并可直接聚焦到对应对象
+- `planet.sh --allow-lan` 打通 Bun + Vite 的局域网开放链路，启动成功后自动打印推荐访问地址与后端健康检查地址
+
+### 🔧 Improvements
+- 前端开发启动链统一改成 Bun 直接执行 Vite 入口，不再依赖 shell 中额外暴露的 Node 路径
+- Earth 搜索结果接入登陆点详情卡片与对象聚焦，搜索后可直接进入对应详情流
+- `planet.sh` 补充局域网 IPv4 自动识别与推荐地址输出，减少 WSL 局域网调试成本
+
+### 🐛 Fixes
+- 修复 `./planet.sh restart --allow-lan` 全量重启时未把 `--allow-lan` 继续传给 `start()`，导致前端退回本机监听的问题
+- 修复 WSL + Bun 环境下前端偶发因 Vite 启动链不稳定而无法正确监听 `0.0.0.0:3000` 的问题
+
+---
+
+## [0.33.0] — 2026-04-22
+
+### ✨ Highlights
+- `news_live_streams` 采集器默认接入 `iptv-org` 频道目录，并将采集结果稳定并入 Earth TV 直播源列表
+- 数据源页支持直接编辑内置数据源 override，并为内置源提供一键恢复默认配置入口
+
+### 🔧 Improvements
+- `News Live Streams` 现在作为可直接触发的内置默认数据源提供，无需先手工补 override 才能采集
+- TV 播放源菜单会直接区分 `[内置]` 和 `[采集]` 来源，频道来源信息也会同步展示
+- 新增 [earth-news-source-configuration-and-collector-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-news-source-configuration-and-collector-plan.md)，正式规划 Earth 态势新闻源配置化与后续采集器化路线
+
+### 🐛 Fixes
+- 修复 `news_live_streams` 采集完成后 `/api/v1/tv/streams` 因读取不存在的 `updated_at` 字段而导致默认频道全部消失的问题
+- 修复内置数据源操作列按钮显示不全，以及编辑抽屉中多个 `Collapse` 紧贴的问题
+
+---
+
+## [0.32.0] — 2026-04-22
+
+### ✨ Highlights
+- Earth 设置新增“地球默认大小”持久化项，重置视角、缩放百分比重置和 BGP 巡航视图现在统一复用这一份默认 zoom
+- 卫星焦点层次继续收口：巡航进入 presentation 前不再过早 dim，非焦点卫星改成“降亮度/尾迹/背板”而不是去饱和度
+
+### 🔧 Improvements
+- Earth 设置面板区块和左右留白进一步收紧，整体更贴近 HUD 面板的密度
+- toolbar 展开边界缓存改为按需刷新，减少 document 级 mousemove 期间的重复布局读取
+- Scrollbar 和 ScrollbarOverlay 收窄 observer 范围，减少大表格和动态菜单下的额外刷新成本
+- 更新 [earth-frontend-context.md](/home/ray/dev/linkong/planet/docs/technical/earth-frontend-context.md)，补充默认视图大小已进入 Earth 设置持久化真源
+
+### 🐛 Fixes
+- 修复开启巡航后，尚未进入连线/presentation 时卫星已经整体变暗的问题
+- 修复默认大小重置链路分散在多个入口、实际 reset/cruise/缩放提示不一致的问题
+- 修复开启地形后卫星反馈层与地球背面可见性之间的一组表现问题，保留正面反馈同时恢复背面轨道遮挡
+
+---
+
+## [0.31.3] — 2026-04-22
+
+### ✨ Highlights
+- Earth 图层注册表和启动任务框架继续收口，启动顺序、启动模式、启动提示和任务注册现在都能从统一入口扩展
+- 修复 Earth 普通旋转模式与巡航模式切换时的一组交互回归，同时让卫星/地形/昼夜模式的表现更稳定
+
+### 🔧 Improvements
+- 新增 [layer-startup-tasks.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/layer-startup-tasks.js) 启动任务注册表，支持 `registerLayerStartupTask(id, taskFactory)`，并拆成海缆 / 卫星 / BGP 独立注册函数
+- Earth 图层控制改成注册表驱动，统一承载 `startupPriority`、`startupMode`、`startupLabel`、`startupMessage` 与图层持久化元信息
+- Earth 设置支持持久化图层开关、旋转模式、HUD 面板显示状态、地形透明度与日夜模式，并提供一键重置
+- 更新 [earth-frontend-context.md](/home/ray/dev/linkong/planet/docs/technical/earth-frontend-context.md) 记录图层注册表、启动任务、设置持久化与巡航适配边界
+
+### 🐛 Fixes
+- 修复普通旋转模式下点击海缆 / 卫星 / BGP 后卡片和选中表现会被异常清空的问题
+- 修复巡航模式切回旋转再切回巡航后无法继续自动巡航的问题
+- 修复开启地形后卫星选中反馈层被高海拔区域吞掉的问题，并恢复轨道只在地球前半侧可见
+- 修复关闭日夜模式后地球照明仍沿真实昼夜切换、亮部过曝和偏色的问题，改成更中性的 inspection lighting
+- 修复 toolbar 收起态仍挡住地球交互，以及首帧短暂展开闪现的问题
+
+---
+
+## [0.31.2] — 2026-04-21
+
+### ✨ Highlights
+- Earth 巡航模式重构为“通用巡航队列 + 通用连线动画 + BGP 业务适配”三层结构，后续扩到海缆、卫星或新闻巡航时不必再复制一套 `main.js` 状态机
+- 修复巡航重构后的交互回归：空白点击重新稳定切到下一项，连线按“起点 → 引导线 → 终点”顺序入场
+
+### 🔧 Improvements
+- 新增 [cruise-sequencer.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/cruise-sequencer.js) 统一管理队列推进、停留时长、打断与恢复
+- 新增 [callout-connector.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/callout-connector.js) 统一管理 SVG 连线、折线路径与描边动画
+- 新增 [bgp-cruise-adapter.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/bgp-cruise-adapter.js) 收口 BGP 巡航目标排序、卡片落点、轮询去重与连线适配
+- 更新 [earth-frontend-context.md](/home/ray/dev/linkong/planet/docs/technical/earth-frontend-context.md) 说明新的巡航分层与复用边界
+
+### 🐛 Fixes
+- 修复巡航模式下点击空白处无法稳定跳转到下一项、切回旋转再切回巡航后直接卡住的问题
+- 修复巡航连线被实时重定位覆盖导致“直接出现”而非绘制动画的问题
+- 修复连线动画节点入场节奏不对的问题，改为先出现起点，再绘制连线，最后出现终点
+
+---
+
+## [0.31.1] — 2026-04-21
+
+### ✨ Highlights
+- Earth 图层开关状态统一成可复用的 `active / loading` 状态机，首次启用地形和卫星时不再像按钮失效
+- 文档目录重构为 `docs/technical`、`docs/plans`、`docs/deprecated`，并吸收 `.sisyphus/plans` 中有价值的 Earth / 卫星 / UE5 草案
+
+### 🔧 Improvements
+- 新增 [layer-button-state.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/layer-button-state.js)，统一按钮 tooltip、`aria-busy`、禁用态和状态文本同步
+- 地形图层支持 hover/focus 预热与空闲预热，首次点击等待前移，加载中状态持续可见
+- 卫星图层启用前会立即切换为 `loading` 中间态，请求完成后再切回正常开关表现
+
+### 🐛 Fixes
+- 修复地形首次加载时通知过早消失、开关仍像关闭状态导致用户误判按钮损坏的问题
+- 修复卫星接口较慢时按钮没有任何中间态反馈的问题
+
+---
+
+## [0.31.0] — 2026-04-21
+
+### ✨ Features
+- Earth 新增"巡航展示"模式：自动轮播 BGP 异常事件，逐帧追踪连接线位置，支持外部交互立即中断序列（cancel notifier 模式）
+- 巡航目标事件点高亮显示：hover 外观 + 锁定脉冲动画，并与点击行为统一展示周边受影响卫星与海缆
+- BGP 事件图标新增填充 W 形波动符号（flap 类型），替换原有难以辨认的贝塞尔细线
+- 巡航/点击激活时其余卫星自动降饱和度 + 增加透明度以突出焦点；海缆未受影响时同步变暗
+
+### 🔧 Improvements
+- 修复巡航轮播期间 BGP 事件 polling 刷新导致标记闪烁消失的问题（clearBGPData 延迟到请求完成后执行）
+- 点击与巡航锁定颜色统一为 hover 色（0.92, 0.98, 1.0 全透明），移除锁定态脉冲动画
+- 巡航连接折线转折点从尖角调整为钝角（linkElbowDropPx），提升连线可读性
+
+---
+
+## [0.30.0] — 2026-04-21
+
+### ✨ Features
+- Earth 新增真实地形图层：后端代理 Terrarium DEM 瓦片（`/api/v1/visualization/terrain/terrarium/{z}/{x}/{y}.png`），前端新增 `terrain.js` 负责瓦片拉取、顶点位移与按海拔着色
+- 设置弹窗新增"地形"分组，支持通过滑块实时调整地形图层透明度
+
+### 🔧 Improvements
+- 地形按钮改为异步加载，首次点击显示进度提示并在失败时自动回退
+- 启动阶段改用 `applyImmediateView` 直接应用初始视角，`showStatusMessage` / `queueStatusMessage` 区分即时与队列态状态消息，加载中不再被临时状态打断
+- 控制面板抽取 `applyTerrainUiState` / `getViewRotation` 收敛地形切换与视角旋转的重复 UI 同步逻辑
+
+---
+
+## [0.29.2] — 2026-04-21
+
+### ✨ Highlights
+- Earth 继续收口 HUD 交互与设置面板表现，设置弹窗改成更接近从按钮展开的窗口感，同时加入系统级 admin 入口
+- 修正天球太阳方向与地球受光解耦后的日照逻辑，地表昼夜判断改为按太阳直射点经纬度落到地球贴图坐标
+
+### 🔧 Improvements
+- toolbar 进一步收成更贴近 hub 的浅弓形排列，并统一成与 HUD panel 一致的液态玻璃配色与透明度
+- 设置弹窗与各 HUD panel 继续统一样式、等比缩放和头部基线，设置列表补充系统分组与 admin 跳转
+- 所有 HUD panel 增加更统一的液态玻璃高光与 hover / press 反馈
+
+### 🐛 Fixes
+- 修复设置弹窗仍像旧圆角矩形、标题文案重复和从底边直直飞出的动画问题
+- 修复天球与太阳方向混用显示校准导致中国白天仍落在夜面的日照错误
+
+---
+
+## [0.29.1] — 2026-04-20
+
+### ✨ Highlights
+- Earth 加载状态条改成单一队列式通知面板，加载阶段不再因为步骤文案变化而回缩，也不会被其他通知打断
+- 调整 brand panel 的呈现方式与昼夜/选中态可读性，让品牌区更自然、交互高亮在白天和黑夜里都更稳定
+
+### 🔧 Improvements
+- 移除旧的地球加载浮层结构，统一由 HUD 状态消息承载三点脉冲加载过程
+- brand panel 改为无边框品牌层，仅保留轻微氛围光，不再因为非常规尺寸显得像第五块功能面板
+- 温和收敛地球昼夜材质与主背光强度，保留昼夜辨识度的同时提升白天地表纹理和夜面交互可见性
+
+### 🐛 Fixes
+- 修复加载地球时通知条在步骤切换中反复缩短、其他状态消息抢占加载流程的问题
+- 修复海缆、登陆点和 BGP 选中高亮在黑夜中过暗、在高光中过亮导致难以辨识的问题
+
+---
+
+## [0.29.0] — 2026-04-20
+
+### ✨ Highlights
+- Earth 新增天球层第一版：引入真实全天星图、亮星层与太阳/月亮位置计算，地球场景首次具备可校准的天文背景
+- 地球昼夜分隔升级为更明显的日夜增强效果，夜面、晨昏带和太阳方向联动更容易直接读出来
+
+### 🔧 Improvements
+- 新增 `celestial.js` 模块和 `assets/celestial/` 资源目录，统一管理星图、亮星数据以及太阳/月亮与光照同步
+- 卫星图例改为按倾角分组，严格固定为“赤道轨道 → 低倾角轨道 → 中倾角轨道 → 高倾角轨道 → 逆行轨道”顺序，并全部中文化
+- 图层面板补齐关闭按钮，拖拽脱离左列后不再被流布局 margin 影响，能够真正贴到品牌面板下沿
+
+### 🐛 Fixes
+- 修复天球球壳放大后被相机 far plane 裁剪导致的外层黑环问题
+- 修复图层面板在左侧上移时始终与 brand panel 保持额外间距的问题
+
+---
+
+## [0.28.2] — 2026-04-20
+
+### ✨ Highlights
+- 修正媒体情报面板在 `电视直播 / 态势聚合` tab 间切换时的尺寸记忆逻辑，切回原 tab 后可恢复各自大小状态
+- 清理 `docs/` 根目录遗留的旧路径文档，只保留新的分组目录和归档目录，结束同一文档双路径并存状态
+
+### 🔧 Improvements
+- `media-panel` 切换逻辑改成按 tab 分别记忆尺寸状态，避免 `A -> B -> A` 时继续共用同一套外层尺寸
+- 目录整理真正完成收尾：旧的 `docs/*.md` 平铺计划文档删除，继续以 `docs/agents / earth / backend / frontend / ops / ue5 / deprecated` 为唯一入口
+
+### 🐛 Fixes
+- 修复拉伸 `media-panel` 后切换 tab 时，`news-panel` 高度回退到旧默认值的问题
+- 修复拉伸后切换 tab 导致面板视觉锚点异常的问题，切换时改为围绕当前卡片自身右下角进行尺寸恢复
+
+---
+
+## [0.28.1] — 2026-04-20
+
+### ✨ Highlights
+- 收口 Earth 媒体情报面板命名，明确外层 `media-panel` 与内部 `tv-panel / news-panel` 的职责边界
+- 整理 `docs/` 目录分组，并将已完成或已废弃的计划文档归档到 `docs/deprecated`
+
+### 🔧 Improvements
+- 底部 tab 语义统一为 `media-panel-tabs / media-panel-tab`，并将文案更新为“电视直播 / 态势聚合”
+- 补充媒体面板、聚合新闻模块的注释说明，减少 `tv-panel` 同时指代外层壳和内层直播 pane 的阅读歧义
+- 更新 README、AI Provider README 与历史文档互链，适配新的 `docs/agents / docs/earth / docs/frontend / docs/backend / docs/ops / docs/ue5` 分组结构
+
+### 🐛 Fixes
+- 修复媒体情报面板标题组在头部撑出多余空白的问题，去掉 `hud-panel__title-group` 的无效弹性占位
+- 修复聚合 tab 头部仍保留冗余固定标题的问题，现在仅显示区域标签
+
+---
+
+## [0.28.0] — 2026-04-20
+
+### ✨ Features
+- 将 Earth 的“新闻直播”和“全球态势聚合”合并为统一的“媒体情报”面板，支持底部 tab 切换与共享标题栏操作区
+- 聚合新闻不再单独占据一个 HUD 面板，而是作为媒体情报面板内的第二视图与直播协同呈现
+
+说明：
+- 当前结构中，外层 HUD 壳为 `media-panel`，内部 tab 内容区分别为 `tv-panel` 和 `news-panel`
+
+### 🔧 Improvements
+- TV / News 面板切换加入底边锚定的 reform 动画，并继续保留拖拽、缩放和共享 HUD 行为收口
+- 聚合新闻视图新增默认高度约束与内部滚动填充逻辑，避免初始高度过度膨胀
+- `tv.js`、`news.js` 进一步清理共享 HUDPanel 迁移后的残留逻辑，收紧 tab / resize / reform 相关局部 helper
+
+### 🐛 Fixes
+- 修复媒体情报面板右下角缩放时高度异常抬升、标题栏被顶出视口的问题
+- 修复直播/聚合 tab 切换时按钮高亮、内容切换和底边基准表现不一致的问题
+
+---
+
+## [0.27.7] — 2026-04-16
+
+## [0.27.8] — 2026-04-20
+
+### 🔧 Improvements
+- Earth HUD 共享 `HUDPanel` 默认展开/收缩逻辑继续收口，图例与图层面板统一使用同一套边缘阈值与箭头状态机
+- 保持新闻直播面板现有特例折叠行为不变，避免播放器区域被默认折叠逻辑影响
+
+### 🐛 Fixes
+- 修复图例与图层面板展开/收缩箭头方向和实际动作不一致的问题
+- 修复拖动到屏幕底边附近时初始箭头、拖动中箭头和点击后动作不同步的问题
+
+---
+
+## [0.27.7] — 2026-04-16
+
+### 🔧 Improvements
+- 用户管理、数据源配置、电视直播源表格统一接入可折叠操作列，窄宽度下自动收起到下拉菜单，减少操作区挤压
+- 电视直播设置改为表格总览 + 弹窗编辑模式，主表内容更紧凑，适合控制台一屏浏览
+- Earth TV 面板新增失败源探测与自动回退恢复标记，便于值班时快速识别异常直播源
+
+### 🐛 Fixes
+- 修复 Settings 电视直播源新增后取消编辑会残留未保存草稿的问题
+- 修复 Settings 删除直播源只改本地状态、刷新后恢复的问题，删除现在会立即持久化
+- 修复 Users / DataSources / Settings 表格“备注/状态”和“操作”之间的空白占位列问题
+- 修复 `useCollapsedActions` 未释放 `ResizeObserver` 导致的潜在内存泄漏与重复回调问题
+
+---
+
+## [0.27.6] — 2026-04-15
+
+### 🔧 Improvements
+- BGP 告警页表格纵向 overflow 修复：补全 flex 布局链，tabs content-holder 正确撑满剩余高度
+- 用户管理表格横向滚动修复：采用 flex-fill 方案替换 `height: auto !important`，自定义滚动条 X 轨道位置对齐表格底部
+- Playground 宽布局隐藏"服务状态"按钮：侧边栏可见时不显示冗余入口
+- AI Chatbox 输入框失焦收起为单行，聚焦或有内容时展开完整 composer
+
+---
+
+## [0.27.4] — 2026-04-14
+
+### 🔧 Improvements
+- info-card 改为懒加载动态挂载：页面初始 DOM 不再含隐藏的 `#info-panel` 节点，仅首次点击交互元素时创建
+
+---
+
+## [0.27.5] — 2026-04-14
+
+### 🔧 Improvements
+- 统一控制台多页面滚动体验：BGP、alerts、采集数据、用户管理、任务、设置、Playground 等区域接入自定义滚动条与表格滚动容器
+- 优化 BGP 与 alerts 页响应式布局：顶部概览卡在窄宽度下优先重排，必要时才启用横向滚动，避免卡片裁切和全局滚动条接管
+- 调整 `situational alerts` 布局策略：统计卡按宽度在单行、两列和横滚之间切换，下方详情卡保持单行高度优先
+- 实时采集进度优化：一键采集完成后在未刷新页面时保留 100% 完成态，不再错误归零
+- 补充 UE5 MVP 融合方案文档，完善后续集成规划沉淀
+
+### 🐛 Fixes
+- 修复 BGP summary 与 alerts 顶部卡片在无真实溢出时误出现横向滚动的问题
+- 修复 alerts 页面缩窄后外层全局竖向滚动被接管的问题，恢复“一屏内、内部滚动”的布局逻辑
+- 修复 `situational alerts` 在两排布局下详情卡竖向溢出的问题，改为更稳定的分区响应式排版
+- 修复自定义滚动条交互反馈，悬停、聚焦、拖拽时颜色加深但不再显示多余外圈
+
+---
+
+## [0.27.3] — 2026-04-14
+
+### 🔧 Improvements
+- TV panel meta 折叠展开方向稳定：底部锚定时向上生长，拖拽后（顶部锚定）通过 JS 补偿 top 保持播放器底部位置不变
+- 修复 TV panel 展开/折叠时视频区域跳动问题：移除面板 min-height，使播放器高度在两种状态下保持一致
+- 修正 TV panel meta toggle 箭头方向：展开朝下，折叠朝上
+- 修复图例面板折叠按钮失效（legend-bar-btn 补充进拖拽排除列表）
+- 调整图层搜索框图标尺寸为 20px，BR 缩放角标改为直角 L 形
+
+---
+
+## [0.27.2] — 2026-04-14
+
+### 🔧 Improvements
+- 修复 brand copy 宽度不随内容收缩的问题，现在与 title 图片宽度保持一致
+- 提取 `--brand-copy-width` CSS 自定义属性，消除 160px / 172px 魔法数字重复
+
+---
+
+## [0.27.1] — 2026-04-14
+
+### 🔧 Improvements
+- 面板拖拽新增 L 形边界约束，其他面板无法覆盖 brand 面板区域，并从右侧/底部自然卡边
+- brand 组件引入 `--brand-scale` 整体缩放变量，padding 与内容尺寸独立控制
+- 图层控制面板宽度收窄（260px），与 brand 面板错落排列，间距调大
+
+### 🐛 Fixes
+- 修复搜索框 `type="search"` 导致清除按钮重复显示的问题
+- 修复 `[hidden]` 属性被组件 `display` 规则覆盖的问题
+
+---
+
 ## 0.27.0
 
 Released: 2026-04-14
@@ -80,7 +577,7 @@ Released: 2026-04-12
 
 - 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.
+- Added [docs/deprecated/earth-tv-live-module-plan.md](/home/ray/dev/linkong/planet/docs/deprecated/earth-tv-live-module-plan.md) and [docs/earth/technical/news-live-streams-collector-format.md](/home/ray/dev/linkong/planet/docs/technical/earth-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
 
@@ -166,7 +663,7 @@ Released: 2026-04-10
 
 - 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.
+- Improved [README.md](/home/ray/dev/linkong/planet/README.md) and [docs/agents/situational-awareness-foundation-plan.md](/home/ray/dev/linkong/planet/docs/plans/agents-situational-awareness-foundation-plan.md) by documenting the current AI/alerts base, planned situational-awareness direction, and the new persistent Playground foundation.
 
 ### Fixed
 
@@ -204,7 +701,7 @@ Released: 2026-04-10
 ### 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.”
+- Improved [frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/technical/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
 
@@ -221,7 +718,7 @@ Released: 2026-04-10
 - 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.
+- Improved [docs/frontend/plans/ai-playground-development-plan.md](/home/ray/dev/linkong/planet/docs/plans/frontend-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
 
@@ -327,8 +824,8 @@ Released: 2026-04-09
 ### 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.
+- Added [docs/frontend/technical/frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/technical/frontend-layout-guidelines.md), documenting the repository standard for one-screen admin workspaces and module-local overflow handling.
+- Added [docs/frontend/plans/ai-playground-development-plan.md](/home/ray/dev/linkong/planet/docs/plans/frontend-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
 
@@ -433,7 +930,7 @@ Released: 2026-04-07
 - Added [backend/app/services/ai_client.py](/home/ray/dev/linkong/planet/backend/app/services/ai_client.py), introducing an internal HTTP client for `backend -> aiprovider` calls with request-id propagation and lightweight retry.
 - Added [aiprovider/main.py](/home/ray/dev/linkong/planet/aiprovider/main.py), [aiprovider/provider_service.py](/home/ray/dev/linkong/planet/aiprovider/provider_service.py), and related config/schema files to stand up the dedicated adapter service.
 - Added [aiprovider/.env.example](/home/ray/dev/linkong/planet/aiprovider/.env.example) and [docker-compose.local-model.yml](/home/ray/dev/linkong/planet/docker-compose.local-model.yml) as ready-to-edit local-model templates.
-- Added [docs/aiprovider.md](/home/ray/dev/linkong/planet/docs/aiprovider.md), documenting architecture, configuration, single-machine and multi-machine deployment, and cross-service calling patterns.
+- Added [docs/agents/aiprovider.md](/home/ray/dev/linkong/planet/docs/technical/agents-aiprovider.md), documenting architecture, configuration, single-machine and multi-machine deployment, and cross-service calling patterns.
 - Added a dedicated `重启 AI Provider` control path in [Dashboard.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Dashboard/Dashboard.tsx), [system_control.py](/home/ray/dev/linkong/planet/backend/app/services/system_control.py), and [system_restart_runner.py](/home/ray/dev/linkong/planet/backend/scripts/system_restart_runner.py).
 
 ### Improved
@@ -559,7 +1056,7 @@ Released: 2026-04-02
 
 - Added a new `IPtoASN Prefix Geography` collector in [iptoasn.py](/home/ray/dev/linkong/planet/backend/app/services/collectors/iptoasn.py) and registered it through [data_sources.yaml](/home/ray/dev/linkong/planet/backend/app/core/data_sources.yaml), [data_sources.py](/home/ray/dev/linkong/planet/backend/app/core/data_sources.py), [datasource_defaults.py](/home/ray/dev/linkong/planet/backend/app/core/datasource_defaults.py), and [collectors/__init__.py](/home/ray/dev/linkong/planet/backend/app/services/collectors/__init__.py).
 - Added country centroid helpers in [countries.py](/home/ray/dev/linkong/planet/backend/app/core/countries.py) so country-level prefix geography can produce map coordinates instead of only labels.
-- Added a dedicated prefix-geography implementation note in [prefix-geography-plan.md](/home/ray/dev/linkong/planet/docs/prefix-geography-plan.md).
+- Added a dedicated prefix-geography implementation note in [prefix-geography-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-prefix-geography-plan.md).
 - Added recent `15m` collector activity dimensions to BGP coverage output in [bgp_collectors.py](/home/ray/dev/linkong/planet/backend/app/services/bgp_collectors.py) and [visualization.py](/home/ray/dev/linkong/planet/backend/app/api/v1/visualization.py).
 - Added additional BGP detector coverage for `route_leak_candidate` and `path_flap` flows in [test_bgp.py](/home/ray/dev/linkong/planet/backend/tests/test_bgp.py).
 - Added a local Earth cloud texture at [earth_clouds_1024.png](/home/ray/dev/linkong/planet/frontend/public/earth/assets/earth_clouds_1024.png) to avoid remote cloud-map dependency failures.
@@ -574,7 +1071,7 @@ Released: 2026-04-02
 - Improved Earth event animation semantics in [bgp.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/bgp.js) by separating icon pulse from ring expansion so the center marker can breathe while the ring expands independently.
 - Improved Earth texture reliability in [earth.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/earth.js) by switching clouds back to a local static asset under the restored `public/earth` runtime.
 - Improved frontend boot noise in [frontend/index.html](/home/ray/dev/linkong/planet/frontend/index.html) by removing the default Vite favicon request that was generating irrelevant `vite.svg` timeouts during Earth debugging.
-- Improved project planning docs in [bgp-context.md](/home/ray/dev/linkong/planet/docs/bgp-context.md) and [TODO.md](/home/ray/dev/linkong/planet/TODO.md) so the roadmap now explicitly prioritizes `activity layer`, `prefix-centric geography`, and follow-up geofeed/whois work.
+- Improved project planning docs in [bgp-context.md](/home/ray/dev/linkong/planet/docs/technical/earth-bgp-context.md) and [TODO.md](/home/ray/dev/linkong/planet/TODO.md) so the roadmap now explicitly prioritizes `activity layer`, `prefix-centric geography`, and follow-up geofeed/whois work.
 
 ### Fixed
 
@@ -732,7 +1229,7 @@ Released: 2026-03-31
 - Added restart-task Redis helpers and whitelist command mapping in [system_control.py](/home/ray/dev/linkong/planet/backend/app/services/system_control.py).
 - Added detached restart runner orchestration in [system_restart_runner.py](/home/ray/dev/linkong/planet/backend/scripts/system_restart_runner.py).
 - Added `-d` / `--database` support to [planet.sh](/home/ray/dev/linkong/planet/planet.sh) for database-only restarts.
-- Added restart control documentation in [system-service-control.md](/home/ray/dev/linkong/planet/docs/system-service-control.md).
+- Added restart control documentation in [system-service-control.md](/home/ray/dev/linkong/planet/docs/technical/backend-system-service-control.md).
 
