AlphaRadar

AlphaRadar is a TUI-first market intelligence terminal for US and HK equities.

It does not place trades. It collects market-relevant information, normalizes it into structured signals, maps signals to sectors and stocks, and lets you inspect the result from the terminal.

What It Does

• Aggregates RSS news, macro feeds, and local knowledge rules into a single snapshot.
• Normalizes multilingual and noisy source text into structured market events.
• Builds two kinds of outputs:
  • signal -> stock
  • signal -> sector -> stocks
• Tracks theme-specific snapshots for ai, semiconductors, cloud, biotech, and china-internet.
• Supports source health, retry workflows, issue reason summaries, and cooldown for repeatedly failing sources.
• Exposes everything through CLI and a curses-based TUI.

Current Scope

Implemented:

• US and HK market snapshots
• RSS feed catalog import from OPML
• Official macro source catalog
• Multi-platform domestic hotlist fetching via NewsNow API (e.g., Toutiao, Weibo, Bilibili)
• Web search integrations for agents via Tavily, Bocha, and Anspire APIs
• High priority tech & finance RSS profile matching (HackerNews, Yahoo Finance)
• Theme-specific snapshot stores
• Source health, issue classification, retry, and cooldown
• TUI for signals, themes, and theme stores

Not implemented yet:

• Real-time Reddit / X ingestion
• Order execution
• Backtesting
• Web UI

Install

Requirements:

• Python >= 3.11

Local editable install:

pip install -e .

Or run directly from source:

PYTHONPATH=src python -m alpharadar.cli.main --help

Project Layout

• src/alpharadar/cli/main.py: CLI entrypoint
• src/alpharadar/tui/app.py: terminal UI
• src/alpharadar/storage/sqlite_store.py: SQLite storage, source catalog, health, cooldown
• src/alpharadar/app/store.py: snapshot and theme-store helpers
• src/alpharadar/knowledge: aliases, event rules, source profiles, macro sources, theme watchlists

快速使用指南 (Quick Start)

AlphaRadar 的核心工作流可以分为：数据初始化 -> 同步刷新 -> 查看信号/大盘情绪。

1. 启动体验 (使用内置数据)

# 生成基准快照
alpharadar ingest
# 查看提取的市场信号 (文本形式)
alpharadar signals
# 查看宏观新闻
alpharadar macro
# 查看聚焦股票 (包含交叉市场逻辑)
alpharadar watch --market US

2. 真实使用流程 (导入高质量信息源)

A. 导入 RSS 数据源 解析 OPML 格式的 RSS 订阅列表。比如你手边有 US 的书签：

alpharadar import-feeds /path/to/United\ States.opml --market US --region "United States"
# 支持带类型前缀的精细导入
alpharadar import-feeds /path/to/feeds.opml --market US --region "Global Tech" --category-prefix Tech

B. 启用内置的权威宏观数据源

alpharadar seed-macro-sources --market US
alpharadar seed-macro-sources --market HK
alpharadar list-macro-sources --market US

C. 数据源自动打分与过滤 基于内置的优质名单（覆盖 HackerNews, Yahoo Finance 等），系统帮你清理垃圾信息：

alpharadar score-feeds --market US
alpharadar autocurate-feeds --market US --threshold 0.5 --disable-others
alpharadar list-feeds --market US --curated-only

3. 生成真正的市场快照

当你订阅好信息源后，随时拉取最新信息，它会自动做清洗、提取信号：

# 拉取被启用的优质数据，并构建你专属的 US 市场全景快照
alpharadar refresh-market --market US --curated-only
alpharadar signals
alpharadar --lang zh signals
alpharadar signals --cross-market-only

(系统会自动记录上一次拉取的 Primary Market，诸如 signals, watch, macro 等命令会自动跟随此偏好；所有 CLI 命令支持 --lang en|zh。)

🎛 终端 UI (TUI) 沉浸式分析

每次刷新好快照后，按以下命令进入终端可视化大屏：

alpharadar tui
alpharadar --lang zh tui

TUI 核心快捷键：

• 0-9: 在全局市场视角与各垂直主题（如 ai、芯片等）之间快速切换。
• a / u / h: 快速筛选查看 全部(All) / 美股(US) / 港股(HK)。
• x: 开启跨市场视角（比如在美股版块中，探求同步关联的港股异动信号）。
• l: 在英文 / 中文显示之间切换。
• s: 唤起侧滑的 Theme-store 浏览器。
• n / o / t / e: 对侧边栏数据库进行状态筛选（全部 / 健康 / 过期 / 错误）。
• Enter: 潜入选中的数据库。
• g: 一键返回全局视角。
• r: 重新从磁盘加载数据库。
• j / k: 上下移动。
• q: 退出。

