A scheduled Telegram bot that pushes Vietnamese market data and developer trends to a channel — gold prices, Bitcoin, stocks, and GitHub trending repos. Each data source runs on its own cron schedule via GitHub Actions, so no server is required.

• 📈 Vietnamese stock prices — VN-Index, VN30, and configurable tickers (via vnstock)
• 🏆 Gold prices — SJC and Kim Long Đồng Tháp jewelry
• ₿ Bitcoin — BTC/USD with 24h change (via CoinGecko)
• 🔥 GitHub Trending — top repos with clickable links
• ⚙️ Plugin architecture — add new sources by implementing a simple 2-method interface
• 🕐 Zero-ops scheduling — runs entirely on GitHub Actions cron
• 🎨 Beautiful messages — HTML formatting with emojis, no link-preview clutter

📈 Giá cổ phiếu Việt Nam

🏛 Chỉ số thị trường
  • VN-Index: 1,819.83 📈 +1.07%
  • VN30: 1,979.19 📈 +0.90%

💹 Cổ phiếu
  • VNM: 61,100 VNĐ 📉 -0.33%
  • HPG: 26,500 VNĐ 📈 +1.85%

🔗 Nguồn: vnstocks

1. Fork this repo (or use it as a template).

2. Create a Telegram bot via @BotFather, then add it as admin to your channel.

3. Add repository secrets — Settings → Secrets and variables → Actions:

• TELEGRAM_BOT_TOKEN — your bot token
• TELEGRAM_CHANNEL_ID — @mychannel or numeric chat ID (e.g. -1001234567890)

4. Customize sources.example.yaml (copy to sources.yaml for local use; GitHub Actions falls back to the example).

5. Trigger manually — go to Actions tab → pick a workflow → Run workflow. Or wait for the cron.

That's it! No server, no Docker, no money spent.

# Clone and install
git clone https://github.com/alanzng/news-bot-say-hi.git
cd news-bot-say-hi
pip install -r requirements.txt

# Configure
cp sources.example.yaml sources.yaml
cp .env.example .env
# edit .env with your bot token + channel ID

# Run all sources on schedule (long-running process)
make run

# Run a single source once (useful for debugging)
python -m src.cli --source gold-price

# Run tests
make test

# Lint
make lint

sources.yaml (or sources.example.yaml as a fallback) controls enabled sources and their options:

sources:
  stock-price:
    enabled: true
    schedule: "0 9 * * 1-5"   # cron (local mode only; GitHub Actions uses .github/workflows/)
    tickers:
      - HPG   # Hoa Phat Group
      - MBB   # MB Bank
      - SSI   # SSI Securities

  github-trending:
    enabled: true
    language: ""              # empty = all languages; or "python", "rust", etc.
    since: "daily"            # daily | weekly | monthly
    limit: 5

Set enabled: false to disable any source.

src/
├── base.py              # DataSource abstract base class
├── cli.py               # Single-source runner (used by GitHub Actions)
├── main.py              # APScheduler long-running process (local mode)
├── config.py            # Loads sources.yaml + env secrets
├── notifier.py          # Telegram Bot API wrapper
├── registry.py          # Wires sources to schedules
└── sources/
    ├── gold_price.py
    ├── kimlong_gold_price.py
    ├── bitcoin_price.py
    ├── stock_price.py
    └── github_trending.py
.github/workflows/       # One cron workflow per source + CI
tests/                   # pytest suite mirroring src/

1. Create src/sources/my_source.py extending DataSource:

   from src.base import DataSource
   
   class MySource(DataSource):
       name = "my-source"
       default_schedule = "0 9 * * *"
   
       def fetch(self) -> list[dict]:
           # return records or [] on soft failure
           ...
   
       def format(self, records: list[dict]) -> str:
           # return HTML Telegram message, or "" to skip sending
           ...

2. Register in src/registry.py.

3. Add a block in sources.example.yaml.

4. Copy an existing .github/workflows/*.yml and update the --source arg + cron.

5. Add tests at tests/sources/test_my_source.py.

See AGENTS.md for the full contributor playbook.

Contributions welcome! See CONTRIBUTING.md for the workflow, coding conventions, and PR checklist. For AI-assisted contributions (Claude Code, Cursor, etc.), see AGENTS.md.

MIT © Alan Ng

• vnstock — Vietnamese stock market data
• SJC, Kim Long Đồng Tháp — gold price data
• CoinGecko — crypto prices
• CafeF — market index data