 ### Improved
 
@@ -935,7 +1432,7 @@ Released: 2026-03-26
 
 ### Added
 
-- Added a dedicated Earth module remediation plan in [earth-module-plan.md](/home/ray/dev/linkong/planet/docs/earth-module-plan.md).
+- Added a dedicated Earth module remediation plan in [earth-module-plan.md](/home/ray/dev/linkong/planet/docs/deprecated/earth-module-plan.md).
 - Added backend TLE helpers in [satellite_tle.py](/home/ray/dev/linkong/planet/backend/app/core/satellite_tle.py).
 - Added backend support for returning `tle_line1` and `tle_line2` from the satellite visualization API.
 

docs/deprecated/README.md

@@ -0,0 +1,22 @@
+# Deprecated Docs
+
+这个目录用于存放两类文档：
+
+1. 已经完成、主要保留为历史记录的实施计划
+2. 已被现有实现或新方案替代的旧计划
+
+放到这里并不代表这些文档“错误”，而是表示：
+
+- 它们不再适合作为当前开发的主指导文档
+- 如果要了解历史决策、演进路径或旧设计背景，仍然可以参考
+
+当前归档原则：
+
+- 明确写明“已完成”的计划，优先归档
+- 已被正式实现替代、继续放在 `docs/` 根目录会误导后续开发的计划，归档
+- 仍然指导未来开发、尚未完成或仍有明确执行价值的文档，继续保留在 `docs/`
+
+补充说明：
+
+- 一部分归档文档来自外部或临时工作流草案，例如 sisyphus 生成的初稿
+- 这类文档如果有可用内容，应先吸收到 `docs/plans/` 或 `docs/technical/`，再归档保留来源记录

docs/deprecated/hud-panel-component-plan.md

@@ -0,0 +1,165 @@
+# HUD Panel Component Plan
+
+## Goal
+
+Unify Earth HUD panels into a reusable component layer so new panels can share:
+
+- a consistent shell
+- a consistent header
+- a consistent action-button system
+- a consistent body and collapse pattern
+
+## Scope
+
+Target panels:
+
+- `tv-panel`
+- `news-panel`
+- `legend`
+- `layer-panel`
+- `earth-stats`
+- `info-card`
+- settings modal header/actions
+
+## Component Model
+
+### Base shell
+
+- `.hud-panel`
+- `.hud-panel--compact`
+- `.hud-panel--media`
+- `.hud-panel--collapsed`
+- `.hud-panel-hidden`
+- `.hud-panel.is-dragging`
+- `.hud-panel.is-layout-animating`
+
+### Header
+
+- `.hud-panel__header`
+- `.hud-panel__title-group`
+- `.hud-panel__title`
+- `.hud-panel__subtitle`
+- `.hud-panel__chip`
+- `.hud-panel__actions`
+
+Header baseline rule:
+
+- Header title styling is fixed by the component layer and should not drift per panel
+- Title font size, font weight, letter spacing, line height, text color, and vertical alignment come from the shared header tokens and structure
+- Header divider, border treatment, inner spacing, and title-to-actions alignment are part of the same shared baseline
+- Panel-specific header differences should be limited to explicit variants such as `compact` or `media`, or token overrides with documented intent
+- “Looks close enough” local header overrides should be treated as temporary compatibility code and removed during migration
+
+### Actions
+
+- `.hud-panel__action`
+- `.hud-panel__action--icon`
+- `.hud-panel__action--collapse`
+- `.hud-panel__action--close`
+- `.hud-panel__action--refresh`
+- `.hud-panel__action--external`
+
+Action-button baseline rule:
+
+- Header action buttons must have one fixed default style baseline across all HUD panels
+- Default width behavior, padding, icon size, radius, alignment, hover, and active feedback all come from `.hud-panel__action`
+- Panel-specific differences must be expressed through explicit variants or token overrides, not ad-hoc local button rewrites
+- `close` buttons are part of the same default action system and must not silently fall back to a separate legacy box model
+
+### Body
+
+- `.hud-panel__body`
+- `.hud-panel__body--scroll`
+- `.hud-panel__body--collapsible`
+
+### Collapse behavior
+
+- `.hud-panel--collapsed`
+- `.hud-panel--expand-up`
+- `.hud-panel--expand-down`
+
+Adaptive collapse / expand rule:
+
+- HUD panels support two expansion directions:
+  - top-to-bottom expansion
+  - bottom-to-top expansion
+- Expansion direction should be decided at runtime from available viewport space rather than hardcoded per panel
+- Use:
+  - `d` = available distance from the header anchor to the viewport bottom edge
+  - `h` = expected expanded panel height
+  - buffer = `20px`
+- Collapsed-state direction rule:
+  - if `d > h + 20px`, the next action direction is `expand-up`
+  - if `d <= h + 20px`, the next action direction is `expand-down`
+- To avoid jitter around the threshold, the shared controller should keep a small hysteresis band:
+  - if the current direction is already `up`, keep it until `d <= h`
+  - if the current direction is already `down`, keep it until `d > h + 20px`
+- The opposite edge is still a safety guard:
+  - if the chosen side cannot fit at all, fall back to the other side if it can fit
+  - if neither side fully fits, choose the side with more space and let the body scroll
+- If neither direction fully fits, choose the direction with more available space and let the body scroll
+- Collapse icon direction must match the active expansion direction so the icon always describes the real open/close motion
+- The collapse icon describes the next action, not the current state
+- This mapping is fixed component behavior and must not drift per panel:
+  - collapsed + expand-down => `expand_more`
+  - expanded + expand-down => `expand_less`
+  - collapsed + expand-up => `expand_less`
+  - expanded + expand-up => `expand_more`
+- Panels must not combine icon-name swapping with extra CSS rotation for the same collapse control
+- Expansion direction and icon direction must come from one shared source of truth in the component controller
+- The direction decision should be recomputed when opening, resizing the viewport, or restoring a dragged panel near another edge
+
+## Tokens
+
+Promote panel differences into CSS variables instead of duplicating selectors:
+
+- `--hud-panel-padding`
+- `--hud-header-padding`
+- `--hud-header-gap`
+- `--hud-action-padding`
+- `--hud-action-gap`
+- `--hud-action-icon-size`
+- `--hud-body-gap`
+- `--hud-body-max-height`
+- `--hud-chip-radius`
+- `--hud-title-font-size`
+- `--hud-title-font-weight`
+- `--hud-title-letter-spacing`
+- `--hud-title-line-height`
+- `--hud-title-color`
+- `--hud-header-border-color`
+- `--hud-header-divider-opacity`
+- `--hud-expand-direction`
+
+## Migration Order
+
+1. Build the shared component layer in `frontend/public/earth/css/hud.css`
+2. Migrate `tv-panel` and `news-panel` first as the reference implementation
+3. Migrate `legend` and `layer-panel` into a compact variant
+4. Migrate `earth-stats` and `info-card`
+5. Align settings modal header/actions with the same action system
+6. Remove legacy one-off button selectors after verification
+
+## Guardrails
+
+- Do not change panel behavior and data flow during the first pass
+- Keep old class names temporarily as compatibility hooks
+- Prefer variable overrides over per-panel reimplementation
+- Treat header action-button default styling as fixed component API, not per-panel design space
+- Treat header title typography, border, and divider styling as fixed component API, not per-panel design space
+- Treat collapse direction as a component behavior contract, not a one-off panel trick
+- Treat collapse icon semantics as a component behavior contract, not a per-panel visual preference
+- Verify header alignment and drag/collapse behavior after each migration batch
+
+## First Implementation Batch
+
+Batch 1 should only do:
+
+- shared header structure
+- shared action-button system
+- shared title typography and header border/divider baseline
+- shared collapsible body pattern
+- adaptive collapse direction logic and direction-aware collapse icons
+- migration of `tv-panel` and `news-panel`
+
+That keeps risk low while giving the rest of the HUD a stable target to migrate toward.

docs/deprecated/sisyphus-earth-architecture-refactor.md

@@ -1,3 +1,5 @@
+> Archived note: this document was originally created by sisyphus and later reviewed against the main docs set. Useful content has been absorbed into `docs/plans/` where appropriate.
+
 # 地球3D可视化架构重构计划
 
 ## 背景

docs/deprecated/sisyphus-predicted-orbit.md

@@ -1,3 +1,5 @@
+> Archived note: this document was originally created by sisyphus and later reviewed against the main docs set. Useful content has been absorbed into `docs/plans/` where appropriate.
+
 # 卫星预测轨道显示功能
 
 ## TL;DR

docs/deprecated/sisyphus-ue5-client.md

@@ -1,3 +1,5 @@
+> Archived note: this document was originally created by sisyphus and later reviewed against the main docs set. Useful content has been absorbed into `docs/plans/` where appropriate.
+
 # UE5 3D 大屏客户端开发计划
 
 ## 项目概述

docs/deprecated/sisyphus-webgl-instancing-satellites.md

@@ -1,3 +1,5 @@
+> Archived note: this document was originally created by sisyphus and later reviewed against the main docs set. Useful content has been absorbed into `docs/plans/` where appropriate.
+
 # WebGL Instancing 卫星渲染优化计划
 
 ## 背景

docs/plans/README.md

@@ -0,0 +1,37 @@
+# Plans Docs
+
+这里放“未来实施方案和未完成计划”的文档，重点回答：
+
+- 我们准备做什么
+- 为什么要做
+- 分几期做
+- 当前差距和下一步是什么
+
+适合放入这里的内容：
+
+- Earth / BGP / 地形 / 天球实施方案
+- AI Playground 发展计划
+- backend / datasource / agent roadmap
+- UE5 MVP 方案
+
+当前重点入口：
+
+- [earth-mobile-drawer-ui-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-mobile-drawer-ui-plan.md)
+- [earth-compute-center-bgp-style-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-compute-center-bgp-style-plan.md)
+- [earth-renderer-architecture-separation-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-renderer-architecture-separation-plan.md)
+- [earth-predicted-orbit-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-predicted-orbit-plan.md)
+- [earth-webgl-instancing-satellites-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-webgl-instancing-satellites-plan.md)
+- [earth-real-terrain-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-real-terrain-plan.md)
+- [earth-news-source-configuration-and-collector-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-news-source-configuration-and-collector-plan.md)
+- [frontend-ai-playground-development-plan.md](/home/ray/dev/linkong/planet/docs/plans/frontend-ai-playground-development-plan.md)
+- [ue5-mvp-fused-plan.md](/home/ray/dev/linkong/planet/docs/plans/ue5-mvp-fused-plan.md)
+
+不适合放入这里的内容：
+
+- 当前代码结构说明
+- 组件现状和实现入口
+- 已经落地的技术上下文说明
+
+这些应放入：
+
+- [docs/technical/README.md](/home/ray/dev/linkong/planet/docs/technical/README.md)

docs/plans/agents-agent-runtime-roadmap.md

@@ -10,9 +10,9 @@ This document connects three existing planning threads into one implementation r
 
 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)
+- [aiprovider](/home/ray/dev/linkong/planet/docs/technical/agents-aiprovider.md)
+- [datasource-health-plan](/home/ray/dev/linkong/planet/docs/plans/agents-datasource-health-plan.md)
+- [agent-architecture-plan](/home/ray/dev/linkong/planet/docs/plans/agents-agent-architecture-plan.md)
 
 
 ## Big Picture

docs/plans/earth-bgp-region-aggregation-plan.md

@@ -17,7 +17,7 @@ It is an aggregation/view-model layer:
 
 ## Why This Layer Exists
 
-Current product gap from [bgp-context.md](/home/ray/dev/linkong/planet/docs/bgp-context.md):
+Current product gap from [bgp-context.md](/home/ray/dev/linkong/planet/docs/technical/earth-bgp-context.md):
 
 - incident density is naturally low
 - anomaly density is higher, but still not enough to keep the globe expressive all the time
@@ -290,7 +290,7 @@ Each feature should include:
 
 ## Earth Rendering Plan
 
-Detailed visual layering guidance is expanded in [bgp-earth-rendering-plan.md](/home/ray/dev/linkong/planet/docs/bgp-earth-rendering-plan.md).
+Detailed visual layering guidance is expanded in [bgp-earth-rendering-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-bgp-earth-rendering-plan.md).
 
 ### Layer Relationship
 

docs/plans/earth-celestial-background-plan.md

@@ -0,0 +1,715 @@
+# Earth 天球背景与日月位置实施方案
+
+## 目标
+
+为 Earth 大屏增加一套真正可用的天文背景层，覆盖三件事：
+
+1. 用真实天球背景替换当前随机星点
+2. 在当前时间下显示太阳与月亮的相对位置
+3. 让太阳方向同时驱动地球受光，形成更可信的昼夜关系
+
+本方案优先追求：
+
+- 与当前 Three.js Earth 架构兼容
+- 风险可控
+- 先落地一版真实感明显提升的 V1
+- 为后续更严格的天文参考系升级预留余地
+
+## 当前现状
+
+当前 Earth 的基础条件已经具备：
+
+- 地球、云层、地形、网格都基于 Three.js，主渲染入口在 [frontend/public/earth/js/main.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/main.js)
+- 地球实体创建在 [frontend/public/earth/js/earth.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/earth.js)
+- 当前所谓“宇宙背景”只是 `createStars()` 生成的随机星点，不是真实星图
+- Earth 已有倾角常量 `EARTH_CONFIG.tiltRad`，位于 [frontend/public/earth/js/constants.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/constants.js)
+- 主循环 `animate()` 已稳定运行，可在其中接入天体更新逻辑
+
+这意味着：
+
+- 不需要重写 Earth
+- 可以在现有 scene/world 层新增一个 celestial layer
+- 第一阶段不必拆 Earth / satellite / cable 的参考系
+
+## 总体策略
+
+采用“两层现实”设计：
+
+### 1. 世界层（world-space celestial layer）
+
+用于放置：
+
+- 天球背景
+- 太阳
+- 月亮
+- 太阳光方向
+
+这些对象不挂在 `earthObj` 上，而是直接放在 `scene` 中。
+
+### 2. 地球层（earth-fixed layer）
+
+继续保持当前结构：
+
+- 海缆
+- 登陆点
+- 卫星点与轨迹
+- BGP 覆盖
+- 地球纹理、云层、地形
+
+这些对象继续挂在 `earthObj` 下，不打断现有交互。
+
+## 为什么先这样做
+
+当前用户交互是“拖动地球本体”，而不是“移动相机绕惯性系观测”。  
+如果现在直接做严格惯性参考系改造，会同时影响：
+
+- `earthObj.rotation`
+- 卫星轨迹与锁定逻辑
+- 海缆与登陆点附着关系
+- resetView / autoRotate / hover / click 等交互链路
+
+所以第一阶段只做：
+
+- 真正的天空
+- 真正的日月方向
+- 不碰现有 Earth 附着对象的语义
+
+## 推荐技术选型
+
+### 天文计算库
+
+推荐：
+
+- [Astronomy Engine](https://github.com/cosinekitty/astronomy)
+
+原因：
+
+- 有 JavaScript 版本
+- 支持 Sun / Moon 的矢量与坐标变换
+- 精度、可扩展性都比轻量太阳高度角库更适合本项目
+- 后续若要加行星、月相、黄道、赤道网，也能继续沿用
+
+不作为主选的库：
+
+- [SunCalc](https://github.com/mourner/suncalc)
+
+原因：
+
+- 更偏本地观察者视角的太阳/月亮高度角
+- 用于“地面日出日落”很好
+- 但不如 Astronomy Engine 适合做真实天球与后续空间参考系扩展
+
+### Three.js 表现层
+
+推荐组合：
+
+- 天球：内翻球壳 + 星图纹理
+- 太阳：`THREE.Sprite`
+- 月亮：`THREE.Sprite` 或小型 `THREE.Mesh`
+- 太阳光：`THREE.DirectionalLight`
+
+参考：
+
+- [Three.js SpriteMaterial](https://threejs.org/docs/pages/SpriteMaterial.html)
+
+## 天球背景资源与星体数据来源
+
+为避免把“视觉背景”和“可计算天体位置”混为一谈，本方案明确分成两类资源：
+
+### 1. 背景资源：全天星图贴图
+
+用于 Phase 1 的“真实天空背景”。
+
+推荐优先来源：
+
+- NASA SVS 的 Tycho 全天星图
+  - [The Tycho Catalog Skymap - Version 2.0](https://svs.gsfc.nasa.gov/3572/)
+- NASA Deep Star Maps 2020
+  - SatelliteMap.space 在 credits 中明确提到其使用了 `NASA Deep Star Maps 2020 - High-resolution star field (1.7 billion stars from Gaia DR2)` 作为星空视觉资源
+  - 这说明行业内成熟实现并不一定直接渲染全部星表点，而很可能先使用一张高质量官方深空星图作为背景层
+- 如需后续替换，也可评估 ESA / Gaia 的全天 sky map 资源
+  - [Gaia DR3 stories](https://www.cosmos.esa.int/web/gaia/dr3-stories)
+
+建议要求：
+
+- 使用官方来源或官方衍生可复用资源
+- 等距矩形投影（equirectangular）
+- 坐标定义尽量明确为赤道坐标展开
+- 分辨率建议至少 `4k`
+- 颜色不要过亮，避免压过 Earth HUD 前景
+- 尽量优先选择官方天文机构已经生产好的深空图，而不是自行拼接低质量星空纹理
+
+建议本地资源目录：
+
+- `frontend/public/earth/assets/celestial/starmap_equatorial_4k.jpg`
+
+### 2. 位置数据：星表与天体计算
+
+用于 Phase 2+ 的“位置正确的星体”。
+
+推荐来源分两层：
+
+- 太阳、月亮位置
+  - 使用 [Astronomy Engine](https://github.com/cosinekitty/astronomy)
+- 恒星位置
+  - 第一优先：Hipparcos / Tycho
+    - [Hipparcos overview](https://www.cosmos.esa.int/web/Hipparcos)
+    - [Hipparcos catalogues](https://www.cosmos.esa.int/web/hipparcos/catalogues)
+  - 第二优先：Gaia
+    - [Gaia DR3 stories](https://www.cosmos.esa.int/web/gaia/dr3-stories)
+
+建议策略：
+
+- V1：背景球壳只用全天星图，不立即生成全量恒星点
+- V2：只挑选亮星（例如星等 `< 5.5`）生成恒星点层
+- V3：如果确实需要更丰富的星场，再逐步扩展到更深星等
+
+这样做的原因：
+
+- 背景球壳负责“天球真实感”
+- 亮星点负责“位置正确、可后续标注和高亮”
+- 不需要一开始就处理数十万甚至数百万颗星
+
+### 3. 对外部成熟实现的参考结论
+
+`SatelliteMap.space` 的公开 credits 提供了一个很有价值的参考样板：
+
+- 图形渲染使用 `TWGL.js`
+- 天文计算使用 `Skyfield` 与 `Astronomia`
+- 星空/天球视觉资源使用 `NASA Deep Star Maps 2020`
+
+这给本项目的启发是：
+
+- “真实感强的天球背景”完全可以先依赖官方高质量深空图
+- “位置正确的动态天体”则应依赖单独的天文计算链路
+- 没有必要在第一版就直接渲染完整星表
+
+因此本项目推荐继续坚持两层拆分：
+
+- 背景层：官方深空图 / 全天星图
+- 计算层：太阳、月亮与后续亮星点
+
+## 如何保证星体位置正确
+
+位置正确不是只看“图看起来像”，而是要统一参考系和转换链路。
+
+### 1. 统一坐标基准
+
+本方案推荐统一使用：
+
+- `J2000` 赤道坐标系作为恒星位置基准
+
+原因：
+
+- Hipparcos / Tycho 资料和大量天文可视化都容易映射到该基准
+- 太阳、月亮也可以通过 Astronomy Engine 转到同一坐标系
+- 这样背景、恒星点、太阳、月亮就能共用一套 sky orientation
+
+### 2. 背景贴图与点位必须使用同一展开逻辑
+
+如果背景球壳使用赤道坐标全天图，那么：
+
+- 亮星点也必须按赤道坐标贴到同一球面方向
+- 太阳/月亮 sprite 也必须按赤道坐标转换后落到同一 world-space
+
+否则会出现：
+
+- 背景银河带是对的
+- 但太阳/月亮或亮星点飘到不匹配的位置
+
+### 3. RA / Dec 到 Three.js 坐标的落点方式
+
+亮星点和日月方向最终都要转成单位球面向量。
+
+概念步骤：
+
+1. 读取赤经 `RA`
+2. 读取赤纬 `Dec`
+3. 转成弧度
+4. 映射到单位球面向量
+5. 再根据 Three.js 当前世界坐标定义做轴向映射
+
+参考公式：
+
+```text
+x = cos(dec) * cos(ra)
+y = sin(dec)
+z = cos(dec) * sin(ra)
+```
+
+实际接入 Three.js 时，需要做一次项目内坐标轴校准：
+
+- 验证 `RA = 0h`
+- 验证 `RA = 6h`
+- 验证北天极
+- 验证银河带主方向
+
+然后确定最终的：
+
+- `x/y/z` 对应 Three.js 哪个轴
+- 是否需要 `z` 取反
+- 是否需要整体再做一个固定 `rotation`
+
+建议把这层显式封装在：
+
+```js
+function equatorialToWorldVector(raRad, decRad)
+```
+
+不要把轴映射散落在不同模块里。
+
+### 4. 背景球壳与恒星点的关系
+
+推荐最终组合：
+
+- 背景层：全天星图球壳
+- 点位层：亮星点
+- 动态层：太阳 / 月亮
+
+这样有三个好处：
+
+- 背景层提供密集真实的天空纹理
+- 亮星点提供位置正确、可扩展的标注基础
+- 太阳/月亮提供与时间相关的真实动态对象
+
+## 数据与资源建议清单
+
+### 推荐首批引入资源
+
+1. 全天星图
+- 来源：NASA Tycho all-sky map
+- 用途：背景球壳纹理
+
+2. 月亮纹理
+- 用途：Phase 4 月相表现
+- 路径建议：
+  - `frontend/public/earth/assets/celestial/moon_albedo_2k.jpg`
+
+3. 太阳 glow 贴图
+- 用途：太阳 sprite halo
+- 路径建议：
+  - `frontend/public/earth/assets/celestial/sun_glow.png`
+
+### 推荐首批数据文件
+
+如果要上亮星层，建议新增一个预处理后的轻量数据文件：
+
+- `frontend/public/earth/assets/celestial/bright-stars.json`
+
+建议字段：
+
+```json
+[
+  {
+    "id": 32349,
+    "name": "Sirius",
+    "raDeg": 101.2875,
+    "decDeg": -16.7161,
+    "mag": -1.46,
+    "colorIndex": 0.00
+  }
+]
+```
+
+建议不要在浏览器里直接吞原始 Gaia 大表，而是先离线裁剪成：
+
+- 只保留亮星
+- 只保留渲染必需字段
+- JSON 或二进制轻量格式
+
+## 资源与数据实施路线
+
+### 路线 A：先做可用版本（推荐）
+
+1. 引入 NASA Tycho 全天图
+   - 或评估替换为更接近 SatelliteMap.space 路线的 `NASA Deep Star Maps 2020`
+2. 实现背景球壳
+3. 用 Astronomy Engine 计算太阳/月亮方向
+4. 暂不做亮星点
+
+优点：
+
+- 最快见效
+- 风险最低
+- 就能明显提升天球真实感
+
+### 路线 B：在 A 基础上增强
+
+1. 离线生成 `bright-stars.json`
+2. 浏览器端渲染亮星点
+3. 后续可加：
+   - 星座线
+   - 亮星名称
+   - 特定星体高亮
+
+优点：
+
+- 背景真实感和“位置正确的可交互星体”同时兼顾
+
+## 代码模块建议细化
+
+### 新增模块
+
+- `frontend/public/earth/js/celestial.js`
+  - 管理天球背景
+  - 管理太阳/月亮
+  - 管理亮星层（后续）
+
+- `frontend/public/earth/js/celestial-data.js`
+  - 资源路径
+  - 星图方向配置
+  - 亮星数据加载（后续）
+
+### 建议函数设计
+
+```js
+export function initCelestialLayer(scene)
+export function updateCelestialLayer(date)
+export function setCelestialVisibility(visible)
+export function disposeCelestialLayer()
+
+function loadStarMapTexture()
+function createSkySphere(texture)
+function createSunSprite()
+function createMoonSprite()
+function getSunEquatorialPosition(date)
+function getMoonEquatorialPosition(date)
+function equatorialToWorldVector(raRad, decRad)
+```
+
+### 推荐后续预处理脚本
+
+如要引入亮星层，建议单独做离线脚本：
+
+- `scripts/build_bright_stars.py`
+
+职责：
+
+- 从 Hipparcos / Tycho 源数据读取
+- 过滤亮星
+- 生成 `bright-stars.json`
+
+这样浏览器端只消费轻量结果，不承担大表解析成本。
+
+## 分阶段实施
+
+## Phase 1：真实天球背景
+
+### 目标
+
+用真实全天星图替换当前随机星点背景。
+
+### 做法
+
+1. 新增一张全天星图纹理
+
+建议路径：
+
+- `frontend/public/earth/assets/celestial/starmap_equatorial_4k.jpg`
+
+纹理要求：
+
+- 等距矩形投影
+- 赤经/赤纬坐标展开
+- 无地平线、无地景遮挡
+- 尽量深色、弱干扰，适合大屏 HUD 叠加
+
+2. 新增天球球壳
+
+新增模块：
+
+- [frontend/public/earth/js/celestial.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/celestial.js)
+
+建议接口：
+
+```js
+export function initCelestialLayer(scene)
+export function updateCelestialLayer(date, camera, earth)
+export function disposeCelestialLayer()
+```
+
+3. 实现一个大半径内翻球体
+
+建议参数：
+
+- 半径：`600 ~ 900`
+- 材质：`MeshBasicMaterial`
+- `side: THREE.BackSide`
+- 不受场景光照影响
+- 始终围绕场景中心
+
+### 验收标准
+
+- 初始加载后背景不再是随机星点
+- 旋转地球时，背景保持为稳定天球而不是跟地球一起转
+- 不明显干扰海缆/卫星/BGP 的前景识别
+
+## Phase 2：太阳与月亮真实位置
+
+### 目标
+
+在当前 UTC 时间下，计算太阳与月亮在天球中的方向，并显示出来。
+
+### 做法
+
+1. 在 `celestial.js` 内封装天体位置计算
+
+建议函数：
+
+```js
+function getSunDirection(date)
+function getMoonDirection(date)
+```
+
+输出统一为 world-space `THREE.Vector3`
+
+2. 太阳显示
+
+- 一个暖色发光 sprite
+- 比月亮更大、更亮
+- 可选添加柔和 halo
+
+3. 月亮显示
+
+- 一个较小 sprite 或 sphere
+- 灰白偏冷色
+- 后续 Phase 3 再做月相
+
+4. 更新频率
+
+不要每帧重新做完整天文计算，建议：
+
+- 每 30 秒或 60 秒重算一次真实位置
+- 渲染帧内做平滑过渡
+
+### 验收标准
+
+- 页面可见太阳与月亮两个对象
+- 时间变化时位置会更新
+- 日月不会跟随地球局部旋转而错误附着
+
+## Phase 3：太阳驱动地球受光
+
+### 目标
+
+让地球光照方向与太阳方向一致，不再使用写死的固定主光。
+
+### 做法
+
+1. 替换或接管当前主定向光
+
+当前 [frontend/public/earth/js/main.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/main.js) 中 `addLights()` 里用了固定方向的 `DirectionalLight`。
+
+建议改为：
+
+- 保留环境补光
+- 主太阳光方向由 `sunDirection` 决定
+
+2. 太阳光参数建议
+
+- `DirectionalLight` 颜色偏暖白
+- 强度略高于当前主光
+- 保留一个弱背光作为氛围补偿，避免背面过死黑
+
+3. 先不做物理级大气散射
+
+第一版只要求：
+
+- 亮面与暗面方向真实
+- 云层和大气仍保持当前风格
+
+### 验收标准
+
+- 地球明暗面会随太阳方向改变
+- 太阳 sprite 和地球亮面方向一致
+- 不破坏现有海缆、卫星、BGP 的可见性
+
+## Phase 4：月相与天文细节增强
+
+### 目标
+
+在日月真实位置基础上增加更强的“天文可信度”。
+
+### 可选项
+
+1. 月相
+
+- 根据日月夹角计算 illuminated fraction
+- 用月相纹理或 shader 表达盈亏
+
+2. 赤道/黄道辅助线
+
+- 可作为开发调试层，不默认显示
+
+3. 太阳 terminator 增强
+
+- 给地球夜面加入更自然的 night tint
+- 未来可叠加城市夜光纹理
+
+4. 天文时间入口
+
+- 设置中加入“当前时刻 / 指定时刻 / 加速时间”模式
+
+### 验收标准
+
+- 月亮不再只是一个静态圆点
+- 后续扩展行星或观测模式时无需推倒重来
+
+## Phase 5：严格参考系升级（可选，不作为 V1 必做）
+
+### 目标
+
+把 Earth 从“用户旋转球体”升级为“真实地球姿态 + 用户观察姿态”的双层模型。
+
+### 需要处理的问题
+
+- 地球自转角与 UTC 的一致性
+- 赤道坐标系、地固坐标系、相机交互层分离
+- 卫星轨道显示与 Earth 旋转同步关系
+- resetView 和 autoRotate 的语义重定
+
+### 风险
+
+这一步会影响：
+
+- [frontend/public/earth/js/main.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/main.js)
+- [frontend/public/earth/js/satellites.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/satellites.js)
+- [frontend/public/earth/js/cables.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/cables.js)
+- [frontend/public/earth/js/controls.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/controls.js)
+
+因此不建议与 V1 同时推进。
+
+## 代码改造清单
+
+## 1. 新增文件
+
+- [frontend/public/earth/js/celestial.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/celestial.js)
+
+职责：
+
+- 管理天球背景、太阳、月亮
+- 对外暴露 init/update/dispose
+
+## 2. 修改 `constants.js`
+
+文件：
+
+- [frontend/public/earth/js/constants.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/constants.js)
+
+新增：
+
+```js
+export const CELESTIAL_CONFIG = {
+  sphereRadius: 800,
+  updateIntervalMs: 60000,
+  sunSpriteScale: 28,
+  moonSpriteScale: 16,
+  sunLightIntensity: 1.25,
+  ambientIntensity: 0.28,
+  backLightIntensity: 0.18,
+};
+```
+
+## 3. 修改 `main.js`
+
+文件：
+
+- [frontend/public/earth/js/main.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/main.js)
+
+主要改动：
+
+1. `init()` 中：
+   - 初始化 celestial layer
+2. `addLights()` 中：
+   - 把固定太阳光改成可更新的 celestial sun light
+3. `animate()` 中：
+   - 每帧调 `updateCelestialLayer()`
+4. `destroy()` 中：
+   - 清理 celestial 资源
+
+## 4. 修改 `earth.js`
+
+文件：
+
+- [frontend/public/earth/js/earth.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/earth.js)
+
+主要改动：
+
+- `createStars()` 逐步退役
+- 第一阶段可先保留作为 fallback
+- 当真实星图加载成功后，不再显示随机星点
+
+## 5. 新增资源
+
+目录建议：
+
+- `frontend/public/earth/assets/celestial/`
+
+建议至少包含：
+
+- `starmap_equatorial_4k.jpg`
+- `sun_glow.png`
+- `moon_albedo_2k.jpg`
+
+## 数据流设计
+
+```mermaid
+flowchart TD
+    A["main.js:init()"] --> B["initCelestialLayer(scene)"]
+    B --> C["创建天球球壳"]
+    B --> D["创建太阳 sprite + 主定向光"]
+    B --> E["创建月亮 sprite"]
+
+    F["animate()"] --> G["updateCelestialLayer(now, camera, earth)"]
+    G --> H["Astronomy Engine 计算 Sun/Moon 方向"]
+    H --> I["更新 sun sprite / moon sprite 位置"]
+    H --> J["更新太阳 DirectionalLight 方向"]
+    J --> K["地球昼夜方向变化"]
+```
+
+## 风险与注意事项
+
+### 1. 星图投影方向容易反
+
+这会表现为：
+
+- 星图左右镜像
+- 赤经方向颠倒
+- 日月位置和背景对不上
+
+建议：
+
+- 先做一个开发调试模式
+- 显示赤经/赤纬参考点，快速校正纹理朝向
+
+### 2. 不要让天球跟随 Earth 旋转
+
+天球背景和日月必须属于 scene/world，而不是 `earthObj`。
+
+### 3. 不要每帧做重型天文计算
+
+真实位置更新应节流，否则会浪费 CPU。
+
+### 4. 月亮先求“方向正确”，再求“月相精致”
+
+月相属于第二步优化，不应阻塞 V1 上线。
+
+## 推荐实施顺序
+
+1. 新建 `celestial.js`
+2. 用星图球壳替换随机星点
+3. 接入 Astronomy Engine
+4. 加太阳/月亮 sprite
+5. 用太阳方向驱动主光
+6. 再决定要不要做月相和更严格参考系
+
+## 最终建议
+
+对于当前 Planet Earth，最稳妥的方案是：
+
+- 先做真实天球背景
+- 再做真实太阳/月亮方向
+- 再让太阳驱动地球受光
+- 暂时不做 Earth 参考系重构
+
+这样可以在不破坏现有 Earth 交互和图层系统的前提下，显著提升空间感、真实感和演示说服力。