高级分析：聚焦垂直主题 (Themes Workflow)

Theme stores 是独立的 SQLite 库，专门用于存放特定赛道（如 ai, semiconductors, cloud, biotech 等）的切片快照。

# 查看支持的所有主题
alpharadar themes

# 自动筛选并汇聚有关 AI 的专属数据库
alpharadar refresh-theme ai --market US --include-macro

# 终端纯净态查看主题切片
alpharadar theme-watch ai --market US --use-theme-store
alpharadar theme-status ai --market US

# 直接进入这块主题的专属 TUI 终端大屏
alpharadar theme-tui ai --market US

Source Operations

Feed catalog

List feeds:

alpharadar list-feeds --market US
alpharadar list-feeds --market US --curated-only
alpharadar list-feeds --market US --enabled-only --include-cooling

Manually change feed state:

alpharadar set-feed techcrunch --enable --priority 2.0
alpharadar set-feed cnn --disable

Fetch RSS without rebuilding the snapshot:

alpharadar fetch-feeds --market US --curated-only --market-signal-only

Macro source catalog

List macro sources:

alpharadar list-macro-sources --market US
alpharadar list-macro-sources --market US --enabled-only --include-cooling

Fetch macro feeds without rebuilding the snapshot:

alpharadar fetch-macro --market US

Health, Retry, and Cooldown

AlphaRadar tracks:

• last_fetch_status
• last_fetch_error
• consecutive_failures
• cooldown_until
• cooldown_reason

After repeated failures, a source enters cooldown and is skipped by default during refresh/fetch. Use --include-cooling to inspect or explicitly include those sources.

Health views:

alpharadar feed-health --market US
alpharadar macro-health --market US
alpharadar list-theme-stores --status stale
alpharadar theme-status ai --market US

Retry failed theme sources:

alpharadar retry-theme-sources ai --market US --dry-run
alpharadar retry-theme-sources ai --market US --dry-run --rss-only --limit-targets 2
alpharadar retry-theme-sources ai --market US --dry-run --rss-only --reason-contains 403
alpharadar retry-theme-sources ai --market US --macro-only --reason-contains timeout
alpharadar retry-theme-sources ai --market US

retry-theme-sources supports:

• --rss-only
• --macro-only
• --limit-targets
• --reason-contains
• --retry-status error|empty

TUI

Open the global TUI:

alpharadar tui
alpharadar --lang zh tui

Open a theme-specific TUI:

alpharadar theme-tui ai --market US

Main controls:

• 0-9: switch between global view and theme views
• a / u / h: filter by ALL / US / HK
• x: show cross-market signals relative to the active global snapshot's primary market
• l: switch display language between English and Chinese
• s: open the theme-store browser
• n / o / t / e: filter theme stores by all / ok / stale / error
• enter: open the selected theme store
• g: return to the global snapshot
• r: reload the active store from disk
• j / k: move selection
• q: quit

The theme-store view shows:

• snapshot status
• signal and event counts
• RSS and macro message counts
• issue counts and issue reasons
• per-source issue summaries

CLI Reference

Core analysis:

• ingest
• signals
  • --include-cross-market
  • --cross-market-only
• explain
• macro
• watch
  • --include-cross-market
  • --cross-market-only
• themes
• theme-watch
• tui
• theme-tui

Global options:

• --store
• --lang en|zh

Catalog management:

• import-feeds
• list-feeds
• score-feeds
• set-feed
• autocurate-feeds
• seed-macro-sources
• list-macro-sources

Fetching and snapshot rebuild:

• fetch-feeds
• fetch-macro
• build-from-feeds
• build-from-all
• refresh-rss
• refresh-market
• refresh-theme

Health and maintenance:

• feed-health
• macro-health
• list-theme-stores
• theme-status
• retry-theme-sources

Development

Run tests:

PYTHONPATH=src python -m unittest discover -s tests -v

Compile check:

python -m py_compile $(find src -name '*.py' -print)

Notes

• The project is still MVP-stage and rule-heavy by design.
• Theme stores are isolated from the global snapshot on purpose.
• Source issue summaries and cooldown are meant to reduce repeated fetch noise, not permanently disable sources.
• Most operational state lives in local SQLite files under .alpharadar/.