docs/plans/earth-compute-center-bgp-style-plan.md

@@ -0,0 +1,372 @@
+# Earth Compute Center BGP-Style Plan
+
+## Goal
+
+这份文档定义如何按照 BGP 模块的产品方式，把“算力中心”提升为 Earth 上的一级能力。
+
+这里的“按 BGP 方式”指的是：
+
+- 有独立的数据语义和接口入口
+- 有独立的 Earth 图层与图例
+- 有独立的 hover / click / 选中态 / 详情卡
+- 有独立的统计口径与后续专题页扩展空间
+
+这里的“按 BGP 方式”不指：
+
+- 机械复制 BGP 的 anomaly / incident / collector 三层事件模型
+- 为静态算力设施强行引入不必要的复杂告警语义
+
+算力中心本质上更接近“长期基础设施分布层”，不是“高频动态异常层”。
+因此应该复用 BGP 的模块化方法，而不是照搬 BGP 的事件结构。
+
+## Why
+
+当前仓库里已经有算力相关基础：
+
+- 后端已有 `top500` 和 `epoch_ai_gpu` 数据采集
+- 可视化接口已有 `/api/v1/visualization/geo/supercomputers` 和 `/api/v1/visualization/geo/gpu-clusters`
+- Earth 信息卡已对 `supercomputer` 和 `gpu_cluster` 做了基础类型兼容
+
+但当前能力还停留在“数据可取到”的阶段，没有形成像 BGP 那样完整的可视化模块：
+
+- Earth 缺少独立的算力图层加载模块
+- 缺少算力 marker 体系和视觉层级
+- 缺少算力图例、统计、开关和搜索接入
+- 缺少与海缆、BGP、卫星的关系表达
+- 缺少算力专题页和后续告警/研判扩展入口
+
+所以当前真正的缺口不是“有没有数据”，而是“有没有产品级模块”。
+
+## Core Principle
+
+算力中心应当采用和 BGP 一致的模块化分层：
+
+1. 数据层：稳定的数据契约和 GeoJSON 输出
+2. 渲染层：独立的 Earth 图层、marker 和视觉状态管理
+3. 交互层：hover、click、锁定态、详情卡、图例和统计
+4. 扩展层：后续专题页、关系分析、告警和 AI 研判
+
+但语义上必须保持算力中心自身的特点：
+
+- `site / center` 是主对象，不是事件
+- `capacity / rank / vendor / operator / status` 是主信息，不是异常严重度
+- `distribution / concentration / dependency` 是后续分析方向，不是第一阶段必须项
+
+## Recommended Scope
+
+第一版“算力中心”建议统一承载两类对象：
+
+- `supercomputer`
+- `gpu_cluster`
+
+并在 Earth 上收口为一个主题层：`compute_centers`
+
+这样做有几个好处：
+
+- 用户看到的是统一的“算力基础设施”语义，而不是零散数据源
+- 后端仍可保留 `top500` 和 `epoch_ai_gpu` 的来源差异
+- 前端可以在一个图层里再细分两种 marker 语言
+
+## Current Gap
+
+和 BGP 对比，当前差距主要在下面几层。
+
+### 1. Data Contract Gap
+
+现在的算力 GeoJSON 还是通用 `collected_data` 输出思路，字段较轻：
+
+- `gpu_cluster` 只有基础名称和地点
+- `supercomputer` 只暴露一部分性能字段
+- 缺少统一的 `site_type / operator / capacity_band / source / updated_at / confidence`
+- 缺少统一的算力层聚合出口
+
+### 2. Earth Rendering Gap
+
+当前 Earth 里没有类似 `bgp.js` 的算力模块：
+
+- `constants.js` 没有算力 API 路径和视觉配置
+- `main.js` 没有算力加载、拾取、状态同步和 HUD 更新
+- `controls.js` 没有算力图层开关和启动加载优先级
+- `layer-startup-tasks.js` 没有算力启动任务
+- `legend.js` / `ui.js` 没有算力统计与图例模式
+
+### 3. Interaction Gap
+
+虽然 `info-card.js` 支持基础字段，但还没有形成 BGP 那种完整交互链路：
+
+- 没有 hover / selected / dimmed 的视觉状态
+- 没有算力对象专属 tooltip 与摘要文案
+- 没有锁定后与其他基础设施的联动高亮
+- 没有搜索、统计卡和详情组织方式
+
+### 4. Product Expansion Gap
+
+当前还没有“算力中心”专题页与分析语义：
+
+- 没有全球分布/国家聚合/厂商聚合视图
+- 没有算力与海缆/BGP/区域的关系表达
+- 没有 AI brief / assessment 的后续落点
+
+## Architecture Direction
+
+推荐把算力中心做成“BGP 同级能力”，但采用更适合静态基础设施的结构。
+
+### Backend
+
+建议新增统一聚合接口，例如：
+
+- `/api/v1/visualization/geo/compute-centers`
+
+它的职责是把：
+
+- `top500`
+- `epoch_ai_gpu`
+
+统一转换成一个主题层输出，同时保留对象细分类型：
+
+- `site_type: supercomputer | gpu_cluster`
+
+建议统一字段至少包括：
+
+- `id`
+- `name`
+- `site_type`
+- `country`
+- `city`
+- `latitude`
+- `longitude`
+- `operator`
+- `vendor`
+- `capacity_value`
+- `capacity_unit`
+- `capacity_band`
+- `rank`
+- `source`
+- `updated_at`
+- `location_precision`
+- `geography_mode`
+- `is_estimated`
+- `estimated_reason`
+- `metadata`
+
+这里建议优先做“统一聚合出口”，而不是一开始就新增独立数据库表。
+
+原因：
+
+- 当前源数据更新频率低，先复用 `collected_data` 成本更低
+- 可以先把 Earth 产品体验做完整
+- 如果后续要做历史趋势、关系推断、告警，再评估是否拆成独立模型
+
+### Frontend Earth
+
+建议新增独立模块，例如：
+
+- `frontend/public/earth/js/compute-centers.js`
+
+职责参照 `bgp.js`：
+
+- 拉取算力中心 GeoJSON
+- 创建 marker
+- 管理 hover / selected / dimmed 状态
+- 输出图例项
+- 输出统计摘要
+- 提供 overlay 和详情格式化辅助函数
+
+推荐视觉分层：
+
+1. `supercomputer` 用更稳定、更规整的设施型符号
+2. `gpu_cluster` 用更活跃、更现代的密度型符号
+3. 选中态通过 halo / ring / related infrastructure highlight 表达
+
+视觉上应避免把算力中心做成“BGP 事件点”那种高频脉冲风格。
+它应该更像长期存在的高价值设施。
+
+## Phases
+
+## Phase 1: Unified Earth Layer
+
+目标：
+
+- 先把算力中心做成 Earth 上可用、可点、可解释的一级图层
+
+工作项：
+
+- 新增统一算力 GeoJSON 接口
+- 新增 `compute-centers.js`
+- 在 `constants.js` 增加 API 路径和视觉配置
+- 在 `controls.js` 增加算力图层开关与启动元数据
+- 在 `layer-startup-tasks.js` 增加算力启动加载任务
+- 在 `main.js` 接入算力拾取、hover、click、锁定态和 HUD 统计
+- 在 `ui.js` / `legend.js` / `index.html` 增加算力统计与图例入口
+- 在 `info-card.js` 提升算力详情字段组织
+- 对无法精确定位、但可按国家或弱线索推测的大概位置，仍然生成地图点位
+- 这类对象必须带显式“估算位置”状态，例如图标问号角标与详情说明
+
+完成标准：
+
+- Earth 上能独立显示/隐藏算力中心
+- 两类对象有可区分的视觉表达
+- hover / click / 详情卡 / 图例 / 统计全部打通
+- 精确位置与估算位置在图标或文案上可区分，不会误导为同一精度
+- 不干扰现有海缆、卫星、BGP 的交互链路
+
+## Phase 2: Relationship Layer
+
+目标：
+
+- 让算力中心不只是“点”，而是和其他基础设施产生上下文关系
+
+工作项：
+
+- 建立算力中心与国家/区域聚合摘要
+- 增加与附近海缆登陆点的关系提示
+- 增加与 BGP 事件/观测范围的空间邻近提示
+- 增加与卫星覆盖或区域连通性的实验性提示
+
+完成标准：
+
+- 点击算力中心时，用户能看到“它和哪些基础设施相关”
+- 信息表达以辅助判断为主，不做夸张推断
+
+## Phase 3: Compute Center Observatory
+
+目标：
+
+- 把算力中心从 Earth 图层扩展成独立专题观测能力
+
+工作项：
+
+- 新增算力中心专题页
+- 提供国家/厂商/类型/容量分布统计
+- 支持列表、筛选、详情和历史快照
+- 预留 AI brief / assessment 入口
+
+完成标准：
+
+- 算力中心不再只是 Earth 上的视觉点位
+- 能作为独立业务上下文进入日常观察与研判
+
+## Phase 4: Alerts And Assessment
+
+目标：
+
+- 在不滥造“假动态告警”的前提下，引入真正有价值的变化感知
+
+候选方向：
+
+- 新增大规模算力中心
+- 既有中心容量显著变化
+- 国家/区域集中度显著变化
+- 高价值中心与关键网络基础设施关系变化
+
+完成标准：
+
+- 告警来自可解释的结构变化
+- 不把静态数据硬做成噪声式实时事件流
+
+## Implementation Notes
+
+建议按下面顺序推进：
+
+1. 先统一 GeoJSON 契约
+2. 再做 Earth 独立模块和图层开关
+3. 再补详情卡、图例和统计
+4. 最后才做关系层和专题页
+
+这样可以避免一开始把范围摊得过大。
+
+## Unknown Location Strategy
+
+由于部分算力数据源不会直接提供经纬度，未知位置补全不能只依赖“继续找 API 字段”。
+更稳妥的方式是做成一条分层富化链路，而不是单一猜测规则。
+
+推荐按下面优先级推进：
+
+1. 直接源信息
+
+- 源记录显式给出 `latitude / longitude`
+- 源记录给出 `city / region / facility / campus / operator`
+- 源页面详情、内嵌 JSON、结构化元数据、新闻稿链接里能抽出地点线索
+
+2. 名称与机构归一化
+
+- 建立 `canonical_name / aliases / operator / facility` 归一化表
+- 把 `cluster name`、`operator`、`campus name` 归一到同一个实体
+- 优先解决同一对象多写法导致的命中失败，而不是先扩大猜测范围
+
+3. 本地位置注册表
+
+- 用仓库内可维护的 registry 保存高价值对象的位置知识
+- 每条记录至少包含：`canonical_name`、`aliases`、`operator`、`country`、`region`、`city`、`lat`、`lon`、`confidence`、`source_note`
+- 转换层优先读取 registry，避免地点知识长期散落在转换代码里
+
+4. 分层回退定位
+
+- `precise`
+- `estimated_site`
+- `estimated_city`
+- `estimated_region`
+- `estimated_national_hub`
+- `estimated_country`
+
+这里建议把“国家内主要算力城市”作为国家质心之前的一层。
+例如没有美国精确位置时，优先考虑已知的主要算力/数据中心城市候选，而不是直接落在几何质心。
+
+5. 候选证据富化
+
+- 如果源 API 无地点信息，可以允许采集链路读取公开辅助证据
+- 例如机构官网、数据中心介绍页、新闻稿、百科型页面、公开 PDF
+- 但只提取“地点线索”，不把外部页面上的经纬度当真值直接写回
+
+6. 人工校验闭环
+
+- 对高价值且仍然未知的对象输出待核验清单
+- 把人工确认结果回写到位置注册表
+- 后续采集继续优先复用这层人工确认结果
+
+### Additional Solution Paths
+
+除了静态映射表，还可以考虑下面这些办法：
+
+- 基于国家和运营方建立“主要园区候选集”，用稳定散列把同国未知节点分散到若干可信城市，而不是全部压到一个点
+- 基于数据中心/云厂商公开 region 列表建立 `operator -> city set` 候选映射，用于云 GPU 集群类对象
+- 把“估算依据”结构化，例如 `matched_alias`、`matched_operator`、`matched_city_text`、`fallback_country_hub`
+- 给位置补全增加 `last_verified_at`，便于后续按时间重新校验老旧映射
+- 单独维护“不可可靠定位”状态；这类对象仍可在国家级聚合统计中出现，但可以允许用户在地图上过滤掉
+- 后续如果你们愿意投入更多，可把这条链路做成小型 enrichment pipeline，而不是仅在 API 转换时临时判断
+
+## Non-Goals
+
+第一阶段不建议做这些内容：
+
+- 不复制 BGP 巡航模式到算力中心
+- 不先做复杂实时 websocket 推送
+- 不先引入独立 `compute_center_incident` 一类模型
+- 不先做全量 AI 分析面板
+
+原因是算力中心的第一需求是“被看清楚”，不是“被实时播报”。
+但“被看清楚”不等于“只显示精确坐标对象”。
+对于没有精确经纬度、但能推测到国家或区域级位置的算力中心，应优先以上图并标注估算状态的方式处理，而不是直接在地图上消失。
+
+## Acceptance Checklist
+
+- 后端存在统一的算力中心 GeoJSON 出口
+- Earth 有独立算力图层模块，而不是散落在 `main.js`
+- 页面上有清晰的算力开关、图例和统计
+- `supercomputer` 和 `gpu_cluster` 在视觉和详情上都可区分
+- 估算位置对象在地图和详情中都有明确状态提示
+- 现有 BGP / 海缆 / 卫星功能无回归
+- 代码结构上为后续专题页和关系分析留出了明确扩展点
+
+## Summary
+
+这项工作的本质不是“再多画几个点”。
+
+它应该把算力中心从已有数据源，升级成与 BGP 同级的 Earth 观测主题：
+
+- 有独立语义
+- 有独立图层
+- 有独立交互
+- 有后续分析扩展能力
+
+推荐先完成 Phase 1，把算力中心做成真正可用的 Earth 一级模块，再继续推进关系层和专题页。

docs/plans/earth-mobile-drawer-ui-plan.md

@@ -0,0 +1,400 @@
+# Earth Mobile Drawer UI Plan
+
+## 背景
+
+当前 Earth 移动端已经补上了基础触控能力，例如：
+
+- 单指拖拽旋转地球
+- 双指缩放
+- 点击阈值和基础事件隔离
+
+但移动端 UI 仍然存在一个根本问题：
+
+它还在沿用桌面 HUD 的内容切分方式，只是把原来的 panel、modal、toolbar 改位置、改层级、改容器。这样虽然能快速复用旧代码，但手机端体验仍然是生硬的，因为：
+
+- 信息密度和结构是按桌面设计的
+- 面板标题、关闭、折叠、开关项是桌面心智，不是手机心智
+- 很多内容只是“被塞进抽屉”，而不是为抽屉重新设计
+- 设置里仍然带有“显示/隐藏某些 panel”的思路，但移动端本来就不应该存在那些独立 panel
+
+因此本计划进一步收紧：
+
+移动端不只是“底部抽屉化”，而是**重新设计一套 fit 抽屉体系的 mobile-first UI**。
+
+## 新目标
+
+1. 手机端不再使用现有 `toolbar` 作为主入口。
+2. 手机端不再使用现有独立 `panel / modal / sheet` 作为直接 UI 单元。
+3. 手机端统一采用“底部抽屉 + 顶部标题 + tab 切换 + 卡片内容”的单前景模式。
+4. 抽屉内部每个 tab 页面都按移动端重新设计内容结构，而不是直接复用旧 panel 结构。
+5. 设置页移除“显示/隐藏 panel”的桌面遗留配置。
+6. 媒体页拆成两个移动端页面：`新闻` 与 `TV`，都归入抽屉体系。
+7. 桌面端保持现有 HUD 体系，不回退。
+
+## 核心原则
+
+### 1. 只复用数据和状态，不复用桌面 UI 结构
+
+可复用：
+
+- 图层注册表
+- 搜索结果数据
+- BGP / 海缆 / 卫星详情数据
+- 媒体数据
+- 旋转、缩放、选择、高亮等运行时状态
+
+不直接复用：
+
+- 桌面 panel DOM 结构
+- 桌面 panel header / close / collapse 交互
+- 桌面 settings 项里的“显示某 panel”逻辑
+- 桌面媒体面板布局
+
+### 2. 抽屉是唯一主前景层
+
+移动端同一时刻只有一个主前景层：底部抽屉。
+
+抽屉内部切换内容页，而不是多个悬浮层互相覆盖。
+
+### 3. 每个 tab 都是移动端页面，而不是 panel 容器
+
+抽屉中的每一项都应视为一个移动端子页面：
+
+- 有自己的标题
+- 有自己的内容层次
+- 有自己的滚动区域
+- 有自己的主操作
+
+而不是简单挂一个旧面板进去。
+
+### 4. 移动端状态提示不占据屏幕正中
+
+桌面端当前很多通知、状态提示、胶囊消息更适合在屏幕上方居中出现，但移动端不应继续沿用这套布局。
+
+移动端统一改为：
+
+- 通知栏放在右上角安全区
+- 胶囊提示放在右上角堆叠
+- 不遮挡地球中心视野
+- 不与底部抽屉主交互区冲突
+
+## 交互模型
+
+### 默认态
+
+移动端默认只显示：
+
+- 地球主画布
+- 底部半露出的抽屉头部
+
+不再单独显示上箭头按钮。
+
+### 展开态
+
+用户从底边直接上拉抽屉，或点击抽屉头部展开。
+
+展开后显示：
+
+- 当前页面标题
+- tab 导航
+- 当前页面内容
+
+### 收起态
+
+用户下拉抽屉头部收起，或点击背景收起。
+
+## 信息架构
+
+移动端抽屉内的一级页面重定为：
+
+1. 图层
+2. 搜索
+3. 态势
+4. 新闻
+5. TV
+6. 设置
+7. 详情（按需出现，不固定常驻 tab）
+
+其中 `新闻` 和 `TV` 不再共享同一个移动端媒体面板。
+
+## 页面重设计要求
+
+### 图层页
+
+目标：
+
+- 成为移动端最核心的控制页
+- 强调快速开关，不强调桌面 panel 感
+
+内容建议：
+
+- 顶部摘要：当前已启用图层数量
+- 图层列表卡片
+- 每个图层项只保留：
+  - 图标
+  - 中文名
+  - 英文副标题
+  - 开关
+- 去掉桌面式 header / collapse / close 结构
+
+### 搜索页
+
+目标：
+
+- 成为抽屉中的完整搜索页
+- 避免看起来像桌面 modal 被塞进抽屉
+
+内容建议：
+
+- 顶部搜索输入框
+- 搜索提示文案
+- 结果列表
+- 结果项更适合手指点击
+- 结果点击后：
+  - 聚焦地球对象
+  - 自动切换到详情页
+
+### 态势页
+
+目标：
+
+- 合并原来的 `stats + legend` 思路
+- 成为移动端全局态势页
+
+内容建议：
+
+- 顶部核心统计卡
+  - 海缆数量
+  - 登陆点数量
+  - 卫星数量
+  - BGP 事件数量
+- 当前关注层图例
+- BGP 状态摘要
+- 不再出现独立 legend 面板和独立 stats 面板
+
+### 新闻页
+
+目标：
+
+- 从原媒体面板中拆出单独的移动端新闻页
+
+内容建议：
+
+- 当前区域焦点
+- 新闻源数量
+- 新闻卡片列表
+- 卡片内显示标题、来源、时间、区域
+- 外链操作更清晰
+
+### TV 页
+
+目标：
+
+- 从原媒体面板中拆出单独的移动端 TV 页
+
+内容建议：
+
+- 顶部频道选择
+- 直播状态
+- 当前频道说明
+- 视频播放器区域
+- 刷新和外链按钮
+
+不再保留桌面式“新闻/TV tab 共处一个 panel”的结构。
+
+### 设置页
+
+目标：
+
+- 只保留对移动端仍有意义的系统配置
+
+必须移除：
+
+- 图层控制 panel 显示/隐藏
+- 图例 panel 显示/隐藏
+- 全球态势 panel 显示/隐藏
+- 媒体 panel 显示/隐藏
+
+保留项建议：
+
+- 旋转模式
+- 日夜模式
+- 地球默认大小
+- 地形透明度
+- 系统入口
+
+原因：
+
+移动端已经没有这些独立 panel 了，所以继续保留这些开关会制造错误心智。
+
+### 详情页
+
+目标：
+
+- 成为海缆 / BGP / 卫星对象的统一移动端详情页
+
+内容建议：
+
+- 标题区
+- 类型标签
+- 关键属性列表
+- 相关对象摘要
+- 相关图层或态势提示
+
+行为建议：
+
+- 点击对象后自动切入详情页
+- 搜索结果点击后也切入详情页
+
+## 阶段重定义
+
+### 阶段 2：抽屉壳层
+
+目标：
+
+1. 实现底部抽屉基本壳层。
+2. 支持上拉展开、下拉收起、背景点击关闭。
+3. `mobile` 模式下隐藏旧 toolbar。
+4. `mobile` 模式下不再直接显示旧 panel。
+
+完成标准：
+
+1. 手机端只有地球主视图和抽屉。
+2. 抽屉开合稳定。
+
+### 阶段 3：基础页面重做
+
+目标：
+
+1. 重新设计并实现图层页。
+2. 重新设计并实现搜索页。
+3. 重新设计并实现设置页。
+
+完成标准：
+
+1. 这三个页面不再是旧 panel 原样移植。
+2. 设置页已移除 panel 可见性开关。
+
+### 阶段 4：态势与详情重做
+
+目标：
+
+1. 将 stats 和 legend 合并为新的态势页。
+2. 实现统一详情页。
+3. 对象点击与搜索结果点击都可切入详情页。
+
+完成标准：
+
+1. 不再存在移动端独立 legend / stats 面板。
+2. 详情页成为统一对象信息入口。
+
+### 阶段 5：媒体拆分重做
+
+目标：
+
+1. 将原媒体面板拆成两个移动端页面：新闻页、TV 页。
+2. 分别重做这两个页面的布局。
+3. 保留各自必要操作，但不继续共享桌面 panel 结构。
+
+完成标准：
+
+1. 新闻与 TV 各自成为独立移动端页面。
+2. 不再使用桌面媒体 panel 的 tab 结构作为移动端主体。
+
+### 阶段 6：手感与真机修正
+
+目标：
+
+1. 调整抽屉高度、节奏、手势阈值。
+2. 调整 tab 密度与文字层级。
+3. 优化 iPhone / Android 安全区。
+4. 优化抽屉滚动与地球拖拽边界。
+
+完成标准：
+
+1. 抽屉和地球不会抢手势。
+2. 手机端各页面信息层次清晰。
+3. 真机下无遮挡、无死层、无错误交互心智。
+
+## 技术落点调整
+
+### [frontend/public/earth/index.html](/home/ray/dev/linkong/planet/frontend/public/earth/index.html)
+
+职责：
+
+- 只保留移动端抽屉壳层
+- 为各页面提供新的页面容器
+
+不再把旧 panel 作为最终结构直接塞进抽屉。
+
+### [frontend/public/earth/js/controls.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/controls.js)
+
+职责：
+
+- 管理抽屉开合
+- 管理 tab 切换
+- 管理详情页切入
+- 管理 mobile / desktop 分流
+
+### [frontend/public/earth/js/search.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/search.js)
+
+职责：
+
+- 保留搜索能力和结果逻辑
+- 输出给新的移动端搜索页
+
+### [frontend/public/earth/js/info-card.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/info-card.js)
+
+职责：
+
+- 从桌面 info-card 逻辑中提取可复用的数据层
+- 服务新的移动端详情页
+
+### [frontend/public/earth/js/tv.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/tv.js)
+
+职责：
+
+- 为新的 TV 页面提供数据和状态
+- 不再直接主导移动端媒体 panel 壳层
+
+### [frontend/public/earth/js/news.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/news.js)
+
+职责：
+
+- 为新的新闻页面提供列表和区域焦点数据
+
+### CSS
+
+需要新增真正的移动端页面样式，而不是继续在旧 panel class 上堆条件分支：
+
+- 图层页样式
+- 搜索页样式
+- 态势页样式
+- 新闻页样式
+- TV 页样式
+- 设置页样式
+- 详情页样式
+- 移动端右上角通知 / 胶囊提示样式
+
+## 验收标准
+
+1. `mobile` 模式下不再显示旧 toolbar。
+2. `mobile` 模式下不再把旧 panel 直接作为最终 UI。
+3. 图层、搜索、态势、新闻、TV、设置都是重新设计的移动端页面。
+4. 设置页不再包含移动端无意义的 panel 显示/隐藏项。
+5. 新闻与 TV 已拆分为两个移动端页面。
+6. legend / stats 已整合为态势页。
+7. 详情页成为统一对象详情入口。
+8. 移动端通知栏和胶囊提示已统一放到右上角安全区，而不是屏幕正中。
+
+## 结论
+
+本计划进一步明确：
+
+移动端目标不是“把桌面 HUD 放进抽屉”，而是“以抽屉为载体，重做一套适合手机端的信息页面”。
+
+后续开发必须以此为准：
+
+- 复用数据
+- 重做界面
+- 清除桌面遗留心智

docs/plans/earth-news-source-configuration-and-collector-plan.md

@@ -0,0 +1,156 @@
+# Earth News Source Configuration And Collector Plan
+
+## Why
+
+当前 Earth 的“态势新闻”由 [earth_news.py](/home/ray/dev/linkong/planet/backend/app/services/earth_news.py) 直接在请求时抓取 RSS / Google News feed，再按当前地球视角中心区域聚合返回。
+
+这条链已经可用，但存在两个明显限制：
+
+- 新闻源写死在代码里，不能像 TV 直播源一样从后台维护
+- 新闻并未进入统一采集体系，没有采集状态、失败监控、历史数据和后续 AI 复用能力
+
+因此这块更合理的路线不是一步到位重写，而是分阶段推进：
+
+1. 先做“新闻源配置化”
+2. 再做“新闻采集器化”
+
+## Current State
+
+当前实现分布在：
+
+- 新闻接口
+  - [news.py](/home/ray/dev/linkong/planet/backend/app/api/v1/news.py)
+- 实时聚合逻辑
+  - [earth_news.py](/home/ray/dev/linkong/planet/backend/app/services/earth_news.py)
+- 前端消费
+  - [news.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/news.js)
+
+当前新闻源包含：
+
+- `BBC World` RSS
+- `DW Top Stories` RSS
+- 按区域关键词拼出来的 `Google News RSS`
+  - `Global`
+  - `Americas`
+  - `Europe`
+  - `Middle East / Africa`
+  - `Asia Pacific`
+
+当前不是采集器，也不落库，只做内存缓存。
+
+## Phase 1: Source Configuration
+
+### Goal
+
+把 `NEWS_FEED_SOURCES` 从硬编码列表升级成可配置新闻源目录，但继续保留当前“实时聚合”的工作方式。
+
+### Scope
+
+- 为 Earth news 建立独立配置结构
+- 支持后台维护 feed 源
+- 支持启用/禁用、优先级、区域、源类型
+- 保持现有 `/api/v1/news/earth-feed` 输出协议不变
+
+### Proposed Shape
+
+建议配置字段至少包括：
+
+- `id`
+- `name`
+- `region`
+- `feed_url`
+- `homepage_url`
+- `source_type`
+- `priority`
+- `is_enabled`
+- 可选 `query_profile`
+- 可选 `language`
+- 可选 `notes`
+
+### Suggested Storage
+
+优先走系统设置或单独的 news source settings payload，而不是先建复杂新表。
+
+推荐原因：
+
+- 改动小
+- 易上线
+- 和当前 TV settings 维护体验更接近
+- 先解决“写死在代码里”的问题
+
+### Non-goals
+
+这一阶段不做：
+
+- 新闻入库
+- 新闻历史回看
+- 新闻采集任务监控
+- 新闻去重流水线
+
+## Phase 2: News Collectorization
+
+### Goal
+
+把“态势新闻”升级为真正的采集器链路，使其进入采集系统和数据层。
+
+### Scope
+
+- 新增专用 news collector
+- 按配置源定时采集 RSS / feed
+- 做标题/链接级去重
+- 建立统一新闻记录模型
+- 为 Earth、控制台、AI 研判复用同一份新闻数据
+
+### Benefits
+
+- 有采集状态
+- 有失败监控
+- 有历史缓存
+- 可以做时间轴 / 区域新闻基线
+- 可以作为 AI 引用证据
+
+### Required Design Work
+
+需要提前明确：
+
+- 新闻数据模型
+- 去重策略
+- 过期清理策略
+- 区域映射策略
+- 聚合排序策略
+- 新闻与 Earth 当前视角/区域的关联方式
+
+### Candidate Output Model
+
+至少应包含：
+
+- `source_id`
+- `headline`
+- `summary`
+- `url`
+- `publisher`
+- `region`
+- `published_at`
+- `language`
+- `tags`
+- `raw_feed_source`
+- `reference_date`
+
+## Recommended Order
+
+推荐执行顺序：
+
+1. 先完成 Phase 1 配置化
+2. 保持 Earth 继续实时聚合，但改为读取配置源
+3. 等新闻源稳定后，再设计 Phase 2 的 collector / storage / dedupe
+
+## Decision
+
+当前结论：
+
+- TV 直播源：优先采集器化
+- 态势新闻：优先配置化，再采集器化
+
+## Source Note
+
+This plan is newly created for the Planet repo to separate the short-term "configurable source directory" work from the longer-term "collectorized news pipeline" work.

docs/plans/earth-predicted-orbit-plan.md

@@ -0,0 +1,98 @@
+# Earth Predicted Orbit Plan
+
+> Source note: this plan absorbs useful ideas from a sisyphus-created draft formerly stored at `.sisyphus/plans/predicted-orbit.md`.
+
+## Goal
+
+在 Earth 中锁定卫星时，显示“预测轨道”而不是只有历史尾迹：
+
+- 从当前时刻开始
+- 绕地球一圈
+- 当前点最亮
+- 向后沿轨道逐步衰减
+
+## Current State
+
+当前已经有：
+
+- 卫星历史轨迹
+- 锁定卫星
+- 轨道高亮与相关联动
+
+但“预测轨道”仍然不是一套稳定、可验证的单独功能计划。
+
+## Why It Is Valuable
+
+预测轨道可以明显提升：
+
+- 锁定卫星后的空间可读性
+- 轨道类型辨识
+- 演示解释力
+
+相比短历史尾迹，预测轨道更符合用户对“这颗卫星接下来会怎么走”的预期。
+
+## Scope
+
+### Phase 1
+
+- 锁定卫星时显示一整圈预测轨道
+- 解锁时隐藏
+- 不替代现有普通轨迹系统
+
+### Phase 2
+
+- 根据轨道类型调整采样率
+- GEO / MEO / LEO 不同密度
+- 进一步减少 fallback 轨迹的比例
+
+## Implementation Direction
+
+### 1. Orbit period
+
+基于 `meanMotion` 估算轨道周期。
+
+### 2. Predicted samples
+
+以固定采样步长从 `now -> now + period` 推算轨迹点。
+
+### 3. Render object lifecycle
+
+预测轨道应是一个独立渲染对象：
+
+- show
+- update
+- hide
+- dispose
+
+### 4. Visual semantics
+
+预测轨道不应与普通尾迹混淆：
+
+- 更稳定
+- 更完整
+- 透明度沿轨道衰减
+- 当前点附近更亮
+
+## Known Risks
+
+### 1. TLE propagation gaps
+
+部分卫星可能出现 SGP4 计算不足，需要 fallback。
+
+### 2. Multiple orbit lines
+
+必须确保：
+
+- 锁定切换前先清旧轨道
+- 页面隐藏/销毁时清理
+
+### 3. Performance
+
+GEO 轨道点数高，采样率需要按轨道类型分层。
+
+## Acceptance
+
+1. 锁定单颗卫星时只显示一条预测轨道
+2. 解锁后轨道立即清除
+3. 不同轨道类型下点数可控
+4. 页面切换回来不会闪出旧轨道残留

docs/plans/earth-real-terrain-plan.md

@@ -0,0 +1,472 @@
+# Earth Real Terrain Plan
+
+## Goal
+
+将 Earth 页当前的“程序噪声假地形”替换成基于真实 DEM 的可用地形层，使 `地形 terrain` 开关真正显示全球海拔起伏，而不是占位效果。
+
+当前占位实现位于：
+
+- [earth.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/earth.js)
+
+具体问题：
+
+- `createTerrain()` 直接对球体顶点应用 `simplex noise`
+- 没有真实海拔数据来源
+- 没有分辨率分层
+- 没有和当前相机/视角配套的性能控制
+
+## Constraints
+
+本计划必须贴合当前 Earth 架构，而不是引入一套全新的地形引擎：
+
+- 地球主体仍然是一个 Three.js sphere
+- 海缆、登陆点、卫星、BGP 都已经建立在当前球体坐标系之上
+- 不能为了地形把整页改成 Cesium/MapLibre Globe 之类的全栈替换
+- 第一阶段优先做“真实可用”，不是一步到位做摄影测量级地形
+
+## Recommended Data Source
+
+### Primary recommendation
+
+使用公开的 Terrarium 编码高程瓦片作为浏览器端高度来源，第一阶段优先接入：
+
+- Mapzen/AWS `Terrarium` elevation tiles  
+  参考：[Mapzen terrain tile format / Terrarium](https://www.mapzen.com/blog/terrain-tile-service/)
+
+原因：
+
+- 已经是全球瓦片化高程
+- 浏览器端按 tile 请求，最适合当前 Earth 这种在线 globe
+- 编码简单稳定：
+  - `heightMeters = (R * 256 + G + B / 256) - 32768`
+- 不需要我们先离线拼整球 DEM
+
+### Data quality upgrade path
+
+如果后面第一阶段效果确认可用，再逐步升级到底层源：
+
+- Copernicus DEM GLO-30  
+  参考：[Copernicus DEM docs](https://documentation.dataspace.copernicus.eu/APIs/SentinelHub/Data/DEM.html)
+- 或用 Copernicus / SRTM / ASTER 等离线切成我们自己的 terrain tiles
+
+这条升级路径适合第二阶段，不建议一开始就直接自建全球瓦片服务。
+
+## Why Not Replace the Engine
+
+不建议为了地形直接切到 Cesium terrain / quantized mesh 引擎，原因：
+
+- 现有 Earth 业务对象都依附当前球面坐标
+- 切引擎会同时波及：
+  - 海缆绘制
+  - 卫星/轨迹
+  - BGP 标记
+  - HUD 与交互
+- 这是“重做一页”，不是“给地形层接真实数据”
+
+所以推荐路线是：
+
+- 保持当前 sphere globe
+- 为 sphere 增加真实高度位移层
+
+## Implementation Strategy
+
+分三期推进。
+
+### Phase 1 — Global Heightmap Terrain Overlay
+
+目标：
+
+- 地形层切换后显示真实海拔起伏
+- 全球范围可用
+- 性能可控
+
+做法：
+
+1. 新增 terrain 数据模块
+
+建议文件：
+
+- `frontend/public/earth/js/terrain.js`
+
+职责：
+
+- 选择 DEM zoom level
+- 请求 Terrarium tiles
+- 解码 tile 高程
+- 将高程重采样到当前地形球体网格
+
+2. 替换 `createTerrain()`
+
+当前：
+
+- 在 [earth.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/earth.js) 中同步生成噪声地形
+
+调整后：
+
+- `createTerrain()` 只负责创建 terrain mesh 骨架
+- 真正的顶点位移由 terrain 模块异步注入
+
+3. 第一阶段采用“整球低分辨率位移”
+
+不要一上来做动态 patch stitching。第一阶段更稳的办法是：
+
+- 保留一张全球 terrain sphere
+- 使用较低分辨率几何
+  - 例如 `SphereGeometry(radius, 192, 192)` 或 `256/256`
+- 运行时按一个固定地形 zoom（如 `z=4` 或 `z=5`）抓取覆盖全球的 Terrarium tiles
+- 将 tile 解码后重投影到经纬度采样网格
+- 将每个球面顶点按真实高度抬升
+
+这样第一阶段就能做到：
+
+- 有真实地形
+- 不需要复杂的局部 LOD
+- 不会让现有球体对象体系爆炸
+
+### Phase 2 — View-Aware Refinement
+
+目标：
+
+- 正面可见区域更精细
+- 背面与远处维持低成本
+
+做法：
+
+- 引入“基础全球地形 + 当前视角高分局部补丁”
+- 正面区域额外抓更高 zoom 的高程 tile
+- 只替换局部顶点位移或局部 overlay mesh
+
+这一阶段适合在第一阶段稳定后做。
+
+### Phase 3 — Normals / Shading / Terrain UX
+
+目标：
+
+- 地形不仅有起伏，还更好看、更可读
+
+包括：
+
+- 根据高度生成更合理的 normals
+- 调整 terrain material，使山脉/高原更易读
+- 可选加入：
+  - hillshade
+  - contour lines
+- snowline / bathymetry tint
+
+## Calibration Overlay Before More Terrain Tuning
+
+在当前项目里，terrain 看起来“不像真地形”，不一定只是 DEM 或 exaggeration 不够，也可能是因为缺少稳定参照物。
+
+没有清晰的海岸线、国界线和地表分层时，人眼很难判断：
+
+- 山脉是不是在应该高的地方高
+- terrain 是否真的贴在正确的大陆位置上
+- 地球纹理、本初子午线、terrain 采样之间是否存在偏移
+
+这里要明确区分两件事：
+
+- 国界线不会修好错误的 terrain
+- 但海岸线 / 国界线会让我们更容易判断 terrain 有没有贴准
+
+所以在继续盲调 terrain 参数之前，建议先插入一个“校准参照层”阶段。
+
+### Recommended order for the calibration layer
+
+1. 海岸线
+2. 国界线
+3. 再继续调 terrain
+
+原因：
+
+- 海岸线比国界线更基础，也更接近真实地表边界
+- 判断 terrain 是否贴准，最重要的是大陆边缘和山脉/海岸关系
+- 国界线更多是政治边界，只能作为辅助参照
+
+如果只加国界线，不加海岸线，效果仍然可能会怪，因为：
+
+- 很多国界线本来就是人为直线
+- 它们并不总是跟真实地形走
+
+### Suggested layer order during debugging
+
+建议调试期临时把地球层次明确成：
+
+1. base earth texture
+2. coastline / borders overlay
+3. terrain relief
+4. cables / landing points / bgp / satellites
+
+这样会比现在更容易判断：
+
+- 山脉是否位于正确区域
+- terrain 是否和地表对齐
+- 国界/海岸是否漂移
+
+### Suggested data source for the calibration overlay
+
+优先用 `Natural Earth` 的轻量全球矢量数据：
+
+- 海岸线（coastline）
+- Admin 0 国界线（country borders）
+
+优点：
+
+- 全球一致
+- 轻量
+- 很适合当前 Three.js globe 做 overlay
+
+### Recommended execution path
+
+#### Phase A — Add reference overlays
+
+先加两层可开关的参考线：
+
+- 海岸线
+- 国界线
+
+这两层的目标不是最终美术表现，而是调试 / 校准。
+
+#### Phase B — Recalibrate terrain against coastline
+
+有了海岸线以后，再重新看 terrain：
+
+- terrain 是否和大陆边缘错位
+- 地球纹理、本初子午线、terrain 采样之间是否有固定偏移
+
+#### Phase C — Decide whether to keep the current terrain path
+
+这时再决定后面的路线：
+
+- 如果发现真实高程整体是对的，只是缺少 shading / readability  
+  继续保留当前 DEM + terrain overlay 路线
+- 如果发现整球采样投影、本初子午线或 overlay 关系本身就很别扭  
+  再考虑重做 terrain pipeline
+
+### Practical recommendation
+
+当前阶段不建议“从头开始重做 terrain”。
+
+更稳的策略是：
+
+- 暂停继续盲调 terrain 参数
+- 先补海岸线 / 国界线作为校准参照层
+- 再基于参照层判断 terrain 是“参数没调好”，还是“整条实现路径有偏移”
+
+## Recommended Geometry Model
+
+### First usable model
+
+保留一层独立 terrain sphere：
+
+- base earth sphere：贴纹理、昼夜、海洋
+- terrain sphere：略高于地球半径，真实高程位移
+
+建议：
+
+- `terrainBaseRadius = CONFIG.earthRadius + 0.2`
+- 高度缩放使用真实米制换算，再乘一个可调 exaggeration
+
+示例关系：
+
+- `heightWorld = (elevationMeters / 6371000) * CONFIG.earthRadius * exaggeration`
+
+建议第一阶段 `exaggeration = 1.3 ~ 1.8`
+
+因为完全真实比例在全球球体上会太平，看不出来。
+
+## Tile Decoding Plan
+
+### Terrarium decode
+
+对于每个高程 tile 像素：
+
+```text
+heightMeters = (R * 256 + G + B / 256) - 32768
+```
+
+### Sampling path
+
+对于 terrain mesh 上每个顶点：
+
+1. 将顶点方向转成经纬度
+2. 将经纬度映射到 Web Mercator tile 坐标
+3. 找到对应的 tile 和像素
+4. 解码高程
+5. 将顶点沿法线方向抬升
+
+### Needed helpers
+
+建议新增：
+
+- `latLonToTileXY(lat, lon, z)`
+- `tilePixelFromLatLon(lat, lon, z, tileSize)`
+- `decodeTerrariumHeight(r, g, b)`
+
+## Caching Strategy
+
+为了不让地形开关每次重开都重新抓全量 tile：
+
+- terrain tile 按 `z/x/y` 存到内存缓存
+- terrain mesh 结果也缓存一份
+- 当用户关闭/开启 terrain：
+  - 直接复用已有位移结果
+
+建议：
+
+- `Map<string, Float32Array | ImageBitmap>`
+
+## Material Strategy
+
+第一阶段不要复杂化。
+
+建议 terrain material：
+
+- 半透明低饱和地形色
+- 比 base earth 稍亮或稍偏冷
+- 保留当前 HUD 风格下的可读性
+
+第一阶段不需要：
+
+- 真实土地覆被纹理
+- 独立卫星影像贴 terrain
+
+因为那会和现有地球纹理、云层、昼夜 shader 打架。
+
+## Integration Points
+
+### Files to change
+
+- [earth.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/earth.js)
+  - 重写 `createTerrain()`
+  - 删除 simplex noise 占位逻辑
+- [main.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/main.js)
+  - 初始化 terrain 数据加载
+  - 控制 terrain readiness / loading message
+- [controls.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/controls.js)
+  - `toggleTerrain` 逻辑保持，但应能区分：
+    - mesh 已就绪
+    - 正在加载
+- [constants.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/constants.js)
+  - 新增 `TERRAIN_CONFIG`
+- 新文件：
+  - `frontend/public/earth/js/terrain.js`
+
+### Suggested new config
+
+建议新增：
+
+```js
+export const TERRAIN_CONFIG = {
+  enabled: true,
+  tileSize: 256,
+  baseZoom: 4,
+  baseRadiusOffset: 0.2,
+  exaggeration: 1.5,
+  opacity: 0.55,
+  color: 0x6c876f,
+  maxConcurrentRequests: 8,
+  cacheEnabled: true,
+};
+```
+
+## Loading UX
+
+地形第一次开启时，不能像现在一样瞬时切换。
+
+建议：
+
+- 如果地形数据尚未准备：
+  - 顶部状态条显示：`正在加载真实地形数据...`
+- 完成后：
+  - `真实地形已就绪`
+
+如果加载失败：
+
+- 保留 base earth
+- 显示轻量错误提示
+- 不要让 terrain 开关卡死在“开”状态
+
+## Risks
+
+### 1. Global tile count too high
+
+即使 `z=5` 全球 tile 数也不少。
+
+缓解：
+
+- 第一阶段限定低 zoom
+- 并发上限
+- 缓存
+
+### 2. Mesh resolution too low
+
+如果球面分段太低，山脉会被抹平。
+
+缓解：
+
+- 第一阶段先选一个中等分辨率
+- 用 exaggeration 保证可见性
+
+### 3. Existing overlays may z-fight with terrain
+
+海缆、登陆点、BGP、卫星相关对象都假设地球半径固定。
+
+缓解：
+
+- terrain sphere 单独作为 overlay
+- overlay 保持略低或略高的固定 offset
+- 必要时局部调整 landing point / cable altitude offset
+
+### 4. Mercator sampling distortion near poles
+
+Web Mercator 在高纬会有失真。
+
+缓解：
+
+- 第一阶段接受
+- 后续若需要更严格极区质量，再上 geodetic reprojection pipeline
+
+## Acceptance Criteria
+
+第一阶段完成后，应满足：
+
+1. `地形 terrain` 开关开启时，地表起伏明显不再是随机噪声
+2. 喜马拉雅、安第斯、落基山、东非高原等全球大尺度地形可辨认
+3. 关闭/重新开启 terrain 不重复全量请求
+4. 不破坏：
+   - 海缆
+   - 卫星
+   - BGP
+   - 地球昼夜
+   - 天球层
+
+## Suggested Execution Order
+
+1. 引入 `TERRAIN_CONFIG`
+2. 新建 `terrain.js`
+3. 实现 Terrarium tile 请求与 decode
+4. 用低 zoom 全球 tile 构建真实 terrain sphere
+5. 接管 `toggleTerrain()`
+6. 调整 terrain material 和高度 exaggeration
+7. 做缓存
+8. 再考虑第二阶段局部高分 refinement
+
+## Source References
+
+- Mapzen Terrarium / AWS terrain tiles  
+  [Mapzen Terrain Tile Service](https://www.mapzen.com/blog/terrain-tile-service/)
+- Terrarium tile experiments / format background  
+  [mapzen/terrarium](https://github.com/mapzen/terrarium)
+- Copernicus DEM overview  
+  [Copernicus DEM docs](https://documentation.dataspace.copernicus.eu/APIs/SentinelHub/Data/DEM.html)
+
+## Recommendation Summary
+
+如果现在就要开始做，我建议直接按这条路线开工：
+
+- 第一阶段接入 Terrarium 全球高程 tile
+- 替换掉当前 simplex 假地形
+- 先做一层真实可见的全球 terrain overlay
+- 等第一阶段稳定，再做视角高分 refinement
+
+这是对当前项目风险最低、最贴合现有 Earth 架构的一条路。

docs/plans/earth-renderer-architecture-separation-plan.md

@@ -0,0 +1,111 @@
+# Earth Renderer / Logic Separation Plan
+
+> Source note: this plan absorbs useful ideas from a sisyphus-created draft formerly stored at `.sisyphus/plans/earth-architecture-refactor.md`.
+
+## Goal
+
+将 Earth 前端继续往“逻辑层 / 状态层 / 渲染层”分离推进，降低后续这几类工作的耦合成本：
+
+- Three.js 渲染重构
+- 部分图层替换实现
+- 未来 UE / Cesium 客户端迁移
+- Earth 行为逻辑复用
+
+## Why This Matters
+
+当前 Earth 已经有一些良好分层，例如：
+
+- 图层显隐入口
+- Cable state 枚举与状态 map
+- 交互逻辑与实际视觉效果的部分分离
+
+但还没有形成一套更明确的统一规则。现在的风险是：
+
+- 同一类对象的 hover / locked / hidden / loading 语义不一致
+- 状态和渲染更新散落在多个模块
+- 后续再加新图层时容易复制旧逻辑
+
+## Target Architecture
+
+Earth 对每类对象都尽量拆成三层：
+
+1. `state layer`
+   - 保存对象状态
+   - 例如：`normal / hovered / locked / hidden / loading`
+
+2. `logic layer`
+   - 处理点击、悬停、锁定、过滤、显隐切换
+   - 不直接关心 Three.js 具体材质怎么改
+
+3. `renderer layer`
+   - 根据状态更新 Three.js / HUD 外观
+   - 是最容易针对不同渲染引擎替换的一层
+
+## Current Good Signals
+
+当前已经接近这条方向的地方：
+
+- cable 状态管理
+- 部分 landing point 状态同步
+- layer button 的统一状态入口
+- tooltip / legend / info-card 开始朝状态驱动靠拢
+
+## Next Steps
+
+### 1. Standardize object state enums
+
+优先为这些对象建立更稳定的状态语义：
+
+- cables
+- satellites
+- landing points
+- BGP markers
+- media / news 面板入口按钮
+
+### 2. Unify state-to-visual adapters
+
+为各模块建立更清晰的渲染适配函数，例如：
+
+- `applyCableVisualState()`
+- `applySatelliteVisualState()`
+- `applyBGPVisualState()`
+
+要求：
+
+- 逻辑层只改状态
+- 视觉层负责把状态映射到材质、透明度、发光、尺寸、文字
+
+### 3. Separate Earth UI state from render state
+
+HUD / 面板 / 图层按钮状态也需要和渲染状态分离：
+
+- `loading`
+- `active`
+- `locked`
+- `hidden`
+- `error`
+
+不要再让 UI 通过“猜渲染结果”推导业务状态。
+
+### 4. Prepare migration-safe boundaries
+
+后续如果做 UE / Cesium 客户端，尽量保留：
+
+- 状态枚举
+- 交互规则
+- 数据层接口
+
+只替换：
+
+- Three.js 具体渲染实现
+- HUD 展示实现
+
+## Practical Rule
+
+后续 Earth 新功能开发时，优先问三个问题：
+
+1. 这个状态由谁持有？
+2. 这个交互逻辑在哪一层处理？
+3. 这个视觉变化是否能在不改逻辑的情况下单独替换？
+
+如果答不上来，就说明还在把状态、逻辑、渲染揉在一起。

docs/plans/earth-webgl-instancing-satellites-plan.md

@@ -0,0 +1,82 @@
+# Earth WebGL Instancing Satellites Plan
+
+> Source note: this plan absorbs useful ideas from a sisyphus-created draft formerly stored at `.sisyphus/plans/webgl-instancing-satellites.md`.
+
+## Goal
+
+把 Earth 卫星渲染从当前方案继续推进到更适合高数量卫星的 instancing 方向，目标是：
+
+- 支持更多卫星
+- 降低渲染压力
+- 仍然保留当前数据层和交互层
+
+## Why It Matters
+
+当前卫星系统已经具备：
+
+- 数据加载
+- 轨迹
+- 选择/锁定
+- 图例
+- 相关区域联动
+
+但当卫星数量持续增加时，渲染层会越来越接近瓶颈。
+
+## Recommended Direction
+
+优先调研并原型验证：
+
+- `InstancedBufferGeometry + custom shader`
+
+而不是一开始就推倒重写成 raw WebGL。
+
+原因：
+
+- 仍能保留 Three.js 主架构
+- 更容易渐进迁移
+- 比继续堆普通点渲染更有上限
+
+## What Should Stay
+
+尽量保留这些层：
+
+- 卫星数据获取
+- 位置计算
+- 锁定/悬停逻辑
+- legend / info-card / 相关联动
+
+主要替换的是：
+
+- 卫星点渲染实现
+- 颜色/大小等实例属性更新方式
+
+## Phases
+
+### Phase 1: Prototype
+
+- 用 instancing 做最小原型
+- 先只渲染卫星点
+- 不碰轨迹系统
+
+### Phase 2: Integrate
+
+- 接入当前 `satellites.js` 数据层
+- 保留当前选择和高亮语义
+
+### Phase 3: Tune
+
+- 调整可视大小
+- 调整选中高亮方式
+- 评估是否需要分层 LOD
+
+## Risks
+
+1. 透明度排序更复杂
+2. Shader 调试成本更高
+3. 选中态和 hover 态不能简单复用旧材质逻辑
+
+## Acceptance
+
+1. 在更高卫星数量下保持可接受帧率
+2. 不破坏现有锁定/高亮语义
+3. 图例、信息卡、相关卫星联动仍然成立

docs/plans/enterprise-logging-system-plan.md

@@ -0,0 +1,793 @@
+# Planet 企业级日志系统实施计划
+
+## Goal
+
+把 Planet 当前“能看一点运行输出”的日志能力，升级为一套真正可用、可定位、可纠错、可追责、可演进的企业级日志系统。
+
+这里的“企业级”不是指一上来就接入很重的外部平台，而是指这套系统需要同时满足下面五件事：
+
+1. 排障可用
+2. 历史可查
+3. 业务可解释
+4. 权限操作可追责
+5. 出错后能够反向定位到请求、任务、模块和操作者
+
+最终目标不是“把更多 stdout 放到日志页里”，而是建立一套统一的日志契约与落地链路：
+
+- 统一日志字段
+- 统一事件命名
+- 统一采集入口
+- 统一查询视图
+- 清晰的实时日志、持久化事件、审计日志分层
+
+## Why
+
+当前仓库已经有一些日志基础，但离真正可用的日志系统还有明显距离。
+
+已有基础：
+
+- 后端运行日志可通过 `/tmp/planet_backend.log` 查看
+- 前端开发服务日志可通过 `/tmp/planet_frontend.log` 查看
+- AI Provider 可从 Docker 容器读取日志
+- Earth 浏览器端关键日志可上报到后端并进入 Redis 缓冲
+- 已有 `system_logs` / `audit_logs` 持久化能力
+- 管理台已有“系统日志”页面，支持来源、级别、日期、搜索
+
+当前缺口：
+
+- 后端日志仍以 `uvicorn` / 文本输出为主，不是统一结构化事件流
+- 不同模块的日志格式不一致，很多地方只有 message，没有 event 语义
+- 还没有统一的后端 logger 封装与字段注入机制
+- 前端虽然能上报错误，但还没有统一 logger API 和统一事件词汇
+- Earth 与管理台之间的错误事件还没有形成可串联的事件链路
+- 历史持久化还偏点状，很多高价值失败并没有系统性落库
+- 系统日志页当前更像“运行输出查看器”，不是“多层日志查询台”
+- 审计日志与运行日志尚未形成明确的产品级联动
+
+所以当前真正的问题不是“有没有日志页”，而是：
+
+**当前系统能看见输出，但还不能稳定回答“发生了什么、影响了谁、在哪条链路上坏了、是否已修复、是谁触发的”。**
+
+## Current State
+
+截至 2026-04-23，当前代码中的日志相关能力大致如下。
+
+### 1. 日志来源
+
+当前系统日志页主要读取以下来源：
+
+- `backend`
+  读取 `/tmp/planet_backend.log`
+- `frontend`
+  读取 `/tmp/planet_frontend.log`
+- `ai-provider`
+  读取 Docker 容器日志
+- `earth-client`
+  读取 Redis 缓冲的浏览器端日志
+
+这些来源定义在：
+
+- [backend/app/services/system_logs.py](/home/ray/dev/linkong/planet/backend/app/services/system_logs.py)
+
+### 2. 当前日志读取模型
+
+当前 `read_log_snapshot()` 的职责是：
+
+- 读取某个来源的最近若干行
+- 解析基础级别与时间
+- 按级别、日期、搜索进行过滤
+- 返回用于日志页展示的快照
+
+这个模型适合“运维查看器”，但不适合企业级日志系统，原因是：
+
+- 读取基于文本尾部扫描，不是基于事件模型
+- 不同来源的结构粒度完全不同
+- 过滤依赖文本解析，准确率有限
+- 没有请求、任务、用户、资源、动作等核心关联字段
+
+### 3. 已有持久化能力
+
+当前已经存在两个持久化入口：
+
+- `record_system_log(...)`
+- `record_audit_log(...)`
+
+位置：
+
+- [backend/app/services/persistent_logs.py](/home/ray/dev/linkong/planet/backend/app/services/persistent_logs.py)
+
+这说明系统并不是从 0 开始，但也说明当前最大的问题是：
+
+**持久化能力存在，但没有成为统一默认路径。**
+
+### 4. 已有 request_id 基础
+
+当前系统已具备 `request_id` 相关基础，部分持久化能力也会尝试写入 `request_id`。
+
+这为后续做：
+
+- 请求链路排障
+- 前后端关联查询
+- 任务执行追踪
+
+提供了很好的基础。
+
+### 5. 当前日志页定位
+
+当前日志页已经具备：
+
+- 来源切换
+- 级别筛选
+- 日期筛选
+- 搜索
+- 文本控制台视图
+
+但它仍然是“单层视图”：
+
+- 上面是筛选器
+- 下面是一块文本控制台
+
+它还不是：
+
+- 运行日志 + 事件日志 + 审计日志 的统一入口
+- 也没有事件详情、关联跳转、纠错建议、链路追踪能力
+
+## Core Principles
+
+这套日志系统后续必须遵循下面几个原则。
+
+### 1. 分层，而不是混存
+
+日志必须拆成三层：
+
+1. 运行日志
+2. 持久化事件日志
+3. 审计日志
+
+它们的用途不同，绝不能继续混成一个概念。
+
+#### 运行日志
+
+用于：
+
+- 实时排障
+- 观察服务运行状态
+- 看 stdout / stderr / exception / collector 输出
+
+特点：
+
+- 数据量大
+- 时效性强
+- 保留周期短
+- 不要求每条都落库
+
+#### 持久化事件日志
+
+用于：
+
+- 记录高价值错误
+- 记录关键业务失败
+- 支撑历史追溯
+- 支撑趋势分析
+
+特点：
+
+- 只持久化有价值事件
+- 必须结构化
+- 必须有统一 event 命名
+
+#### 审计日志
+
+用于：
+
+- 留痕
+- 追责
+- 还原高权限操作
+
+特点：
+
+- 必须单独建模
+- 不与普通运行日志混用
+
+### 2. 结构化优先
+
+正式日志必须可拆字段，不能长期依赖自由文本。
+
+最低要求至少能拿到：
+
+- `timestamp`
+- `level`
+- `service`
+- `module`
+- `event`
+- `message`
+- `request_id`
+- `trace_id`
+- `user_id` / `actor`
+- `context`
+
+### 3. 事件命名优先于 message 命名
+
+人看的 message 可以变化，但机器查询和跨模块关联必须依赖稳定事件名。
+
+例如：
+
+- `collector.run.started`
+- `collector.run.completed`
+- `collector.run.failed`
+- `earth.layer.load_failed`
+- `earth.cruise.route_build_failed`
+- `system.restart_task.failed`
+- `auth.websocket.invalid_token`
+
+### 4. 查询链路必须可串联
+
+企业级日志系统的核心不是“有很多日志”，而是“能串起来”。
+
+最终一条高价值事件，至少要能回链到下面任意几类对象：
+
+- 某个请求
+- 某个任务
+- 某个用户
+- 某个数据源
+- 某个 Earth 模块
+- 某个管理动作
+
+### 5. 默认脱敏
+
+日志体系必须明确禁止记录：
+
+- token
+- password
+- Authorization header
+- cookie
+- session
+- 明文敏感个人信息
+
+并且需要有统一脱敏器，而不是靠调用者自觉。
+
+### 6. “可纠错”不是一句口号
+
+这里的“可纠错”至少包含三层：
+
+1. 日志字段足够解释错误，方便人排查
+2. 系统能识别常见错误模式并给出纠偏建议
+3. 关键错误支持闭环动作，例如重试、重建索引、重新触发采集、跳转到对应对象
+
+也就是说，这套日志系统最终不只是“告诉你出错了”，而要尽量接近“告诉你为什么出错、怎么修、去哪修”。
+
+## Non-Goals
+
+第一阶段不追求：
+
+- 全量接入 ELK / Loki / Datadog / OpenTelemetry 全家桶
+- 做分布式 trace 全链路可视化大屏
+- 把所有历史日志都迁进数据库
+- 先做特别复杂的规则引擎
+
+第一阶段追求的是：
+
+- 在当前仓库和当前部署方式下，先把基础日志体系做正确
+- 再为后续平台化接入预留好接口
+
+## Target Architecture
+
+推荐目标架构如下。
+
+### Layer 1: Runtime Logs
+
+职责：
+
+- 承载后端、前端开发服务、容器输出、浏览器端缓冲事件
+- 提供最近窗口内的实时查看能力
+
+来源：
+
+- 文件
+- Docker
+- Redis 缓冲
+- 后续可扩展到 stdout collector
+
+接口：
+
+- `GET /api/v1/system/logs/sources`
+- `GET /api/v1/system/logs/{source_id}`
+
+这层继续保留，但需要做结构化增强和来源补强。
+
+### Layer 2: Persistent System Events
+
+职责：
+
+- 只存高价值事件
+- 供历史追溯、事件列表、趋势和纠错使用
+
+数据来源：
+
+- 后端关键异常
+- 浏览器端关键失败
+- 采集器/调度器关键失败
+- 业务关键告警与降级事件
+
+接口建议：
+
+- `GET /api/v1/system/events`
+- `GET /api/v1/system/events/{id}`
+- `POST /api/v1/system/events/{id}/actions/...`（后续）
+
+### Layer 3: Audit Logs
+
+职责：
+
+- 留痕高权限操作
+- 记录操作者、对象、结果、请求号
+
+接口建议：
+
+- `GET /api/v1/system/audit-logs`
+
+### Layer 4: Error Intelligence / Triage
+
+职责：
+
+- 对高频错误做归类
+- 对已知错误给出解释与建议动作
+- 对相同错误进行 fingerprint 聚合
+
+这是“可纠错”能力的关键层。
+
+建议字段：
+
+- `fingerprint`
+- `root_cause_type`
+- `known_fix_hint`
+- `runbook_url`
+- `related_resource_type`
+- `related_resource_id`
+
+## Canonical Event Model
+
+推荐统一事件字段模型如下。
+
+### Runtime Log Record
+
+```json
+{
+  "timestamp": "2026-04-23T10:15:30Z",
+  "level": "error",
+  "service": "backend",
+  "module": "app.services.scheduler",
+  "event": "collector.run.failed",
+  "message": "Collector bgp_news failed",
+  "request_id": "req_xxx",
+  "trace_id": "trace_xxx",
+  "user_id": null,
+  "actor": null,
+  "resource_type": "collector",
+  "resource_id": "bgp_news",
+  "context": {
+    "datasource_id": 12,
+    "exception_type": "TimeoutError"
+  }
+}
+```
+
+### Persistent System Event
+
+```json
+{
+  "id": 1024,
+  "event": "earth.layer.load_failed",
+  "level": "error",
+  "source": "earth-client",
+  "service": "earth",
+  "module": "cables",
+  "message": "Failed to load cable layer",
+  "fingerprint": "earth.layer.load_failed:cables:network_timeout",
+  "request_id": "req_xxx",
+  "trace_id": null,
+  "user_id": 1,
+  "resource_type": "earth_layer",
+  "resource_id": "cables",
+  "category": "visualization",
+  "status": "open",
+  "context": {
+    "url": "/api/v1/visualization/geo/cables"
+  },
+  "created_at": "2026-04-23T10:15:30Z"
+}
+```
+
+### Audit Log
+
+```json
+{
+  "id": 88,
+  "action": "system.restart_task.requested",
+  "actor_id": 1,
+  "actor_name": "root",
+  "target_type": "restart_task",
+  "target_id": "restart_20260423_xxx",
+  "result": "success",
+  "request_id": "req_xxx",
+  "ip": "127.0.0.1",
+  "details": {
+    "action": "restart_backend"
+  },
+  "created_at": "2026-04-23T10:15:30Z"
+}
+```
+
+## Implementation Plan
+
+## Phase 0: Logging Inventory And Naming Freeze
+
+目标：
+
+- 先统一“记录什么”和“怎么命名”，避免后面越做越乱
+
+工作项：
+
+- 盘点当前所有 `logging.getLogger` 使用点
+- 盘点裸 `print`
+- 盘点 `record_system_log` / `record_audit_log` 已落点位
+- 建立统一事件命名表
+- 定义 service / module / category / resource 字段枚举
+- 输出日志字段白名单和脱敏规范
+
+完成标准：
+
+- 有一份稳定的事件命名清单
+- 有一份字段规范清单
+- 后续新增日志不再“临时起名”
+
+## Phase 1: Backend Structured Logging Foundation
+
+目标：
+
+- 把后端从“散落 logging + 文本输出”升级成“统一结构化 logger”
+
+工作项：
+
+- 新增统一后端 logger helper，例如 `app/core/logging.py`
+- 自动注入：
+  - `service`
+  - `module`
+  - `request_id`
+  - `trace_id`
+- 增加统一脱敏 filter
+- 把关键模块先切到统一 logger：
+  - API 层
+  - scheduler
+  - collectors
+  - websocket
+  - visualization
+  - system control
+- 约束：
+  - 正式路径禁止裸 `print`
+  - 正式异常优先 `logger.exception(..., extra={...})`
+
+完成标准：
+
+- 后端关键模块都有稳定 `event`
+- request 日志和异常日志能挂上 `request_id`
+- 不再依赖只看 `uvicorn` 原生文本输出来定位问题
+
+## Phase 2: Persistent Event Layer
+
+目标：
+
+- 把“值得长期保留的错误和关键事件”系统性落库
+
+工作项：
+
+- 重新定义 `record_system_log()` 的使用边界
+- 明确哪些事件必须持久化：
+  - API 关键失败
+  - 调度器失败
+  - 采集器失败
+  - Earth 客户端关键错误
+  - 数据源不可用
+  - 业务降级与恢复
+- 补齐字段：
+  - `event`
+  - `resource_type`
+  - `resource_id`
+  - `category`
+  - `fingerprint`
+  - `status`
+- 增加高频错误去重/聚合策略
+
+完成标准：
+
+- 高价值错误不再只存在于运行日志里
+- 能查询最近一周/一月的关键失败事件
+- 相同错误具备聚合基础
+
+## Phase 3: Frontend And Earth Unified Logger
+
+目标：
+
+- 把前端从“点状 error 上报”升级成统一前端事件流
+
+工作项：
+
+- 在前端新增统一 logger API
+- 统一方法：
+  - `debug`
+  - `info`
+  - `warn`
+  - `error`
+- 统一字段：
+  - `page`
+  - `module`
+  - `event`
+  - `message`
+  - `url`
+  - `user_agent`
+  - `context`
+- Earth 模块优先接入：
+  - layer load failed
+  - cruise build failed
+  - popup render failed
+  - connector render failed
+  - websocket dropped
+- 管理台优先接入：
+  - settings save failed
+  - datasource toggle failed
+  - restart task submit failed
+
+完成标准：
+
+- 前端日志事件名与后端可对齐
+- Earth 和管理台关键失败不再只停留在 console
+- 浏览器端关键问题能进入统一系统日志/事件层
+
+## Phase 4: Audit Logging Completion
+
+目标：
+
+- 把管理员与高权限操作真正做成企业级审计
+
+工作项：
+
+- 扩大审计覆盖面：
+  - 系统重启
+  - 数据源启停
+  - 调度规则变更
+  - 配置变更
+  - 人工触发采集
+  - 删除/修改关键配置
+- 增加字段：
+  - actor
+  - target
+  - before / after
+  - request_id
+  - IP
+- 审计页支持：
+  - 动作筛选
+  - 操作者筛选
+  - 时间筛选
+  - 目标对象筛选
+
+完成标准：
+
+- 所有高权限操作都能追到人、时间、对象、结果
+
+## Phase 5: Log Console To Enterprise Observability UI
+
+目标：
+
+- 把当前“系统日志”页升级为真正的多层日志工作台
+
+工作项：
+
+- 将页面拆为三个主视图：
+  1. 运行日志
+  2. 关键事件
+  3. 审计日志
+- 运行日志视图：
+  - 保留大控制台
+  - 支持来源、级别、日期、搜索
+- 关键事件视图：
+  - 列表化展示高价值事件
+  - 支持聚合、状态、指纹、对象筛选
+- 审计视图：
+  - 列表化展示管理员动作
+- 增加详情抽屉：
+  - 原始 message
+  - context
+  - request_id
+  - related resource
+  - recommended action
+
+完成标准：
+
+- 日志页不再只是“终端文本窗口”
+- 运维排障、历史追溯、审计留痕三者分层清晰
+
+## Phase 6: Corrective Intelligence
+
+目标：
+
+- 让系统从“能看日志”进化到“能辅助修错”
+
+工作项：
+
+- 引入错误 fingerprint
+- 对已知错误配置：
+  - 根因类型
+  - 修复建议
+  - runbook 链接
+  - 推荐动作
+- 支持常见纠错动作：
+  - 重试采集任务
+  - 重载配置
+  - 跳转到对应模块/资源
+  - 打开相关日志过滤视图
+- 高频错误支持聚合与静默窗口
+
+完成标准：
+
+- 已知错误能给出明确建议
+- 运维不需要每次都从零猜
+
+## Recommended Module Changes
+
+### Backend
+
+建议新增/增强的模块：
+
+- `backend/app/core/logging.py`
+  - 统一 logger 封装
+  - formatter
+  - filter
+  - request/trace 注入
+- `backend/app/services/persistent_logs.py`
+  - 扩展字段
+  - 统一持久化策略
+- `backend/app/services/system_logs.py`
+  - 逐步从“文本尾部查看器”升级为“运行日志聚合器”
+- `backend/app/services/log_classification.py`
+  - 指纹
+  - 根因分类
+  - 纠错建议
+- `backend/app/api/v1/system_control.py`
+  - 补充事件 / 审计 / 日志多视图接口
+
+### Frontend
+
+建议新增/增强：
+
+- `frontend/src/lib/logger.ts`
+  - 统一前端 logger API
+- `frontend/src/pages/Logs/Logs.tsx`
+  - 升级为多层工作台
+- `frontend/public/earth/js/...`
+  - 各 Earth 模块接入统一事件 logger
+
+## Event Naming Convention
+
+建议采用：
+
+`<domain>.<resource>.<action>.<result>`
+
+示例：
+
+- `collector.datasource.run.started`
+- `collector.datasource.run.failed`
+- `earth.layer.cables.load.failed`
+- `earth.cruise.route.build.failed`
+- `system.restart_task.requested`
+- `system.restart_task.completed`
+- `auth.websocket.connect.failed`
+- `settings.datasource.priority.updated`
+
+规则：
+
+- 不用自然语言句子
+- 不把 ID 塞进 event 名里
+- 资源对象通过字段承载，不通过 event 名承载
+
+## Query Model
+
+最终推荐支持的查询维度：
+
+- 时间范围
+- level
+- source
+- service
+- module
+- event
+- request_id
+- trace_id
+- user_id / actor
+- resource_type / resource_id
+- category
+- fingerprint
+- status
+- full-text search
+
+## Retention Strategy
+
+推荐保留策略：
+
+- 运行日志：
+  - 文件 / 容器 / Redis 缓冲保留短周期
+- 持久化事件：
+  - 保留中长期
+- 审计日志：
+  - 长期保留
+
+初版可以先这样：
+
+- 运行日志：7 到 14 天
+- 关键事件：90 到 180 天
+- 审计日志：180 天以上
+
+后续再根据存储与合规要求调整。
+
+## Security And Compliance
+
+必须落实：
+
+- 敏感字段脱敏
+- 前端上报白名单
+- 防止日志注入
+- 审计日志不可被普通管理员随意篡改
+- 高敏感纠错动作必须再次鉴权
+
+## Success Criteria
+
+当下面这些条件成立时，才算这套日志系统真的“成了”：
+
+1. 一个后端请求失败时，能通过 `request_id` 在运行日志、持久化事件、审计日志之间串联查询
+2. 一个 Earth 前端错误能定位到页面、模块、事件名和上下文
+3. 一个采集器失败能同时看到运行日志、持久化事件和可执行纠错动作
+4. 一个管理员操作能查到操作者、目标对象、结果和 request_id
+5. 日志页不再只是文本控制台，而是完整的“运行日志 / 关键事件 / 审计日志”工作台
+6. 高频已知错误能聚合并给出修复建议
+
+## Delivery Order
+
+推荐严格按下面顺序做，不要乱跳：
+
+1. Phase 0 命名与字段规范冻结
+2. Phase 1 后端结构化 logging 基础
+3. Phase 2 高价值事件持久化
+4. Phase 3 前端 / Earth 统一 logger
+5. Phase 4 审计覆盖补齐
+6. Phase 5 日志工作台 UI 重构
+7. Phase 6 指纹 / 纠错 / runbook
+
+原因：
+
+- 如果不先统一字段和命名，后面 UI 和持久化会越来越乱
+- 如果不先做后端结构化基础，前端上报再多也串不起来
+- 如果不先补持久化层，就只有“实时可看”，没有“历史可查”
+
+## First Actionable Milestone
+
+如果要从明天就开始做，最合理的第一个里程碑是：
+
+### M1: 让后端关键路径全部拥有统一结构化事件
+
+范围：
+
+- API 请求入口/出口
+- scheduler
+- collectors
+- websocket
+- visualization
+- system control
+
+交付物：
+
+- 统一 logger helper
+- 统一 event naming 表
+- 统一 request_id 注入
+- 统一脱敏策略
+- 关键模块替换完成
+
+完成这个里程碑后，Planet 才算真正拥有了“企业级日志系统的地基”。
+

docs/plans/frontend-ai-playground-development-plan.md

@@ -30,7 +30,7 @@
 - [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)
+- [docs/technical/agents-aiprovider.md](/home/ray/dev/linkong/planet/docs/technical/agents-aiprovider.md)
 
 ### 2. 本地运行与配置打通
 
@@ -77,7 +77,7 @@
 
 相关文件：
 
-- [docs/frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/frontend-layout-guidelines.md)
+- [docs/technical/frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/technical/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

docs/technical/README.md

@@ -0,0 +1,27 @@
+# Technical Docs
+
+这里放“当前实现和当前结构”的文档，重点回答：
+
+- 现在代码是怎么组织的
+- 当前入口在哪
+- 状态和组件如何工作
+- 后续改动应该沿着哪条实现边界继续走
+
+适合放入这里的内容：
+
+- 前端上下文
+- Earth 前端结构
+- Earth 卫星 footprint 策略
+- 后端运行控制
+- collector 现状
+- 采集格式约定
+
+不适合放入这里的内容：
+
+- 尚未完成的 roadmap
+- 未来迭代方案
+- 大范围重构计划
+
+这些应放入：
+
+- [docs/plans/README.md](/home/ray/dev/linkong/planet/docs/plans/README.md)

docs/technical/earth-bgp-context.md

@@ -187,7 +187,7 @@ Current reality:
 - that is expected, because incidents are aggregated and de-noised
 - but incident-first rendering makes the Earth view look too quiet unless there is another always-available activity layer
 
-Implementation detail for the recommended `activity layer` is expanded in [bgp-region-aggregation-plan.md](/home/ray/dev/linkong/planet/docs/bgp-region-aggregation-plan.md).
+Implementation detail for the recommended `activity layer` is expanded in [bgp-region-aggregation-plan.md](/home/ray/dev/linkong/planet/docs/plans/earth-bgp-region-aggregation-plan.md).
 
 So the immediate next milestone is:
 

docs/technical/earth-frontend-context.md

@@ -0,0 +1,381 @@
+# Earth Frontend Context
+
+本文件描述当前 Earth 大屏前端的真实结构，重点是帮助后续继续改 HUD、图层、媒体面板、真实地形、BGP 可视化时，不再重复踩结构和状态同步上的坑。
+
+相关规则建议一起参考：
+
+- [rules.md](/home/ray/dev/linkong/planet/rules.md)
+- [frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/technical/frontend-layout-guidelines.md)
+
+## 当前目标
+
+Earth 前端不是普通管理页，它是独立的大屏展示前端。当前产品目标是：
+
+- 维持地球视图的空间感和可读性
+- 让 HUD、图层、媒体面板、BGP、卫星、海缆等保持统一交互
+- 把加载中、已启用、已隐藏、锁定中这类状态做清楚
+
+## 当前入口
+
+React 路由入口：
+
+- [Earth.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Earth/Earth.tsx)
+
+当前做法很简单：
+
+- React 页面只负责提供一个全屏 `iframe`
+- 真正的 Earth 应用运行在：
+  - [index.html](/home/ray/dev/linkong/planet/frontend/public/earth/index.html)
+
+所以 Earth 前端本质上是 `public/earth` 下的一套独立静态应用。
+
+## 当前文件分层
+
+### 1. 页面入口与结构
+
+- [index.html](/home/ray/dev/linkong/planet/frontend/public/earth/index.html)
+
+职责：
+
+- HUD 基础 DOM
+- 图层面板
+- 媒体面板
+- 工具栏
+- 设置弹窗
+- 兼容旧元素 id
+
+### 2. 主运行时
+
+- [main.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/main.js)
+
+职责：
+
+- 地球初始化
+- Three.js 场景组装
+- 数据加载与刷新
+- 各图层集成
+- Earth 级别状态同步
+
+### 3. 地球控制层
+
+- [controls.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/controls.js)
+
+职责：
+
+- 工具栏交互
+- 图层面板交互
+- 旋转/缩放/布局
+- HUD 面板拖拽
+- 图层开关状态机
+- Earth 设置读取、持久化与重置
+
+这份文件是 Earth 前端当前最核心的 UI 控制入口。
+
+### 4. UI 与状态消息
+
+- [ui.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/ui.js)
+
+职责：
+
+- loading 面板
+- status message
+- tooltip / error / 清理逻辑
+
+### 5. 地球与地形
+
+- [earth.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/earth.js)
+- [terrain.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/terrain.js)
+
+职责：
+
+- 地球球体、云层、大气
+- 真实地形 mesh
+- terrain tile 拉取、解码、位移、着色
+
+### 6. 图层模块
+
+- [satellites.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/satellites.js)
+- [cables.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/cables.js)
+- [bgp.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/bgp.js)
+- [bgp-cruise-adapter.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/bgp-cruise-adapter.js)
+- [news.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/news.js)
+- [tv.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/tv.js)
+- [layer-startup-tasks.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/layer-startup-tasks.js)
+- [cruise-sequencer.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/cruise-sequencer.js)
+- [callout-connector.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/callout-connector.js)
+
+职责：
+
+- 各自的数据层
+- 开关行为
+- 面板内容
+- hover/lock/selection 语义
+
+其中 Earth 启动加载链现在也拆成了两层：
+
+- `controls.js`
+  - 提供图层注册表与启动元信息
+- `layer-startup-tasks.js`
+  - 提供图层启动任务注册表
+  - 通过 `registerLayerStartupTask(id, taskFactory)` 扩展启动任务
+- `main.js`
+  - 只负责读取排序后的启动图层，再按映射执行队列
+
+其中巡航模式现在已经拆成两层：
+
+- `cruise-sequencer.js`
+  - 负责目标队列顺序、停留时长、切换节奏、打断与恢复
+- `callout-connector.js`
+  - 负责卡片连线 SVG、路径计算与绘制动画
+- `bgp-cruise-adapter.js`
+  - 负责 BGP 巡航展示适配：目标排序、卡片落点、连线路径、focus/overlay/info-card 时序
+
+当前 BGP 巡航只是这套能力的一个调用方，不应再把“按队列巡航”和“BGP 事件展示”混写在同一个状态机里。
+
+## 当前样式分层
+
+Earth 的 CSS 不是一份大样式表，而是分层管理：
+
+- [base.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/base.css)
+- [hud.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/hud.css)
+- [toolbar.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/toolbar.css)
+- [layer-panel.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/layer-panel.css)
+- [info-panel.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/info-panel.css)
+- [legend.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/legend.css)
+- [earth-stats.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/earth-stats.css)
+- [coordinates-display.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/coordinates-display.css)
+- [tv-panel.css](/home/ray/dev/linkong/planet/frontend/public/earth/css/tv-panel.css)
+
+当前建议：
+
+- 通用 HUD 壳层写进 `hud.css`
+- 单一面板特性写进各自子文件
+- 不要把业务状态样式再散回 `index.html`
+
+## 当前图层开关状态语义
+
+Earth 图层按钮现在不应再只有“开/关”两态，而应支持：
+
+- `inactive`
+- `active`
+- `loading`
+
+当前入口在：
+
+- [controls.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/controls.js)
+- [layer-button-state.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/layer-button-state.js)
+
+关键函数：
+
+- `updateLayerButtonState(button, isActive)`
+- `setLayerButtonState(button, options)`
+
+`setLayerButtonState` 负责：
+
+- `loading` 样式
+- `aria-busy`
+- 按钮禁用
+- tooltip 更新
+- 绑定状态文本更新
+- 可选同步 `active`
+
+因此后续如果别的图层也需要异步启用，应该直接走这套状态机，而不是再手写一套临时 loading class。
+
+另外，Earth 图层控制现在已经收成“注册表驱动”：
+
+- 图层元数据
+  - `id`
+  - `icon`
+  - `label`
+  - `meta`
+  - `buttonId`
+  - `persist`
+  - `startupPriority`
+  - `startupMode`
+  - `startupLabel`
+  - `startupMessage`
+- 图层行为
+  - `getVisible()`
+  - `setVisible(next, options)`
+
+当前入口仍在 [controls.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/controls.js)。
+
+这意味着后续新增图层时，优先应补一条图层注册定义，而不是同时去改：
+
+- 图层面板 HTML
+- 持久化快照
+- 初始化恢复
+- click 绑定
+
+这四处现在都应该由注册表派生。
+
+其中：
+
+- `startupPriority`
+  - 描述图层参与启动加载时的顺序
+- `startupMode`
+  - `visible`
+    - 仅当前图层处于启用/可见状态时，才加入启动加载队列
+  - `preload`
+    - 即使当前图层未显示，也会参与启动预加载
+
+当前 `main.js` 会通过注册表读取排序后的启动图层列表，再动态拼装启动加载队列，而不是手写一串固定步骤。像 BGP 这类需要尽早准备数据、但不一定默认显示的图层，应该优先走 `startupMode: "preload"`，而不是在启动流程里写隐式特判。
+
+此外，启动阶段给用户看的提示文案也应尽量从注册表派生：
+
+- `startupLabel`
+  - 用于描述当前启动任务的业务名称
+- `startupMessage`
+  - 用于描述启动中的提示文案
+  - 可以是字符串
+  - 也可以是对象，用于像海缆这种“准备阶段 / 主加载阶段”两段式文案
+
+这样后续新增会参与启动加载的图层时，顺序、模式和提示文案都在同一处定义，不需要再去 `main.js` 里补第二套常量。
+
+### `data-status-target`
+
+图层按钮可以通过：
+
+- `data-status-target`
+
+指向一个状态文本节点。当前 terrain 已接入：
+
+- 按钮：`#toggle-terrain`
+- 状态节点：`#terrain-status`
+
+以后别的异步图层也可以沿用这套约定。
+
+## 当前设置持久化
+
+Earth 设置面板当前由 [controls.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/controls.js) 统一负责：
+
+- 捕获默认值
+- 从 `localStorage` 读取上次设置
+- 初始化应用当前设置
+- 用户变更后即时持久化
+- 一键重置回默认值
+
+当前持久化的范围是：
+
+- 旋转模式
+- 地球默认大小（作为重置视角、缩放重置和巡航视图的默认 zoom 真源）
+- HUD 面板显示/隐藏
+- 图层控制开关：`地形 / 卫星 / 轨迹 / 海缆 / BGP`
+- 地形透明度
+
+也就是说，Earth 设置不是一次性 UI 状态了，而是本地设备级偏好。后续如果再加入新的设置项，应优先接入同一条持久化链，而不是各自散着写 `localStorage`。
+
+## 当前地形链路
+
+真实地形首次启用会慢，原因不只是一个：
+
+1. 需要拉取 Terrarium 瓦片
+2. 需要解码图片
+3. 需要按顶点采样高程
+4. 需要重新写入 geometry 和 color
+5. 需要重新计算法线与包围体
+
+当前入口在：
+
+- [terrain.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/terrain.js)
+
+当前已经做了两层体验优化：
+
+1. 图层开关 loading 状态持续可见
+2. 页面空闲时会预热 `ensureTerrainReady()`
+
+也就是说，后续再继续优化 terrain 时，优先顺序应该是：
+
+1. 先保证用户感知正确
+2. 再压缩首次等待
+3. 最后才做更激进的几何/瓦片优化
+
+## 当前高频风险点
+
+### 1. 视觉状态和业务状态不同步
+
+Earth 里最常见的 bug 不是“没渲染”，而是：
+
+- 图层关了，tooltip 还在
+- 锁定对象隐藏了，info card 还在
+- legend 没跟图层切换
+- loading 已结束，但按钮还像没开
+
+后续改动必须优先检查状态同步。
+
+### 2. HUD 布局问题先查结构，不要先打 CSS 补丁
+
+Earth HUD 历史上反复出现：
+
+- 面板只剩一条缝
+- markdown 被裁掉
+- tabs/iframe 被 `overflow: hidden` 吃掉
+
+优先检查：
+
+1. 谁负责高度
+2. 谁负责滚动
+3. 哪一层在裁剪
+
+不要上来先加 `overflow: hidden` 或额外包装层。
+
+### 3. Transitional path 必须收口
+
+Earth 已经经历过多轮 HUD、toolbar、media panel 重构，所以最容易积累：
+
+- 旧 helper
+- 旧 class
+- 旧 fallback 逻辑
+- 已废弃变体
+
+每次大功能完成后，都要做一次 cleanup pass。
+
+### 4. 巡航与业务事件不要再深度耦合
+
+当前正确边界应该是：
+
+- 通用巡航层只知道：
+  - 当前目标
+  - 队列顺序
+  - 相机 focus
+  - 停留 / 隐藏 / 切换
+- 业务模块只负责：
+  - 提供目标队列
+  - 提供 focus 坐标
+  - 提供卡片内容
+  - 提供高亮/图层副作用
+
+如果以后再给海缆、卫星或新闻做巡航，不应复制一套新的 `main.js` 状态变量，而应复用：
+
+- [cruise-sequencer.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/cruise-sequencer.js)
+- [callout-connector.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/callout-connector.js)
+- [bgp-cruise-adapter.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/bgp-cruise-adapter.js) 这种业务适配层模式
+
+## 当前推荐改动方式
+
+如果后续继续改 Earth，建议按这个顺序：
+
+1. 先确认改的是：
+   - Three.js 渲染层
+   - HUD 结构层
+   - 图层状态层
+   - 面板内容层
+2. 如果涉及图层按钮，优先接入统一状态机
+3. 如果涉及可见性切换，检查 tooltip / legend / info-card / lock 是否一起收口
+4. 如果涉及面板布局，先查结构再动 CSS
+
+## 当前与控制台前端的边界
+
+Earth 前端和控制台前端不是同一套 UI 系统：
+
+- 控制台前端：React + Ant Design 工作台
+- Earth 前端：`public/earth` 原生 HUD + Three.js 展示面
+
+因此：
+
+- Earth 不应该直接复用 Ant Table / AppLayout 语义
+- 控制台也不应该照搬 Earth HUD 动画和玻璃层语言
+
+控制台相关结构见：
+
+- [admin-frontend-context.md](/home/ray/dev/linkong/planet/docs/technical/frontend-admin-frontend-context.md)

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

@@ -95,3 +95,93 @@
   - 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`
+- `news_live_streams` 在未配置 override 时，默认使用 `iptv-org`：
+  - `channels.json`
+  - `streams.json`
+  - `logos.json`
+  并自动筛出新闻类频道目录
+
+## 采集器配置方式
+
+`news_live_streams` 不需要单独新页面，直接复用现有数据源配置：
+
+- `endpoint`
+  - 频道目录 JSON API 地址
+- `auth_type`
+  - `none` / `bearer` / `api_key` / `basic`
+- `headers`
+  - 额外请求头
+- `config`
+  - 采集器请求与解析行为
+
+### 支持的 `config` 字段
+
+```json
+{
+  "timeout": 30,
+  "method": "GET",
+  "params": {
+    "region": "global"
+  },
+  "body_type": "json",
+  "body": {
+    "include_disabled": false
+  },
+  "response_path": "payload.channels"
+}
+```
+
+- `timeout`
+  - 请求超时秒数
+- `method`
+  - `GET` 或 `POST`
+- `params`
+  - 查询参数对象
+- `body_type`
+  - `json` 或 `form`
+- `body`
+  - 配合 `POST` 使用的请求体
+- `json_body`
+  - 显式 JSON 请求体，优先级高于 `body`
+- `form_body`
+  - 显式表单请求体，优先级高于 `body`
+- `response_path`
+  - 返回 JSON 中频道数组所在路径，支持点路径，例如：
+    - `payload.channels`
+    - `data.items`
+    - `result.streams`
+
+### 认证补充
+
+- `bearer`
+  - 使用 `Authorization: Bearer <token>`
+- `api_key`
+  - 默认作为请求头发送
+  - 如果 `auth_config.in = "query"`，则作为 query param 发送
+- `basic`
+  - 使用 HTTP Basic Authorization
+
+## 兼容的响应结构
+
+采集器会优先读取：
+
+- 顶层数组
+- 或这些常见字段下的数组：
+  - `sources`
+  - `streams`
+  - `channels`
+  - `items`
+  - `results`
+  - `data`
+
+同时会兼容这些字段别名：
+
+- `id` / `source_id` / `slug` / `channel_id` / `code`
+- `name` / `title` / `channel` / `display_name`
+- `provider` / `publisher` / `network`
+- `stream_url` / `stream` / `playback_url` / `hls_url` / `m3u8_url`
+- `embed_url` / `embed` / `page_url`
+- `homepage_url` / `source_url` / `website`
+- `language` / `lang` / `locale`
+- `youtube_video_id` / `video_id`
+- `youtube_channel` / `channel_handle`

docs/technical/earth-satellite-footprint-policy.md

@@ -0,0 +1,198 @@
+# Earth Satellite Footprint Policy
+
+本文件记录 Earth 卫星图层当前关于 `footprint` 的产品边界、资料依据和已落地实现，目标是避免把 Starlink 这套专用地表覆盖模型误用到其它星座上。
+
+相关上下文：
+
+- [earth-frontend-context.md](/home/ray/dev/linkong/planet/docs/technical/earth-frontend-context.md)
+- [backend-collectors.md](/home/ray/dev/linkong/planet/docs/technical/backend-collectors.md)
+- [backend/app/services/collectors/celestrak.py](/home/ray/dev/linkong/planet/backend/app/services/collectors/celestrak.py)
+- [frontend/public/earth/js/satellites.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/satellites.js)
+
+## 当前目标
+
+- 明确哪些非 Starlink 卫星不该显示贴地 footprint
+- 明确哪些星座未来可以有独立 footprint，但不能复用 Starlink bowtie / GSO-gap 模型
+- 把这条策略沉淀成可执行实现边界，而不是继续散落在视觉参数里
+
+## 本地实际类别
+
+当前 CelesTrak 卫星分组在 [backend/app/services/collectors/celestrak.py](/home/ray/dev/linkong/planet/backend/app/services/collectors/celestrak.py) 中包括：
+
+- `starlink`
+- `gps-ops`
+- `galileo`
+- `glonass`
+- `beidou`
+- `leo`
+- `geo`
+- `iridium-next`
+
+其中非 Starlink 类别是：
+
+- `gps-ops`
+- `galileo`
+- `glonass`
+- `beidou`
+- `leo`
+- `geo`
+- `iridium-next`
+
+## 资料结论
+
+### 1. GNSS / RNSS: `gps-ops`, `galileo`, `glonass`, `beidou`
+
+默认不要画局部地表 footprint。
+
+原因：
+
+- 公开资料强调的是 `Earth-pointing`、`Earth coverage`、`continuous global coverage`
+- 这类系统的公开语义是全球导航 / 授时覆盖，不是 Starlink 那种面向终端业务的局部 spot footprint
+
+更合适的表示：
+
+- 默认只显示卫星本体和轨道
+- 如果后续要强调“服务可达性”，只能做很弱的 global coverage 语义，不应画贴地局部光斑
+
+资料：
+
+- [GPS III EC Antenna Patterns](https://www.navcen.uscg.gov/sites/default/files/pdf/gps/GPS_ZIP/GPS_III_EC_Antenna_Patterns_SVN_74_75_76_77_78.pdf)
+- [ESA Galileo satellites](https://www.esa.int/Applications/Satellite_navigation/Galileo/Galileo_satellites)
+- [Navipedia Galileo General Introduction](https://gssc.esa.int/navipedia/index.php/Galileo_General_Introduction)
+- [BeiDou official overview](https://www.beidou.gov.cn/xt/gfxz/201812/P020190117356387956569.pdf)
+- [GPS.gov GNSS overview](https://www.gps.gov/systems/gnss/)
+
+### 2. `iridium-next`
+
+可以有 footprint，但不能复用 Starlink 的单一 bowtie footprint。
+
+原因：
+
+- Iridium NEXT 公开资料强调的是固定多 spot beam 体系
+- 公开示例里常见的是 `48 fixed spot beams in 4 tiers`
+- 这和 Starlink 当前这套“单星、单主 footprint、带 GSO 缺口”的业务可视化不是同一个问题
+
+更合适的表示：
+
+- 默认：仍然不画 Starlink 式地表 footprint
+- 后续如果要做：单独接入 Iridium 多波束适配层
+- 在视觉上更接近多束 cluster / 蜂窝 / 分层束，而不是单个 bowtie 光斑
+
+资料：
+
+- [Iridium Satellite Spot Beam Coverage on the US](https://www.mathworks.com/help/phased/ug/iridium-satellite-spot-beam-coverage-on-the-us-1.html)
+
+### 3. `geo`
+
+默认不要画统一 footprint。
+
+原因：
+
+- GEO 通信星公开上可能是 global beam、zone beam、spot beam、steerable spot beam
+- 没有 operator / payload / beam contour 元数据时，统一画一个 footprint 很容易错
+
+更合适的表示：
+
+- 默认只显示 GEO belt 和卫星驻点语义
+- 只有拿到 beam contour / operator metadata 时才允许画 footprint
+
+资料：
+
+- [ITU Handbook on Satellite](https://www.itu.int/dms_pub/itu-r/opb/hdb/R-HDB-42-2002-PDF-E.pdf)
+
+### 4. `leo`（generic）
+
+默认不要画 footprint。
+
+原因：
+
+- `leo` 组过于混杂，可能同时包含通信、遥感、试验、观测等不同任务
+- 没有 mission / payload / antenna pattern 元数据时，无法判断是否存在可视化意义上的服务覆盖面
+
+更合适的表示：
+
+- 默认只显示卫星和轨道
+- 后续如果按 operator / mission subtype 细分，再决定是否引入独立 coverage mode
+
+## 产品策略
+
+当前统一策略如下：
+
+- `Starlink`
+  - 保留当前专用 `ground_footprint` 逻辑
+- `Iridium NEXT`
+  - 预留独立适配层
+  - 当前不复用 Starlink footprint
+- `GPS / Galileo / GLONASS / BeiDou`
+  - 不显示贴地 footprint
+- `GEO`
+  - 无 beam metadata 不显示 footprint
+- `generic LEO`
+  - 无 mission metadata 不显示 footprint
+
+## 已落地实现
+
+本次实现只做最小可执行版本，不改现有 Starlink 视觉参数：
+
+1. 后端把星座分组和 footprint 策略提示透给前端
+
+- CelesTrak collector 会把 `GROUP` 记入 `metadata.constellation_group`
+- Visualization API 会输出：
+  - `properties.constellation_group`
+  - `properties.footprint_policy`
+
+当前策略值：
+
+- `starlink_ground_footprint`
+- `iridium_coverage_ring`
+- `none`
+
+对应代码：
+
+- [backend/app/services/collectors/celestrak.py](/home/ray/dev/linkong/planet/backend/app/services/collectors/celestrak.py)
+- [backend/app/api/v1/visualization.py](/home/ray/dev/linkong/planet/backend/app/api/v1/visualization.py)
+
+2. 前端把 footprint 变成 capability-gated renderer
+
+- `ground_footprint` 只有在 `footprint_policy === starlink_ground_footprint` 时才真正启用
+- `iridium-next` 不再回退成占位分支，而是走独立的 Iridium coverage ring adapter
+- 其它非 Starlink 即使用户全局选择了 `ground_footprint`，也会自动回退到 `self_glow`
+
+对应代码：
+
+- [frontend/public/earth/js/satellites.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/satellites.js)
+- [frontend/public/earth/js/iridium-footprint-adapter.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/iridium-footprint-adapter.js)
+
+3. 卫星信息卡显示 capability，而不是只显示轨道参数
+
+- 卫星详情现在会明确显示：
+  - `星座/分组`
+  - `覆盖能力`
+  - `当前显示`
+  - `覆盖模型`
+- 这样用户能直接看到：
+  - 当前卫星是否支持 footprint
+  - 当前显示是不是因为 capability gating 被回退
+  - Iridium 和 Starlink 使用的不是同一种模型
+
+对应代码：
+
+- [frontend/public/earth/js/main.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/main.js)
+- [frontend/public/earth/js/info-card.js](/home/ray/dev/linkong/planet/frontend/public/earth/js/info-card.js)
+
+## 当前实现边界
+
+这条边界需要继续保持：
+
+- `Starlink` 的 footprint 参数和 shader 逻辑只服务于 Starlink
+- 非 Starlink 的能力判断属于“策略层 / 适配层”
+- 不要把不同星座的覆盖模型再混写进同一套参数里
+- `iridium-next` 已经切成独立 adapter，应继续沿这条边界演进，而不是给现有 Starlink bowtie 增加更多 if/else
+
+## 后续建议
+
+如果继续往前做，推荐顺序是：
+
+1. 为 `iridium-next` 新建独立 footprint adapter
+2. 在 UI 上补一个只读提示，让用户知道当前卫星是否支持 footprint
+3. 如果未来拿到 GEO beam contour / operator metadata，再为 GEO 开 operator-specific footprint

docs/technical/frontend-admin-frontend-context.md

@@ -0,0 +1,236 @@
+# Admin Frontend Context
+
+本文件描述当前控制台前端的真实结构，目标是帮助后续页面开发、表格改造、布局治理和状态收口时快速找到正确入口。
+
+相关规则建议一起参考：
+
+- [rules.md](/home/ray/dev/linkong/planet/rules.md)
+- [frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/technical/frontend-layout-guidelines.md)
+
+## 当前目标
+
+控制台前端承担的是后台工作台，而不是展示型大屏。当前约束是：
+
+- 页面默认遵循单屏工作区
+- 主交互在内部模块滚动，而不是依赖整页无限变长
+- 列表、表格、分析页优先保证主工作区可见
+- 通用布局、滚动条、表格滚动行为尽量复用，不要每页各写一套
+
+## 当前路由入口
+
+主入口在：
+
+- [App.tsx](/home/ray/dev/linkong/planet/frontend/src/App.tsx)
+
+当前后台相关路由包括：
+
+- `/admin`
+- `/users`
+- `/datasources`
+- `/data`
+- `/alerts/system`
+- `/alerts/bgp`
+- `/alerts/situational`
+- `/bgp`
+- `/playground`
+- `/settings`
+
+`/earth` 是独立展示页，不属于控制台骨架。
+
+## 当前页面骨架
+
+控制台公共壳层在：
+
+- [AppLayout.tsx](/home/ray/dev/linkong/planet/frontend/src/components/AppLayout/AppLayout.tsx)
+
+职责：
+
+- 左侧导航
+- 折叠与展开
+- 当前账号/版本信息
+- 内容区高度闭合
+- 全站统一侧边栏滚动条
+
+当前结构是：
+
+```tsx
+<Layout className="dashboard-layout">
+  <Sider className="dashboard-sider">...</Sider>
+  <Layout>
+    <Content className="dashboard-content">
+      <div className="dashboard-content-inner">{children}</div>
+    </Content>
+  </Layout>
+</Layout>
+```
+
+后续控制台页面应优先适配这套壳层，而不是重新定义全页高度语义。
+
+## 当前共享组件
+
+### 1. `Scrollbar`
+
+文件：
+
+- [Scrollbar.tsx](/home/ray/dev/linkong/planet/frontend/src/components/Scrollbar/Scrollbar.tsx)
+
+用途：
+
+- 控制台侧边栏这类普通内容容器
+- 组件内部管理可见性、thumb 尺寸、拖拽和双轴 overflow 判定
+
+当前约束：
+
+- 滚动条必须是浮层，不参与布局
+- 无 overflow 时不应留下可见痕迹
+- 真实滚动仍交给原生容器，只替换可见层和交互层
+
+### 2. `ScrollbarOverlay`
+
+文件：
+
+- [ScrollbarOverlay.tsx](/home/ray/dev/linkong/planet/frontend/src/components/Scrollbar/ScrollbarOverlay.tsx)
+
+用途：
+
+- Ant Table 这类内部已有滚动容器的区域
+- 不接管滚动语义，只叠加新的滚动条可见层
+
+当前使用场景：
+
+- 数据源
+- 采集数据
+- 用户管理
+- 设置页
+- 告警页
+- BGP 页面
+
+### 3. `TableScrollRegion`
+
+文件：
+
+- [TableScrollRegion.tsx](/home/ray/dev/linkong/planet/frontend/src/components/Scrollbar/TableScrollRegion.tsx)
+
+用途：
+
+- 为表格滚动区提供统一包裹层
+- 后续新表格页优先复用，不要重复写“表格区域 + overlay scrollbar”样板
+
+### 4. 其他共享组件
+
+- [MarkdownRenderer.tsx](/home/ray/dev/linkong/planet/frontend/src/components/MarkdownRenderer/MarkdownRenderer.tsx)
+- [TableActions.tsx](/home/ray/dev/linkong/planet/frontend/src/components/TableActions/TableActions.tsx)
+
+## 当前状态来源
+
+### 1. 认证状态
+
+文件：
+
+- [auth.ts](/home/ray/dev/linkong/planet/frontend/src/stores/auth.ts)
+
+职责：
+
+- token
+- 当前用户
+- 登录/退出
+
+`App.tsx` 用它判断是否进入登录页。
+
+### 2. 业务数据网关
+
+目前 AI / 态势感知相关服务集中在：
+
+- [http-gateway.ts](/home/ray/dev/linkong/planet/frontend/src/services/situational-awareness/http-gateway.ts)
+- [port.ts](/home/ray/dev/linkong/planet/frontend/src/services/situational-awareness/port.ts)
+- [types.ts](/home/ray/dev/linkong/planet/frontend/src/services/situational-awareness/types.ts)
+
+约束：
+
+- 页面不要直接散落拼 URL
+- 先通过 port/types 定义边界
+- 再由 http/mock gateway 实现
+
+## 当前页面分层建议
+
+### 1. 仪表盘和摘要型页面
+
+例如：
+
+- [Dashboard.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Dashboard/Dashboard.tsx)
+
+优先目标：
+
+- 页头稳定
+- 摘要卡片先紧凑化
+- 主工作区占据主要高度
+
+### 2. 表格型页面
+
+例如：
+
+- [DataSources.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/DataSources/DataSources.tsx)
+- [DataList.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/DataList/DataList.tsx)
+- [Users.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Users/Users.tsx)
+- [Settings.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Settings/Settings.tsx)
+
+约束：
+
+- 优先内部滚动
+- 不要让表格撑爆整页
+- 新表格区域优先复用 `TableScrollRegion` / `ScrollbarOverlay`
+
+### 3. 复杂工作区页面
+
+例如：
+
+- [BGP.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/BGP/BGP.tsx)
+- [Playground.tsx](/home/ray/dev/linkong/planet/frontend/src/pages/Playground/Playground.tsx)
+
+约束：
+
+- Tabs 里的内容不能套同一套高度逻辑
+- 表格 tab、Markdown tab、配置 tab 要各自定义滚动责任
+- AI 结果区、长文本区优先保证最小可读高度
+
+## 当前布局约束
+
+这些原则已经在项目里反复验证过：
+
+1. 父容器高度链要闭合
+2. `min-height: 0` 不能漏
+3. overflow 责任必须明确
+4. 不要用 `overflow: hidden` 掩盖结构问题
+5. 不要为了摘要卡完整显示去压缩主工作区
+6. 自定义滚动条必须是浮层，不得挤压内容宽度
+
+详细经验见：
+
+- [frontend-layout-guidelines.md](/home/ray/dev/linkong/planet/docs/technical/frontend-layout-guidelines.md)
+
+## 当前推荐改动方式
+
+如果后续继续改后台页面，建议按这个顺序：
+
+1. 先确认页面属于摘要页、表格页还是复杂工作区
+2. 先接入现有壳层和滚动语义
+3. 优先复用共享滚动组件
+4. 最后再改视觉和细节交互
+
+不要先写局部 CSS 补丁，再回头补结构。
+
+## 当前明显边界
+
+控制台前端和 Earth 前端不是一套系统：
+
+- 控制台前端是 React + Ant Design 工作台
+- Earth 前端是 `public/earth` 下的独立原生 HUD 系统
+
+因此：
+
+- 不要把 Earth 的 HUD/动画/状态机直接挪进控制台
+- 不要把控制台表格/滚动策略硬套到 Earth HUD
+
+Earth 相关结构见：
+
+- [earth-frontend-context.md](/home/ray/dev/linkong/planet/docs/technical/earth-frontend-context.md)

docs/version-history.md

@@ -16,12 +16,44 @@
 ## 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：`0.27.0`
+- `dev` 当前开发分支历史推导到：`0.40.1`
 
 ## Timeline
 
 | Version | Type | Branch | Commit | Summary |
 | --- | --- | --- | --- | --- |
+| `0.40.1` | improvement | `dev` | `pending` | 卫星选中标记配色跟随图例，修复 footprint 遮蔽卫星渲染问题，修复选中海缆误触发卫星高亮 |
+| `0.40.0` | feature | `dev` | `pending` | Earth 卫星 footprint 按星座能力分层，Iridium 独立 coverage ring 落地，卫星详情卡补齐覆盖能力与当前显示说明 |
+| `0.39.0` | feature | `dev` | `pending` | 后端统一结构化日志地基落地，系统日志页重构为紧凑日志工作台，并修复 Earth 移动端态势抽屉与新闻详情同步问题 |
+| `0.38.0` | feature | `dev` | `pending` | Earth 新闻接入通用巡航与专用卡片链路，系统日志页升级为结构化时间/级别过滤与真正字符串检索 |
+| `0.37.2` | bugfix | `dev` | `pending` | Earth 图层系统新增经纬线开关，并将经纬线接入统一 layer registry、移动端抽屉与设置持久化流 |
+| `0.37.1` | bugfix | `dev` | `pending` | 修复 `planet.sh` 在 `uvicorn --reload` 场景下未清理旧 worker 的问题，避免后端重启后仍停留旧实例并导致算力中心聚合接口 404 |
+| `0.37.0` | feature | `dev` | `pending` | Earth 连线系统从巡航语义中完全解耦为通用 callout connector，统一桌面/移动端对象级锚点、临界区锚点滑动与稳定巡航展示链路 |
+| `0.36.0` | feature | `dev` | `pending` | Earth 新增统一算力中心图层与估算位置展示，继续收口拖拽交互，并补充 AI Provider 指纹与 WSL 局域网访问支撑 |
+| `0.35.1` | bugfix | `dev` | `pending` | 收口 Earth 桌面 HUD 与移动端抽屉的统一统计绑定机制，修复态势统计在图层切换后的同步遗漏 |
+| `0.35.0` | feature | `dev` | `pending` | Earth 移动端抽屉系统与悬浮卡片全面上线：手势驱动抽屉、点击物件弹出可拖动详情卡、单指旋转双指缩放地球 |
+| `0.34.0` | feature | `dev` | `pending` | Earth 搜索面板正式接入，`planet.sh --allow-lan` 打通 Bun + Vite 局域网开放链路，并自动输出推荐访问地址与健康检查地址 |
+| `0.33.0` | feature | `dev` | `pending` | `news_live_streams` 默认接入 iptv-org 频道目录，内置数据源支持直接编辑 override，并修复 TV 合并采集源后默认频道消失的问题 |
+| `0.32.0` | feature | `dev` | `pending` | Earth 设置新增默认地球大小真源，并继续收口卫星焦点层次、toolbar/scrollbar 性能与 HUD 设置面板细节 |
+| `0.31.3` | bugfix | `dev` | `pending` | 收口 Earth 图层注册表与启动任务框架，修复旋转/巡航切换、卫星地形遮挡与日夜关闭照明回归 |
+| `0.31.2` | bugfix | `dev` | `pending` | 将 Earth 巡航模式拆成通用 sequencer、通用连线和 BGP 巡航适配层，并修复空白点击推进与连线动画回归 |
+| `0.31.1` | bugfix | `dev` | `pending` | Earth 图层开关统一 loading 状态机，卫星首次加载可见化，并将文档按 technical / plans / deprecated 重构归档 |
+| `0.31.0` | feature | `dev` | `pending` | Earth 巡航展示模式：自动轮播 BGP 事件，连线逐帧追踪，卫星/海缆联动高亮，视觉状态全面统一 |
+| `0.30.0` | feature | `dev` | `pending` | Earth 新增真实地形图层（Terrarium DEM 代理 + 前端瓦片解码着色），设置弹窗支持地形透明度滑块 |
+| `0.29.2` | bugfix | `dev` | `pending` | 修正 Earth 设置弹窗展开表现与系统入口，继续统一液态玻璃 HUD，并校正太阳受光方向 |
+| `0.29.1` | bugfix | `dev` | `pending` | Earth 加载通知条改为队列式单面板显示，brand panel 去框并收敛昼夜与选中态可读性 |
+| `0.29.0` | feature | `dev` | `pending` | Earth 新增天球背景与太阳/月亮位置层，强化昼夜分隔并收口卫星图例与图层面板交互 |
+| `0.28.2` | bugfix | `dev` | `pending` | 修正媒体情报 tab 尺寸记忆与切换锚点逻辑，并清理 docs 根目录遗留旧路径文档 |
+| `0.28.1` | bugfix | `dev` | `pending` | 收口 Earth 媒体情报面板命名与 tab 文案，整理 docs 分组并归档已完成/废弃计划文档 |
+| `0.28.0` | feature | `dev` | `pending` | 合并 Earth 媒体情报面板，整合新闻直播与态势聚合 tab，并稳定 TV/news 的 reform、resize 与共享 HUD 行为 |
+| `0.27.8` | bugfix | `dev` | `pending` | 统一 Earth HUD 默认折叠逻辑，修复图例与图层面板箭头和底边阈值行为 |
+| `0.27.7` | bugfix | `dev` | `pending` | 修复电视直播源编辑持久化问题，清理表格空白占位列并统一可折叠操作列 |
+| `0.27.6` | improvement | `dev` | `pending` | BGP/用户表格滚动条修复，Playground 响应式按钮与输入框收起优化 |
+| `0.27.5` | bugfix | `dev` | `pending` | 统一控制台自定义滚动条，修复 alerts/BGP 响应式滚动与采集进度完成态显示 |
+| `0.27.4` | improvement | `dev` | — | info-card 懒加载动态挂载，页面初始不再有隐藏节点 |
+| `0.27.3` | improvement | `dev` | — | TV panel 折叠方向稳定、视频跳动修复、图例折叠按钮修复、搜索图标调整 |
+| `0.27.2` | improvement | `dev` | — | 修复 brand copy 宽度问题，提取 --brand-copy-width CSS 变量 |
+| `0.27.1` | improvement | `dev` | — | HUD 面板拖拽 L 形边界约束、brand 组件整体缩放、图层面板宽度优化、搜索叉叉修复 |
 | `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 |

frontend/package.json

@@ -1,6 +1,6 @@
 {
   "name": "planet-frontend",
-  "version": "0.27.0",
+  "version": "0.40.1",
   "private": true,
   "packageManager": "bun@1",
   "dependencies": {
@@ -25,8 +25,8 @@
     "vite": "^5.0.10"
   },
   "scripts": {
-    "dev": "vite",
-    "build": "tsc && vite build",
-    "preview": "vite preview"
+    "dev": "bun ./node_modules/vite/bin/vite.js",
+    "build": "bun x tsc && bun ./node_modules/vite/bin/vite.js build",
+    "preview": "bun ./node_modules/vite/bin/vite.js preview"
   }
 }

frontend/public/earth/assets/celestial/bright-stars.json

@@ -0,0 +1,30 @@
+[
+  { "id": 32349, "name": "Sirius", "raDeg": 101.2875, "decDeg": -16.7161, "mag": -1.46, "colorIndex": 0.00 },
+  { "id": 30438, "name": "Canopus", "raDeg": 95.9879, "decDeg": -52.6957, "mag": -0.74, "colorIndex": 0.15 },
+  { "id": 69673, "name": "Arcturus", "raDeg": 213.9153, "decDeg": 19.1824, "mag": -0.05, "colorIndex": 1.23 },
+  { "id": 71683, "name": "Alpha Centauri", "raDeg": 219.9021, "decDeg": -60.8339, "mag": -0.01, "colorIndex": 0.71 },
+  { "id": 91262, "name": "Vega", "raDeg": 279.2347, "decDeg": 38.7837, "mag": 0.03, "colorIndex": 0.00 },
+  { "id": 24608, "name": "Capella", "raDeg": 79.1723, "decDeg": 45.9979, "mag": 0.08, "colorIndex": 0.80 },
+  { "id": 24436, "name": "Rigel", "raDeg": 78.6345, "decDeg": -8.2016, "mag": 0.12, "colorIndex": -0.03 },
+  { "id": 37279, "name": "Procyon", "raDeg": 114.8255, "decDeg": 5.2250, "mag": 0.34, "colorIndex": 0.42 },
+  { "id": 7588, "name": "Achernar", "raDeg": 24.4286, "decDeg": -57.2368, "mag": 0.46, "colorIndex": -0.16 },
+  { "id": 27989, "name": "Betelgeuse", "raDeg": 88.7929, "decDeg": 7.4071, "mag": 0.50, "colorIndex": 1.85 },
+  { "id": 68702, "name": "Hadar", "raDeg": 210.9559, "decDeg": -60.3731, "mag": 0.61, "colorIndex": -0.23 },
+  { "id": 97649, "name": "Altair", "raDeg": 297.6958, "decDeg": 8.8683, "mag": 0.76, "colorIndex": 0.22 },
+  { "id": 60718, "name": "Acrux", "raDeg": 186.6490, "decDeg": -63.0991, "mag": 0.77, "colorIndex": -0.24 },
+  { "id": 21421, "name": "Aldebaran", "raDeg": 68.9800, "decDeg": 16.5093, "mag": 0.85, "colorIndex": 1.54 },
+  { "id": 65474, "name": "Spica", "raDeg": 201.2983, "decDeg": -11.1614, "mag": 0.98, "colorIndex": -0.23 },
+  { "id": 80763, "name": "Antares", "raDeg": 247.3519, "decDeg": -26.4320, "mag": 1.06, "colorIndex": 1.83 },
+  { "id": 37826, "name": "Pollux", "raDeg": 116.3289, "decDeg": 28.0262, "mag": 1.14, "colorIndex": 1.00 },
+  { "id": 113368, "name": "Fomalhaut", "raDeg": 344.4128, "decDeg": -29.6222, "mag": 1.16, "colorIndex": 0.09 },
+  { "id": 102098, "name": "Deneb", "raDeg": 310.3579, "decDeg": 45.2803, "mag": 1.25, "colorIndex": 0.09 },
+  { "id": 49669, "name": "Regulus", "raDeg": 152.0929, "decDeg": 11.9672, "mag": 1.35, "colorIndex": -0.11 },
+  { "id": 65477, "name": "Mimosa", "raDeg": 191.9303, "decDeg": -59.6888, "mag": 1.25, "colorIndex": -0.23 },
+  { "id": 33579, "name": "Alphard", "raDeg": 141.8969, "decDeg": -8.6586, "mag": 1.98, "colorIndex": 1.44 },
+  { "id": 21444, "name": "Bellatrix", "raDeg": 81.2828, "decDeg": 6.3497, "mag": 1.64, "colorIndex": -0.22 },
+  { "id": 25336, "name": "Elnath", "raDeg": 81.5729, "decDeg": 28.6075, "mag": 1.65, "colorIndex": -0.13 },
+  { "id": 26311, "name": "Alnilam", "raDeg": 84.0534, "decDeg": -1.2019, "mag": 1.69, "colorIndex": -0.19 },
+  { "id": 26727, "name": "Alnitak", "raDeg": 85.1897, "decDeg": -1.9426, "mag": 1.77, "colorIndex": -0.19 },
+  { "id": 25930, "name": "Saiph", "raDeg": 86.9391, "decDeg": -9.6696, "mag": 2.06, "colorIndex": -0.20 },
+  { "id": 58001, "name": "Alioth", "raDeg": 193.5073, "decDeg": 55.9598, "mag": 1.76, "colorIndex": -0.02 }
+]

frontend/public/earth/css/base.css

@@ -8,6 +8,10 @@
 
 :root {
   --hud-scale: 1;
+  --safe-top: env(safe-area-inset-top, 0px);
+  --safe-right: env(safe-area-inset-right, 0px);
+  --safe-bottom: env(safe-area-inset-bottom, 0px);
+  --safe-left: env(safe-area-inset-left, 0px);
   --hud-offset: calc(20px * var(--hud-scale));
   --hud-radius: calc(22px * var(--hud-scale));
   --hud-panel-padding: calc(18px * var(--hud-scale));
@@ -19,6 +23,7 @@
   --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-panel-header-title-size: calc(0.82rem * 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);
@@ -55,6 +60,9 @@ body,
   height: 100%;
 }
 
+/* Ensure [hidden] always wins over component display rules */
+[hidden] { display: none !important; }
+
 body.earth-page {
   font-family: "Segoe UI", Tahoma, Geneva, Verdana, sans-serif;
   background-color: #0a0a1a;
@@ -62,12 +70,23 @@ body.earth-page {
   overflow: hidden;
 }
 
+html.is-globe-dragging,
+body.earth-page.is-globe-dragging,
+body.earth-page.is-globe-dragging * {
+  user-select: none !important;
+  -webkit-user-select: none !important;
+}
+
 .earth-app {
   position: relative;
   width: 100vw;
   height: 100vh;
 }
 
+.earth-app canvas {
+  touch-action: none;
+}
+
 .earth-app.dragging {
   cursor: grabbing;
 }
@@ -79,59 +98,17 @@ body.earth-page {
   pointer-events: none;
 }
 
-.earth-loading {
-  position: absolute;
-  top: 50%;
-  left: 50%;
-  transform: translate(-50%, -50%);
-  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;
-  color: #4db8ff;
-}
-
-.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%;
-  animation: spin 1s linear infinite;
-}
-
-@keyframes spin {
-  0% {
-    transform: rotate(0deg);
+@keyframes earthLoadingPulse {
+  0%,
+  80%,
+  100% {
+    opacity: 0.28;
+    transform: scale(0.78);
   }
 
-  100% {
-    transform: rotate(360deg);
+  40% {
+    opacity: 1;
+    transform: scale(1);
   }
 }
 

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

@@ -28,20 +28,22 @@
 
 .stats-kicker {
   color: var(--hud-text-soft);
-  font-size: calc(0.64rem * var(--hud-scale));
-  letter-spacing: 0.16em;
-  text-transform: uppercase;
+  font-size: var(--hud-panel-header-title-size);
+  font-weight: 600;
+  letter-spacing: 0.01em;
+  line-height: 1.2;
 }
 
-/* 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));
+  align-self: auto;
+  width: auto;
+  height: auto;
+  min-width: 0;
+  padding: calc(7px * var(--hud-scale));
 }
 
 .stats-drag-bar .hud-panel-close .material-symbols-rounded {
-  font-size: calc(12px * var(--hud-scale));
+  font-size: calc(16px * var(--hud-scale));
 }
 
 /* ── 2-column KPI grid ────────────────────────────────────────── */
@@ -131,3 +133,20 @@
   right: var(--hud-offset);
   transform: translate(calc(100% - var(--hud-offset)), calc(-100% + var(--hud-offset)));
 }
+
+.layout-mode-mobile .hud-panel-stats {
+  position: fixed;
+  top: calc(8px + var(--safe-top));
+  right: 8px;
+  width: min(180px, calc(100vw - 16px));
+  z-index: 205;
+}
+
+.layout-mode-mobile.earth-search-open .hud-panel-stats,
+.layout-mode-mobile.earth-settings-open .hud-panel-stats,
+.layout-mode-mobile.earth-media-open .hud-panel-stats,
+.layout-mode-mobile.earth-info-open .hud-panel-stats {
+  opacity: 0;
+  pointer-events: none;
+  transform: translateY(-12px);
+}

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

@@ -24,64 +24,87 @@
 /* ── Brand panel ──────────────────────────────────────────────── */
 
 .hud-panel-brand {
-  border-radius: 0;
-  padding: calc(12px * var(--hud-scale)) calc(14px * var(--hud-scale));
+  --brand-scale: 0.88;
+  --brand-copy-width: 160px;
+  padding: calc(10px * var(--hud-scale)) calc(4px * var(--hud-scale)) calc(12px * var(--hud-scale)) 0;
   display: flex;
   align-items: center;
-  justify-content: center;
+  justify-content: flex-start;
   /* Reserve full panel height before brand images load */
   min-height: calc(66px * var(--hud-scale));
+  background: transparent;
+  border: 0;
+  box-shadow: none;
+  backdrop-filter: none;
+  -webkit-backdrop-filter: none;
+  isolation: isolate;
+}
+
+.hud-panel-brand::before,
+.hud-panel-brand::after {
+  display: none;
 }
 
 .hud-panel-brand .earth-brand {
   display: flex;
   align-items: center;
-  gap: calc(10px * var(--hud-scale));
+  gap: calc(10px * var(--hud-scale) * var(--brand-scale));
   min-width: 0;
+  position: relative;
+}
+
+.hud-panel-brand .earth-brand::after {
+  content: "";
+  position: absolute;
+  inset: calc(4px * var(--hud-scale)) calc(-10px * var(--hud-scale)) calc(6px * var(--hud-scale)) calc(-10px * var(--hud-scale));
+  background:
+    radial-gradient(circle at 18% 50%, rgba(145, 186, 255, 0.14), transparent 32%),
+    linear-gradient(90deg, rgba(145, 186, 255, 0.05), transparent 58%);
+  opacity: 0.75;
+  pointer-events: none;
+  filter: blur(14px);
+  z-index: -1;
 }
 
 .hud-panel-brand .earth-brand__logo {
   display: block;
   flex: 0 0 auto;
-  width: calc(128px * var(--hud-scale));
-  height: calc(128px * var(--hud-scale));
+  width: calc(128px * var(--hud-scale) * var(--brand-scale));
+  height: calc(128px * var(--hud-scale) * var(--brand-scale));
   object-fit: contain;
 }
 
 .hud-panel-brand .earth-brand__copy {
   display: flex;
-  flex: 1 1 auto;
-  min-width: 0;
+  flex: 0 0 auto;
   flex-direction: column;
   justify-content: center;
-  gap: calc(5px * var(--hud-scale));
+  gap: calc(5px * var(--hud-scale) * var(--brand-scale));
+  width: calc(var(--brand-copy-width) * var(--hud-scale) * var(--brand-scale));
 }
 
 .hud-panel-brand .earth-brand__title {
   display: block;
-  width: min(100%, calc(160px * var(--hud-scale)));
+  width: min(100%, calc(var(--brand-copy-width) * var(--hud-scale) * var(--brand-scale)));
   max-width: 100%;
   height: auto;
-  min-height: calc(20px * var(--hud-scale));
+  min-height: calc(20px * var(--hud-scale) * var(--brand-scale));
   object-fit: contain;
 }
 
 .hud-panel-brand .earth-brand__meta {
   display: flex;
   flex-direction: column;
-  gap: calc(2px * var(--hud-scale));
+  gap: calc(2px * var(--hud-scale) * var(--brand-scale));
   width: fit-content;
 }
 
 .hud-panel-brand .earth-brand__subtitle {
   color: var(--hud-text-muted);
-  font-size: calc(0.74rem * var(--hud-scale));
+  font-size: calc(0.74rem * var(--hud-scale) * var(--brand-scale));
   line-height: 1.3;
   font-weight: 500;
   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;
@@ -89,18 +112,24 @@
 
 .hud-panel-brand .earth-brand__description {
   color: var(--hud-text-soft);
-  font-size: calc(0.6rem * var(--hud-scale));
+  font-size: calc(0.6rem * var(--hud-scale) * var(--brand-scale));
   line-height: 1.3;
   letter-spacing: 0.08em;
-  width: fit-content;
-  max-width: 100%;
   overflow: hidden;
   white-space: nowrap;
   text-overflow: ellipsis;
 }
 
+.hud-panel-brand .earth-brand--en {
+  --brand-copy-width: 172px;
+}
+
 .hud-panel-brand .earth-brand--en .earth-brand__title {
-  width: min(100%, calc(172px * var(--hud-scale)));
+  width: min(100%, calc(var(--brand-copy-width) * var(--hud-scale) * var(--brand-scale)));
+}
+
+.hud-panel-brand .earth-brand--en .earth-brand__copy {
+  width: calc(var(--brand-copy-width) * var(--hud-scale) * var(--brand-scale));
 }
 
 .hud-panel-brand .earth-brand--en .earth-brand__subtitle,
@@ -124,12 +153,103 @@
   transition:
     opacity 0.22s ease,
     transform 0.22s ease;
+  visibility: hidden;
+  user-select: none;
+  -webkit-user-select: none;
 }
 
 .hud-panel-info.is-visible {
   opacity: 1;
   transform: scale(1) translateY(0);
   pointer-events: auto;
+  visibility: visible;
+}
+
+.hud-panel-info.hud-panel-info--anchor-stable {
+  transform: none;
+  transition: opacity 0.22s ease;
+}
+
+.hud-panel-info.hud-panel-info--anchor-stable.is-visible {
+  transform: none;
+}
+
+.callout-connector {
+  position: absolute;
+  inset: 0;
+  width: 100%;
+  height: 100%;
+  overflow: visible;
+  opacity: 0;
+  pointer-events: none;
+  transition: opacity 0.22s ease;
+  z-index: 49;
+}
+
+.callout-connector polyline {
+  fill: none;
+  stroke: rgba(255, 255, 255, 0.98);
+  stroke-width: 2.15;
+  stroke-linecap: round;
+  stroke-linejoin: round;
+  filter:
+    drop-shadow(0 0 1px rgba(6, 14, 28, 0.72))
+    drop-shadow(0 0 2px rgba(6, 14, 28, 0.56))
+    drop-shadow(0 0 6px rgba(8, 20, 36, 0.1));
+}
+
+.callout-connector circle {
+  fill: rgba(255, 255, 255, 0.98);
+  stroke: rgba(7, 16, 32, 0.72);
+  stroke-width: 1.0;
+  filter:
+    drop-shadow(0 0 1px rgba(6, 14, 28, 0.72))
+    drop-shadow(0 0 2px rgba(6, 14, 28, 0.54))
+    drop-shadow(0 0 6px rgba(8, 20, 36, 0.1));
+  transform-box: fill-box;
+  transform-origin: center;
+}
+
+.callout-connector.is-visible {
+  opacity: 1;
+}
+
+.callout-connector.is-animating polyline {
+  animation: calloutConnectorDraw 0.42s cubic-bezier(0.22, 1, 0.36, 1) forwards;
+}
+
+.callout-connector.is-animating circle {
+  opacity: 0;
+}
+
+.callout-connector.is-animating circle:first-of-type {
+  animation: calloutConnectorNodeIn 0.14s ease forwards;
+  animation-delay: 0.02s;
+}
+
+.callout-connector.is-animating circle:last-of-type {
+  animation: calloutConnectorNodeIn 0.16s ease forwards;
+  animation-delay: 0.34s;
+}
+
+@keyframes calloutConnectorDraw {
+  from {
+    stroke-dashoffset: var(--connector-length, 0px);
+  }
+  to {
+    stroke-dashoffset: 0px;
+  }
+}
+
+@keyframes calloutConnectorNodeIn {
+  from {
+    opacity: 0;
+    transform: scale(0.72);
+  }
+  to {
+    opacity: 1;
+    transform: scale(1);
+  }
 }
 
 /* ── Info Card ────────────────────────────────────────────────── */
@@ -155,7 +275,7 @@
 .info-card-header h3 {
   flex: 1;
   margin: 0;
-  font-size: calc(0.92rem * var(--hud-scale));
+  font-size: var(--hud-panel-header-title-size);
   color: var(--hud-title);
   font-weight: 600;
   white-space: nowrap;
@@ -165,6 +285,15 @@
 
 .info-card-close {
   flex-shrink: 0;
+  align-self: auto;
+  width: auto;
+  height: auto;
+  min-width: 0;
+  padding: calc(7px * var(--hud-scale));
+}
+
+.info-card-close .material-symbols-rounded {
+  font-size: calc(16px * var(--hud-scale));
 }
 
 .info-card-content {
@@ -174,6 +303,8 @@
   scrollbar-width: thin;
   scrollbar-color: rgba(160, 186, 216, 0.34) transparent;
   pointer-events: auto;
+  user-select: none;
+  -webkit-user-select: none;
 }
 
 .info-card-content::-webkit-scrollbar {
@@ -211,6 +342,8 @@
   cursor: pointer;
   flex-shrink: 0;
   transition: color 0.18s ease;
+  user-select: none;
+  -webkit-user-select: none;
 }
 
 .info-card-label:hover {
@@ -225,6 +358,8 @@
   text-align: right;
   max-width: calc(180px * var(--hud-scale));
   word-break: break-word;
+  user-select: none;
+  -webkit-user-select: none;
 }
 
 /* Type-specific header accent colors */
@@ -246,8 +381,151 @@
 }
 .info-card.bgp .info-card-header h3 { color: var(--hud-accent-strong); }
 
+.info-card.news .info-card-header {
+  background: rgba(255, 196, 92, 0.12);
+  border-bottom-color: rgba(255, 196, 92, 0.16);
+}
+.info-card.news .info-card-header h3 { color: #ffd77a; }
+
+.info-card.news .info-card-content {
+  padding-top: calc(10px * var(--hud-scale));
+  padding-bottom: calc(12px * var(--hud-scale));
+}
+
+.info-card-news-layout {
+  display: grid;
+  gap: calc(10px * var(--hud-scale));
+}
+
+.info-card-news-kicker {
+  color: rgba(255, 215, 122, 0.82);
+  font-size: calc(0.62rem * var(--hud-scale));
+  font-weight: 700;
+  letter-spacing: 0.18em;
+  text-transform: uppercase;
+}
+
+.info-card-news-title {
+  color: var(--hud-title);
+  font-size: calc(0.96rem * var(--hud-scale));
+  line-height: 1.45;
+  font-weight: 700;
+  text-wrap: balance;
+}
+
+.info-card-news-summary-shell {
+  position: relative;
+  padding: calc(10px * var(--hud-scale)) calc(12px * var(--hud-scale));
+  border: 1px solid rgba(255, 215, 122, 0.12);
+  background:
+    linear-gradient(180deg, rgba(255, 215, 122, 0.05), rgba(255, 255, 255, 0.02)),
+    radial-gradient(circle at top left, rgba(120, 180, 255, 0.08), transparent 56%),
+    rgba(7, 15, 29, 0.42);
+  box-shadow:
+    inset 0 1px 0 rgba(255, 255, 255, 0.05),
+    0 10px 28px rgba(0, 0, 0, 0.16);
+  overflow: hidden;
+}
+
+.info-card-news-summary-shell::before {
+  content: "";
+  position: absolute;
+  inset: 0;
+  background:
+    linear-gradient(90deg, transparent 0%, rgba(122, 180, 255, 0.12) 50%, transparent 100%);
+  opacity: 0.42;
+  transform: translateX(-100%);
+  animation: infoCardNewsScan 3.2s linear infinite;
+  pointer-events: none;
+}
+
+.info-card-news-summary-label {
+  color: rgba(188, 212, 238, 0.72);
+  font-size: calc(0.6rem * var(--hud-scale));
+  font-weight: 700;
+  letter-spacing: 0.16em;
+  text-transform: uppercase;
+  margin-bottom: calc(6px * var(--hud-scale));
+}
+
+.info-card-news-summary {
+  position: relative;
+  color: #d7e6f7;
+  font-size: calc(0.8rem * var(--hud-scale));
+  line-height: 1.65;
+  min-height: calc(4.8em * var(--hud-scale));
+  white-space: pre-wrap;
+  word-break: break-word;
+}
+
+.info-card-news-summary.is-typing::after {
+  content: "";
+  display: inline-block;
+  width: 0.58em;
+  height: 1.05em;
+  margin-left: 0.16em;
+  vertical-align: -0.14em;
+  background: linear-gradient(180deg, rgba(255, 215, 122, 0.96), rgba(122, 180, 255, 0.78));
+  box-shadow: 0 0 10px rgba(255, 215, 122, 0.28);
+  animation: infoCardNewsCaret 0.9s steps(1, end) infinite;
+}
+
+@keyframes infoCardNewsScan {
+  from {
+    transform: translateX(-100%);
+  }
+  to {
+    transform: translateX(100%);
+  }
+}
+
+@keyframes infoCardNewsCaret {
+  0%, 49% {
+    opacity: 1;
+  }
+  50%, 100% {
+    opacity: 0;
+  }
+}
+
 /* ── Layout-expanded: slide left column off-screen ────────────── */
 
 .earth-app.layout-expanded .earth-left-column {
   transform: translate(calc(-100% + var(--hud-offset)), 0);
 }
+
+.layout-mode-mobile .earth-left-column {
+  top: calc(8px + var(--safe-top));
+  left: 8px;
+  max-width: min(300px, calc(100vw - 16px));
+}
+
+.layout-mode-mobile .hud-panel-info {
+  position: fixed;
+  left: 8px !important;
+  right: 8px !important;
+  top: auto !important;
+  bottom: calc(84px + var(--safe-bottom)) !important;
+  width: auto;
+  max-width: none;
+  max-height: min(58vh, 520px);
+  z-index: 240;
+}
+
+.layout-mode-mobile .info-card-header {
+  cursor: default;
+}
+
+.layout-mode-mobile .info-card-content {
+  max-height: min(46vh, 420px);
+}
+
+.layout-mode-mobile .info-card-property {
+  flex-direction: column;
+  align-items: stretch;
+}
+
+.layout-mode-mobile .info-card-value {
+  max-width: none;
+  text-align: left;
+}

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

@@ -1,13 +1,13 @@
 /* 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 */
+  /* Lives inside .earth-left-column — narrower than brand panel intentionally */
   border-radius: 0;
   padding: 0;
-  width: 100%;
+  width: calc(260px * var(--hud-scale));
   z-index: 10;
   overflow: hidden;
-  margin-top: calc(6px * var(--hud-scale));
+  margin-top: calc(12px * var(--hud-scale));
 }
 
 /* ── Header / drag handle ─────────────────────────────────────── */
@@ -38,8 +38,8 @@
 .layer-panel-title {
   flex: 1 1 auto;
   margin: 0;
-  color: var(--hud-title);
-  font-size: calc(0.82rem * var(--hud-scale));
+  color: var(--hud-text-soft);
+  font-size: var(--hud-panel-header-title-size);
   font-weight: 600;
   letter-spacing: 0.04em;
   line-height: 1.2;
@@ -51,55 +51,71 @@
   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;
+  padding: calc(7px * var(--hud-scale));
+  border: 1px solid transparent;
   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;
+  transition:
+    background 0.18s ease,
+    border-color 0.18s ease,
+    color 0.18s ease,
+    transform 0.18s ease,
+    opacity 0.18s ease;
 }
 
 .layer-panel-btn:hover {
-  background: rgba(255, 255, 255, 0.07);
-  color: var(--hud-text);
+  background: rgba(255, 255, 255, 0.08);
+  border-color: rgba(225, 239, 255, 0.14);
+  color: var(--hud-accent-strong);
 }
 
 .layer-panel-btn .material-symbols-rounded {
-  font-size: calc(14px * var(--hud-scale));
+  font-size: calc(16px * 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);
+  transition: color 0.18s ease;
+  font-variation-settings: 'FILL' 0, 'wght' 400, 'GRAD' 0, 'opsz' 20;
 }
 
 /* ── Search bar ───────────────────────────────────────────────── */
 
 .layer-panel-search {
+  padding: calc(6px * var(--hud-scale)) calc(8px * var(--hud-scale));
+  border-bottom: 1px solid var(--hud-line);
+}
+
+.layer-panel-search-box {
   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);
+  padding: calc(5px * var(--hud-scale)) calc(8px * var(--hud-scale));
+  border: 1px solid rgba(201, 225, 247, 0.14);
+  border-radius: calc(8px * var(--hud-scale));
+  background: rgba(255, 255, 255, 0.04);
+  transition: border-color 0.18s ease;
+  box-sizing: border-box;
+  height: calc(38px * var(--hud-scale));
+}
+
+.layer-panel-search-box:focus-within {
+  border-color: rgba(201, 225, 247, 0.28);
 }
 
 .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-icon.material-symbols-rounded {
+  font-size: calc(20px * var(--hud-scale));
+  font-variation-settings: 'FILL' 0, 'wght' 400, 'GRAD' 0, 'opsz' 20;
+}
+
 .layer-panel-search-input {
   flex: 1 1 auto;
   min-width: 0;
@@ -144,6 +160,23 @@
 .layer-panel-list {
   display: flex;
   flex-direction: column;
+  max-height: calc(5 * (56px * var(--hud-scale)));
+  overflow-y: auto;
+  scrollbar-width: thin;
+  scrollbar-color: rgba(160, 186, 216, 0.34) transparent;
+}
+
+.layer-panel-list::-webkit-scrollbar {
+  width: 4px;
+}
+
+.layer-panel-list::-webkit-scrollbar-track {
+  background: transparent;
+}
+
+.layer-panel-list::-webkit-scrollbar-thumb {
+  background: linear-gradient(180deg, rgba(210, 225, 242, 0.2), rgba(126, 154, 185, 0.28));
+  border-radius: 999px;
 }
 
 .layer-row {
@@ -153,6 +186,7 @@
   padding: calc(9px * var(--hud-scale)) calc(10px * var(--hud-scale));
   border-bottom: 1px solid var(--hud-line);
   transition: background 0.14s ease;
+  min-height: calc(56px * var(--hud-scale));
 }
 
 .layer-row:last-child {
@@ -219,6 +253,7 @@
   border: none;
   background: transparent;
   cursor: pointer;
+  transition: opacity 0.18s ease;
 }
 
 .layer-row-toggle-track {
@@ -231,6 +266,11 @@
   transition: background 0.18s ease, border-color 0.18s ease;
 }
 
+.layer-row-toggle:disabled {
+  cursor: progress;
+  opacity: 1;
+}
+
 /* Thumb */
 .layer-row-toggle-track::after {
   content: "";
@@ -245,6 +285,24 @@
   transition: transform 0.18s ease, background 0.18s ease;
 }
 
+.layer-row-toggle.is-loading .layer-row-toggle-track {
+  background: linear-gradient(
+    90deg,
+    rgba(104, 147, 221, 0.38),
+    rgba(143, 185, 255, 0.72),
+    rgba(104, 147, 221, 0.38)
+  );
+  background-size: 180% 100%;
+  border-color: rgba(223, 236, 252, 0.28);
+  animation: layer-toggle-loading-track 1.2s linear infinite;
+}
+
+.layer-row-toggle.is-loading .layer-row-toggle-track::after {
+  background: #f0f6ff;
+  transform: translateX(calc(7px * var(--hud-scale)));
+  animation: layer-toggle-loading-thumb 1s ease-in-out infinite;
+}
+
 /* 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));
@@ -256,5 +314,55 @@
   transform: translateX(calc(14px * var(--hud-scale)));
 }
 
+@keyframes layer-toggle-loading-track {
+  0% {
+    background-position: 0% 50%;
+  }
+  100% {
+    background-position: 180% 50%;
+  }
+}
+
+@keyframes layer-toggle-loading-thumb {
+  0%,
+  100% {
+    box-shadow: 0 2px 6px rgba(1, 8, 18, 0.3), 0 0 0 rgba(174, 205, 255, 0.18);
+  }
+  50% {
+    box-shadow: 0 2px 6px rgba(1, 8, 18, 0.3), 0 0 calc(10px * var(--hud-scale)) rgba(174, 205, 255, 0.42);
+  }
+}
+
 /* Layout-expanded: layer panel slides off with .earth-left-column — no
    individual rule needed since the whole column translates together. */
+
+.layout-mode-mobile .hud-panel-layers {
+  position: fixed;
+  left: 12px;
+  right: 12px;
+  bottom: calc(88px + var(--safe-bottom));
+  width: auto;
+  max-height: min(60vh, 520px);
+  margin-top: 0;
+  z-index: 220;
+  transform: translateY(calc(100% + 28px));
+  opacity: 0;
+  pointer-events: none;
+  transition: transform 0.24s ease, opacity 0.2s ease;
+}
+
+.layout-mode-mobile .hud-panel-layers.is-mobile-open {
+  transform: translateY(0);
+  opacity: 1;
+  pointer-events: auto;
+}
+
+.layout-mode-mobile .layer-panel-body {
+  max-height: min(52vh, 460px);
+  overflow: auto;
+}
+
+.layout-mode-mobile .layer-panel-list {
+  max-height: none;
+  overflow-y: visible;
+}

frontend/public/earth/css/legend.css

@@ -27,38 +27,38 @@
   cursor: grabbing;
 }
 
-/* ── Mode tabs ────────────────────────────────────────────────── */
+/* ── Current mode label ───────────────────────────────────────── */
 
-.legend-tabs {
+.legend-current {
   display: flex;
-  gap: calc(2px * var(--hud-scale));
+  align-items: center;
+  gap: calc(6px * 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;
+.legend-title {
+  flex: 0 0 auto;
+  color: var(--hud-text-soft);
+  font-size: var(--hud-panel-header-title-size);
+  font-weight: 600;
+  letter-spacing: 0.01em;
+  line-height: 1.2;
   white-space: nowrap;
 }
 
-.legend-tab:hover {
-  background: rgba(255, 255, 255, 0.06);
-  color: var(--hud-text);
-}
-
-.legend-tab--active {
+.legend-current-label {
+  display: inline-flex;
+  align-items: center;
+  min-width: 0;
+  padding: calc(3px * var(--hud-scale)) calc(7px * var(--hud-scale));
+  border-radius: calc(4px * var(--hud-scale));
+  border: 1px solid rgba(120, 180, 255, 0.2);
   background: rgba(120, 180, 255, 0.12);
-  border-color: rgba(120, 180, 255, 0.2);
   color: var(--hud-accent-strong);
+  font-size: calc(0.68rem * var(--hud-scale));
+  letter-spacing: 0.08em;
+  white-space: nowrap;
 }
 
 /* ── Bar action buttons ───────────────────────────────────────── */
@@ -66,7 +66,7 @@
 .legend-bar-actions {
   display: flex;
   align-items: center;
-  gap: calc(2px * var(--hud-scale));
+  gap: var(--hud-gap-xs);
   flex-shrink: 0;
 }
 
@@ -74,54 +74,17 @@
   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;
+  --hud-body-collapse-gap: calc(4px * var(--hud-scale));
+  --hud-body-max-height: calc(220px * var(--hud-scale));
 }
 
 /* ── Item list ────────────────────────────────────────────────── */
@@ -175,3 +138,24 @@
   bottom: var(--hud-offset);
   transform: translate(calc(-100% + var(--hud-offset)), calc(100% - var(--hud-offset)));
 }
+
+.layout-mode-mobile .hud-panel-legend {
+  position: fixed;
+  left: 8px;
+  bottom: calc(84px + var(--safe-bottom));
+  width: min(172px, calc(100vw - 16px));
+  z-index: 205;
+}
+
+.layout-mode-mobile .legend-list {
+  max-height: min(20vh, 180px);
+}
+
+.layout-mode-mobile.earth-search-open .hud-panel-legend,
+.layout-mode-mobile.earth-settings-open .hud-panel-legend,
+.layout-mode-mobile.earth-media-open .hud-panel-legend,
+.layout-mode-mobile.earth-info-open .hud-panel-legend {
+  opacity: 0;
+  pointer-events: none;
+  transform: translateY(12px);
+}

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

@@ -0,0 +1,191 @@
+/* news-panel.css */
+
+.news-panel-title-row {
+  display: flex;
+  align-items: center;
+  min-width: 0;
+  flex: 1 1 auto;
+}
+
+.news-region-chip {
+  --news-accent: #d6e6ff;
+  border: 1px solid color-mix(in srgb, var(--news-accent) 46%, transparent);
+  border-radius: 999px;
+  padding: calc(3px * var(--hud-scale)) calc(8px * var(--hud-scale));
+  color: color-mix(in srgb, var(--news-accent) 82%, white);
+  background: color-mix(in srgb, var(--news-accent) 12%, transparent);
+  font-size: calc(0.62rem * var(--hud-scale));
+  letter-spacing: 0.08em;
+  text-transform: uppercase;
+}
+
+.news-panel-subtitle {
+  color: var(--hud-text-soft);
+  font-size: calc(0.7rem * var(--hud-scale));
+}
+
+.news-panel-body {
+  display: flex;
+  flex: 1 1 auto;
+  flex-direction: column;
+  gap: var(--hud-gap-sm);
+  min-height: 0;
+  overflow: hidden;
+}
+
+.news-panel-focus {
+  display: grid;
+  grid-template-columns: 1fr auto;
+  gap: calc(10px * var(--hud-scale));
+  padding: calc(12px * var(--hud-scale));
+  border-radius: calc(16px * var(--hud-scale));
+  background:
+    radial-gradient(circle at 16% 18%, rgba(123, 205, 255, 0.12), transparent 36%),
+    linear-gradient(180deg, rgba(255, 255, 255, 0.04), rgba(104, 166, 232, 0.04));
+  border: 1px solid rgba(205, 231, 255, 0.08);
+}
+
+.news-focus-kicker,
+.news-board-status {
+  color: var(--hud-text-soft);
+  font-size: calc(0.66rem * var(--hud-scale));
+  letter-spacing: 0.08em;
+  text-transform: uppercase;
+}
+
+.news-focus-label {
+  margin-top: calc(4px * var(--hud-scale));
+  color: var(--hud-text);
+  font-size: calc(1rem * var(--hud-scale));
+  font-weight: 600;
+}
+
+.news-focus-coords {
+  margin-top: calc(3px * var(--hud-scale));
+  color: var(--hud-text-muted);
+  font-size: calc(0.74rem * var(--hud-scale));
+}
+
+.news-source-count {
+  align-self: start;
+  color: var(--hud-accent-strong);
+  font-size: calc(0.72rem * var(--hud-scale));
+}
+
+.news-board {
+  display: flex;
+  flex: 1 1 auto;
+  flex-direction: column;
+  gap: var(--hud-gap-sm);
+  min-height: 0;
+  overflow: hidden;
+}
+
+.news-board-list {
+  display: flex;
+  flex: 1 1 auto;
+  flex-direction: column;
+  gap: calc(8px * var(--hud-scale));
+  min-height: 0;
+  max-height: none;
+  overflow-y: auto;
+  padding-right: calc(4px * var(--hud-scale));
+  scrollbar-width: thin;
+  scrollbar-color: rgba(160, 220, 255, 0.36) transparent;
+}
+
+.news-board-list::-webkit-scrollbar {
+  width: 6px;
+}
+
+.news-board-list::-webkit-scrollbar-track {
+  background: transparent;
+}
+
+.news-board-list::-webkit-scrollbar-thumb {
+  background: linear-gradient(180deg, rgba(210, 237, 255, 0.24), rgba(110, 176, 255, 0.28));
+  border-radius: 999px;
+}
+
+.news-story-card {
+  display: flex;
+  flex-direction: column;
+  gap: calc(8px * var(--hud-scale));
+  text-decoration: none;
+  padding: calc(12px * var(--hud-scale));
+  border-radius: calc(16px * var(--hud-scale));
+  border: 1px solid rgba(201, 225, 247, 0.08);
+  background:
+    linear-gradient(180deg, rgba(255, 255, 255, 0.05), rgba(92, 151, 218, 0.03));
+  transition:
+    border-color 0.18s ease,
+    background 0.18s ease,
+    transform 0.18s ease;
+}
+
+.news-story-card:hover {
+  border-color: rgba(214, 235, 255, 0.16);
+  background:
+    linear-gradient(180deg, rgba(255, 255, 255, 0.08), rgba(92, 151, 218, 0.06));
+  transform: translateY(-1px);
+}
+
+.news-story-card--focus {
+  border-color: rgba(127, 219, 255, 0.22);
+  box-shadow: 0 0 0 1px rgba(122, 214, 255, 0.08) inset;
+}
+
+.news-story-card--cruise {
+  border-color: rgba(255, 213, 128, 0.42);
+  background:
+    linear-gradient(180deg, rgba(255, 248, 220, 0.1), rgba(255, 184, 77, 0.08));
+  box-shadow:
+    0 0 0 1px rgba(255, 213, 128, 0.18) inset,
+    0 0 18px rgba(255, 184, 77, 0.12);
+}
+
+.news-story-meta,
+.news-story-tags {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  gap: calc(8px * var(--hud-scale));
+  flex-wrap: wrap;
+}
+
+.news-story-source,
+.news-story-time,
+.news-story-tag {
+  color: var(--hud-text-soft);
+  font-size: calc(0.66rem * var(--hud-scale));
+}
+
+.news-story-source {
+  color: var(--hud-accent-strong);
+}
+
+.news-story-title {
+  color: var(--hud-text);
+  font-size: calc(0.9rem * var(--hud-scale));
+  font-weight: 600;
+  line-height: 1.4;
+}
+
+.news-story-summary {
+  color: var(--hud-text-muted);
+  font-size: calc(0.74rem * var(--hud-scale));
+  line-height: 1.45;
+}
+
+.news-story-tag {
+  border-radius: 999px;
+  padding: calc(3px * var(--hud-scale)) calc(7px * var(--hud-scale));
+  background: rgba(255, 255, 255, 0.04);
+}
+
+.news-board-empty {
+  color: var(--hud-text-muted);
+  font-size: calc(0.82rem * var(--hud-scale));
+  line-height: 1.5;
+  padding: calc(16px * var(--hud-scale)) calc(4px * var(--hud-scale));
+}

frontend/public/earth/css/toolbar.css

@@ -1,4 +1,4 @@
-/* toolbar.css - bottom dock and floating toolbar primitives */
+/* toolbar.css - orbital hub toolbar */
 
 .earth-toolbar-group {
   position: absolute;
@@ -6,10 +6,10 @@
   left: 50%;
   transform: translateX(-50%);
   display: flex;
-  flex-direction: row;
   align-items: center;
   justify-content: center;
   z-index: 200;
+  pointer-events: none;
 }
 
 .earth-toolbar-group,
@@ -24,119 +24,170 @@
 }
 
 .earth-toolbar {
+  --toolbar-scale: 1;
+  --toolbar-orb-size: calc(46px * var(--toolbar-scale));
+  --toolbar-hub-size: calc(58px * var(--toolbar-scale));
+  --toolbar-arc-width: calc(420px * var(--toolbar-scale));
+  --toolbar-arc-height: calc(160px * var(--toolbar-scale));
+  --toolbar-inner-arc-width: calc(260px * var(--toolbar-scale));
+  --toolbar-inner-arc-height: calc(56px * var(--toolbar-scale));
   position: relative;
+  width: min(620px, calc(100vw - 40px));
+  height: calc(200px * var(--toolbar-scale));
   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 {
+.earth-toolbar-cluster {
+  position: relative;
+  width: 100%;
+  height: 100%;
+  pointer-events: none;
+}
+
+.earth-toolbar-cluster::before {
+  content: "";
+  position: absolute;
+  left: 50%;
+  bottom: calc(14px * var(--toolbar-scale));
+  width: var(--toolbar-arc-width);
+  height: var(--toolbar-arc-height);
+  transform: translateX(-50%);
+  border-radius: 50%;
+  border: 1px solid rgba(145, 186, 255, 0.08);
+  border-bottom-color: transparent;
+  background:
+    radial-gradient(circle at 50% 100%, rgba(145, 186, 255, 0.05), transparent 58%);
+  opacity: 0.9;
+  mask: linear-gradient(180deg, rgba(0, 0, 0, 0.82), transparent 86%);
+  pointer-events: none;
+}
+
+.earth-toolbar-orb {
+  position: absolute;
+  left: 50%;
+  bottom: calc(30px * var(--toolbar-scale));
+  transform: translate(-50%, -50%);
+}
+
+.earth-toolbar-hub {
+  position: absolute;
+  left: 50%;
+  bottom: calc(8px * var(--toolbar-scale));
+  transform: translateX(-50%);
+}
+
+.earth-toolbar-orb {
+  pointer-events: none;
+  opacity: 1;
+  transition:
+    transform 0.36s cubic-bezier(0.34, 1.15, 0.64, 1),
+    opacity 0.24s ease;
+}
+
+.earth-toolbar-cluster.is-expanded .earth-toolbar-orb {
+  transform: translate(calc(-50% + var(--orb-x)), calc(-50% + var(--orb-y)));
+}
+
+.earth-toolbar-cluster.is-collapsed .earth-toolbar-orb {
+  transform: translate(-50%, -50%) scale(0.42);
+  opacity: 0;
+}
+
+.earth-toolbar-orb > *,
+.earth-toolbar-hub > * {
+  pointer-events: auto;
+}
+
+.earth-toolbar-orb:has(#layer-action) {
+  display: none;
+}
+
+.layout-mode-mobile .earth-toolbar-group {
+  display: none;
+}
+
+.earth-toolbar-cluster.is-collapsed .earth-toolbar-orb > * {
+  pointer-events: none;
+}
+
+.earth-toolbar-hub > * {
+  pointer-events: auto;
+}
+
+.earth-toolbar-orb > .liquid-glass-surface {
+  animation: floatDock 4.6s ease-in-out infinite;
+  animation-delay: var(--orb-delay, 0s);
+}
+
+.earth-toolbar-cluster.is-dock-engaged .earth-toolbar-orb > .liquid-glass-surface {
+  animation-play-state: paused;
+}
+
+.earth-toolbar-btn,
+.earth-toolbar-hub-btn {
   position: relative;
   width: 28px;
   height: 28px;
   border: none;
-  border-radius: 0;
   background: transparent;
-  color: #4db8ff;
-  font-size: 14px;
+  color: var(--hud-text-soft);
+  font-size: calc(14px * var(--toolbar-scale));
   cursor: pointer;
-  display: flex;
+  display: inline-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;
+  width: var(--toolbar-orb-size);
+  height: var(--toolbar-orb-size);
+  min-width: var(--toolbar-orb-size);
+  min-height: var(--toolbar-orb-size);
   border-radius: 50%;
   overflow: hidden;
 }
 
-.earth-toolbar-btn:not(.liquid-glass-surface)::after {
-  content: none;
+.earth-toolbar-hub-btn {
+  width: var(--toolbar-hub-size);
+  height: var(--toolbar-hub-size);
+  min-width: var(--toolbar-hub-size);
+  min-height: var(--toolbar-hub-size);
+  border-radius: 50%;
+  overflow: hidden;
+  color: var(--hud-title);
+}
+
+.earth-toolbar-btn .icon,
+.earth-toolbar-hub-btn .material-symbols-rounded {
+  position: relative;
+  z-index: 1;
+  line-height: 1;
 }
 
 .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;
+.earth-toolbar-btn .material-symbols-rounded,
+.earth-toolbar-hub-btn .material-symbols-rounded {
+  font-size: calc(21px * var(--toolbar-scale));
   line-height: 1;
   font-variation-settings:
     'FILL' 0,
@@ -154,28 +205,6 @@
   -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;
@@ -190,12 +219,14 @@
   position: relative;
   isolation: isolate;
   transform-style: preserve-3d;
+  transform-origin: center center;
+  will-change: transform, box-shadow;
   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);
+    radial-gradient(circle at var(--glow-x) var(--glow-y), rgba(255, 255, 255, 0.12), transparent 34%),
+    radial-gradient(circle at 50% 118%, rgba(255, 255, 255, 0.05), transparent 30%),
+    linear-gradient(180deg, var(--hud-surface-top), var(--hud-surface-bottom)),
+    rgba(8, 20, 38, 0.12);
   border: 1px solid var(--hud-border);
   box-shadow:
     inset 0 1px 0 rgba(255, 255, 255, 0.14),
@@ -205,7 +236,11 @@
   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)
+    translate3d(
+      var(--elastic-x),
+      calc(var(--float-offset) + var(--press-offset) + var(--elastic-y)),
+      0
+    )
     scale(var(--btn-scale));
   transition:
     transform 0.22s ease,
@@ -213,17 +248,16 @@
     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: '';
+  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;
+    linear-gradient(180deg, rgba(255, 255, 255, 0.12), rgba(255, 255, 255, 0.04) 28%, transparent 68%);
+  opacity: 0.42;
   pointer-events: none;
   transform:
     perspective(120px)
@@ -234,14 +268,14 @@
 }
 
 .liquid-glass-surface::after {
-  content: '';
+  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;
+  opacity: 0.72;
   pointer-events: none;
   filter: url(#liquid-glass-distortion) blur(0.35px);
   transform:
@@ -260,15 +294,28 @@
   transition: opacity 0.18s ease, transform 0.18s ease;
 }
 
+.earth-toolbar-hub-btn.liquid-glass-surface {
+  background:
+    radial-gradient(circle at 50% 24%, rgba(255, 255, 255, 0.16), transparent 34%),
+    linear-gradient(180deg, rgba(28, 54, 90, 0.26), rgba(11, 24, 43, 0.22)),
+    rgba(10, 28, 52, 0.16);
+  border-color: 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),
+    0 16px 30px rgba(0, 0, 0, 0.24),
+    0 0 30px rgba(104, 181, 247, 0.18);
+}
+
 .liquid-glass-surface:hover {
-  --btn-scale: 1.035;
+  --btn-scale: 1.04;
   --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);
+    radial-gradient(circle at var(--glow-x) var(--glow-y), rgba(255, 255, 255, 0.14), transparent 34%),
+    radial-gradient(circle at 50% 118%, rgba(255, 255, 255, 0.08), transparent 30%),
+    linear-gradient(180deg, rgba(255, 255, 255, 0.12), rgba(128, 198, 255, 0.06)),
+    rgba(8, 20, 38, 0.14);
   border-color: var(--hud-border-hover);
   box-shadow:
     inset 0 1px 0 rgba(255, 255, 255, 0.2),
@@ -289,52 +336,16 @@
 
 .liquid-glass-surface:active,
 .liquid-glass-surface.is-pressed {
-  --btn-scale: 0.942;
+  --btn-scale: 0.95;
   --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);
+    radial-gradient(circle at var(--glow-x) var(--glow-y), rgba(255, 255, 255, 0.14), transparent 34%),
+    linear-gradient(180deg, rgba(255, 255, 255, 0.14), rgba(118, 200, 255, 0.08)),
+    rgba(11, 34, 58, 0.18);
   border-color: var(--hud-border-active);
   box-shadow:
     inset 0 1px 0 rgba(255, 255, 255, 0.22),
@@ -357,151 +368,64 @@
 
 .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 {
+.earth-zoom-group.open > .earth-zoom-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 {
+.earth-zoom-group.force-closed > .earth-zoom-toolbar {
   opacity: 0;
   visibility: hidden;
   pointer-events: none;
-  transform: translate(-50%, 8px);
+  transform: translate(-50%, calc(8px * var(--toolbar-scale)));
 }
 
-.earth-zoom-group > .earth-zoom-toolbar,
-.earth-info-group > .earth-info-toolbar {
+.earth-toolbar-popover::before {
+  content: "";
+  position: absolute;
+  left: 50%;
+  bottom: 100%;
+  transform: translateX(-50%);
+  width: calc(56px * var(--toolbar-scale));
+  height: calc(16px * var(--toolbar-scale));
+  background: transparent;
+}
+
+.earth-toolbar-popover > .earth-stack-toolbar {
+  position: absolute;
+  left: 50%;
   top: auto;
   right: auto;
-  left: 50%;
-  bottom: calc(100% + 12px);
+  bottom: calc(100% + (12px * var(--toolbar-scale)));
+  transform: translate(-50%, calc(10px * var(--toolbar-scale)));
   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;
+  gap: calc(8px * var(--toolbar-scale));
+  opacity: 0;
+  visibility: hidden;
+  pointer-events: none;
+  transition:
+    opacity 0.22s ease,
+    transform 0.22s ease,
+    visibility 0.22s ease;
+  z-index: 220;
 }
 
 .earth-zoom-toolbar .earth-zoom-btn,
 .earth-zoom-toolbar .earth-zoom-value {
-  width: 42px;
-  min-width: 42px;
+  width: calc(42px * var(--toolbar-scale));
+  min-width: calc(42px * var(--toolbar-scale));
   border-radius: 50%;
-  color: #4db8ff;
-  animation: floatDock 3.8s ease-in-out infinite;
+  color: var(--hud-text-soft);
+  animation: none;
 }
 
 .earth-zoom-toolbar .earth-zoom-btn {
-  height: 42px;
-  font-size: 20px;
+  height: calc(42px * var(--toolbar-scale));
+  font-size: calc(20px * var(--toolbar-scale));
   font-weight: 500;
   line-height: 1;
 }
@@ -510,41 +434,12 @@
   display: inline-flex;
   align-items: center;
   justify-content: center;
-  height: 42px;
+  height: calc(42px * var(--toolbar-scale));
   padding: 0;
-  font-size: 0.68rem;
+  font-size: calc(11px * var(--toolbar-scale));
   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%);
@@ -552,21 +447,23 @@
 
 .earth-toolbar-btn .earth-toolbar-tooltip {
   position: absolute;
-  bottom: 56px;
+  bottom: calc(56px * var(--toolbar-scale));
   left: 50%;
   transform: translateX(-50%);
-  background: rgba(10, 10, 30, 0.95);
-  color: #fff;
-  padding: 6px 12px;
-  border-radius: 6px;
-  font-size: 12px;
+  background:
+    linear-gradient(180deg, rgba(18, 31, 52, 0.96), rgba(8, 18, 32, 0.95));
+  color: var(--hud-text);
+  padding: calc(6px * var(--toolbar-scale)) calc(12px * var(--toolbar-scale));
+  border-radius: calc(6px * var(--toolbar-scale));
+  font-size: calc(12px * var(--toolbar-scale));
   white-space: nowrap;
   opacity: 0;
   visibility: hidden;
   transition: all 0.2s ease;
-  border: 1px solid rgba(77, 184, 255, 0.4);
+  border: 1px solid var(--hud-border);
   pointer-events: none;
   z-index: 100;
+  box-shadow: var(--hud-shadow-soft);
 }
 
 .earth-toolbar-btn:hover .earth-toolbar-tooltip,
@@ -574,15 +471,15 @@
 .earth-toolbar-popover:focus-within > .earth-toolbar-btn .earth-toolbar-tooltip {
   opacity: 1;
   visibility: visible;
-  bottom: 58px;
+  bottom: calc(58px * var(--toolbar-scale));
 }
 
 .earth-toolbar-btn .earth-toolbar-tooltip::after {
-  content: '';
+  content: "";
   position: absolute;
   top: 100%;
   left: 50%;
   transform: translateX(-50%);
-  border: 6px solid transparent;
-  border-top-color: rgba(77, 184, 255, 0.4);
+  border: calc(6px * var(--toolbar-scale)) solid transparent;
+  border-top-color: rgba(18, 31, 52, 0.96);
 }

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

@@ -1,22 +1,98 @@
-/* tv-panel */
+/* media-panel
+ * Outer HUD shell: #media-panel
+ * Inner live pane: #tv-panel
+ * Inner news pane: #news-panel
+ */
 
-.hud-panel-tv {
+.hud-panel-media {
   bottom: var(--hud-offset);
   right: var(--hud-offset);
   width: calc(420px * var(--hud-scale));
   max-width: calc(100vw - 32px);
+  max-height: calc(100vh - (2 * var(--hud-offset)));
   min-width: calc(300px * var(--hud-scale));
-  min-height: calc(340px * var(--hud-scale));
-  padding: calc(18px * var(--hud-scale));
+  padding: calc(10px * 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));
+.hud-panel-media[data-active-tab="news"]:not([data-resized="true"]) {
+  max-height: min(
+    var(--tv-news-default-max-height, calc(100vh - (2 * var(--hud-offset)))),
+    calc(100vh - (2 * var(--hud-offset)))
+  );
+}
+
+.hud-panel-media.is-reforming {
+  transition:
+    height 0.24s cubic-bezier(0.22, 1, 0.36, 1),
+    top 0.24s cubic-bezier(0.22, 1, 0.36, 1);
+  will-change: height, top;
+}
+
+.hud-panel-media .hud-panel__header {
+  align-items: center;
+  gap: var(--hud-gap-xs);
+}
+
+.hud-panel-media .hud-panel__title-group {
+  flex: 0 0 auto;
+  min-width: auto;
+}
+
+.tv-panel-header-title {
+  flex: 0 0 auto;
+  white-space: nowrap;
+  margin: 0;
+}
+
+.hud-panel-media .hud-panel__header .hud-panel__action,
+.hud-panel-media .hud-panel__header .hud-panel-close {
+  cursor: pointer;
+  user-select: auto;
+}
+
+.hud-panel-media .hud-panel__header .tv-panel-select,
+.hud-panel-media .hud-panel__header .media-panel-tab {
+  cursor: pointer;
+  user-select: auto;
+}
+
+.tv-panel-header-controls {
+  display: flex;
+  align-items: center;
+  gap: var(--hud-gap-xs);
+  flex: 1 1 auto;
+  min-width: 0;
+}
+
+.tv-panel-header-controls--news {
+  justify-content: flex-end;
+}
+
+.tv-panel-content {
+  display: flex;
+  flex: 1 1 auto;
+  min-height: 0;
+  overflow: hidden;
+}
+
+.tv-tab-pane {
+  display: flex;
+  flex: 1 1 auto;
+  min-height: 0;
+  flex-direction: column;
+  gap: var(--hud-gap-sm);
+  overflow: hidden;
+}
+
+.tv-panel-toolbar-actions {
+  display: inline-flex;
+  align-items: center;
+  gap: var(--hud-gap-xs);
+  flex-shrink: 0;
 }
 
 .tv-panel-status {
@@ -26,13 +102,6 @@
   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;
@@ -53,48 +122,19 @@
   color: #eef5fc;
 }
 
-.tv-panel-actions {
-  display: flex;
-  gap: var(--hud-gap-xs);
+.tv-panel-meta-wrap {
+  overflow: hidden;
+  max-height: calc(120px * var(--hud-scale));
+  opacity: 1;
+  transition: max-height 0.22s ease, opacity 0.18s ease, margin 0.22s ease;
 }
 
-.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;
+.tv-panel-meta-wrap.is-collapsed {
+  max-height: 0;
+  opacity: 0;
   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;
+  margin-top: 0;
+  margin-bottom: 0;
 }
 
 .tv-panel-meta {
@@ -163,45 +203,152 @@
   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;
+.media-panel-tabs {
+  display: grid;
+  grid-template-columns: 1fr 1fr;
+  gap: calc(8px * var(--hud-scale));
 }
 
-.tv-panel-resize-handle::before {
+.media-panel-tab {
+  border: 1px solid rgba(201, 225, 247, 0.12);
+  border-radius: calc(12px * var(--hud-scale));
+  background: rgba(255, 255, 255, 0.03);
+  color: var(--hud-text-muted);
+  padding: calc(9px * var(--hud-scale)) calc(12px * var(--hud-scale));
+  font-size: calc(0.78rem * var(--hud-scale));
+  font-weight: 600;
+  letter-spacing: 0.04em;
+  cursor: pointer;
+  transition: background 0.18s ease, border-color 0.18s ease, color 0.18s ease;
+}
+
+.media-panel-tab:hover {
+  color: var(--hud-text);
+  border-color: rgba(214, 235, 255, 0.18);
+  background: rgba(255, 255, 255, 0.06);
+}
+
+.media-panel-tab--active {
+  color: var(--hud-accent-strong);
+  border-color: rgba(120, 180, 255, 0.24);
+  background: rgba(120, 180, 255, 0.12);
+}
+
+.tv-tab-pane[hidden] {
+  display: none !important;
+}
+
+/* ── Multi-edge resize handles ───────────────────────────────── */
+
+.tv-panel-edge {
+  position: absolute;
+  z-index: 10;
+}
+
+.tv-panel-edge[data-edge="r"] {
+  right: 0;
+  top: calc(12px * var(--hud-scale));
+  bottom: calc(12px * var(--hud-scale));
+  width: calc(6px * var(--hud-scale));
+  cursor: ew-resize;
+}
+
+.tv-panel-edge[data-edge="b"] {
+  bottom: 0;
+  left: calc(12px * var(--hud-scale));
+  right: calc(12px * var(--hud-scale));
+  height: calc(6px * var(--hud-scale));
+  cursor: ns-resize;
+}
+
+.tv-panel-edge[data-edge="l"] {
+  left: 0;
+  top: calc(12px * var(--hud-scale));
+  bottom: calc(12px * var(--hud-scale));
+  width: calc(6px * var(--hud-scale));
+  cursor: ew-resize;
+}
+
+.tv-panel-edge[data-edge="br"] {
+  right: 0;
+  bottom: 0;
+  width: calc(20px * var(--hud-scale));
+  height: calc(20px * var(--hud-scale));
+  cursor: nwse-resize;
+}
+
+.tv-panel-edge[data-edge="bl"] {
+  left: 0;
+  bottom: 0;
+  width: calc(20px * var(--hud-scale));
+  height: calc(20px * var(--hud-scale));
+  cursor: nesw-resize;
+}
+
+.layout-mode-mobile .hud-panel-media {
+  position: fixed;
+  left: 8px;
+  right: 8px;
+  top: calc(8px + var(--safe-top));
+  bottom: calc(84px + var(--safe-bottom));
+  width: auto;
+  max-width: none;
+  max-height: none;
+  min-width: 0;
+  z-index: 230;
+}
+
+.layout-mode-mobile .tv-panel-player {
+  min-height: min(42vh, 360px);
+}
+
+.layout-mode-mobile .tv-panel-edge {
+  display: none;
+}
+
+/* 右下角视觉标记 */
+.tv-panel-edge[data-edge="br"]::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;
+  inset: calc(4px * var(--hud-scale));
+  border-right: 2px solid rgba(223, 235, 248, 0.4);
+  border-bottom: 2px solid rgba(223, 235, 248, 0.4);
+  transition: border-color 0.18s ease;
 }
 
-.tv-panel-resize-handle:hover::before {
-  opacity: 1;
-  border-color: rgba(244, 249, 255, 0.78);
+.tv-panel-edge[data-edge="br"]:hover::before {
+  border-color: rgba(244, 249, 255, 0.75);
 }
 
-.hud-panel-tv.is-resizing {
+.hud-panel-media.is-resizing {
   transition: none !important;
   user-select: none;
 }
 
-.earth-app.layout-expanded .hud-panel-tv:not([data-dragged="true"]) {
+/* Whole panel is draggable; player overrides back to default */
+.hud-panel-media:not(.is-resizing) {
+  cursor: grab;
+}
+
+.hud-panel-media.is-dragging {
+  cursor: grabbing;
+}
+
+.hud-panel-media .tv-panel-player {
+  cursor: default;
+}
+
+/* Disable iframe/video pointer capture while dragging so mouse events pass through */
+.hud-panel-media.is-dragging .tv-panel-iframe,
+.hud-panel-media.is-dragging .tv-panel-video {
+  pointer-events: none;
+}
+
+.earth-app.layout-expanded .hud-panel-media: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.
+/* Media panel keeps its fixed width on all screen sizes.
    Responsive stretching removed — width only changes if user manually resizes. */