FeedForBot 📡

RSS/Atom in, Telegram out. 🪄 Fresh entries arrive, templated into tidy messages, deduped via your cache of choice — on a cron schedule, with a healthcheck and graceful shutdown baked in.

✨ Features

• 📬 Multiple feeds — any number of listeners/transports, each on its own cron
• 🎨 Jinja2 templates — sandbox + HTML autoescape, full access to article fields
• 💾 Pluggable cache — files, in-memory, or Redis; first run silent to avoid flood
• 🩺 Healthcheck — HTTP /health for Docker/Kubernetes probes
• 🛡️ Resilient delivery — auto-retry on Telegram flood-wait, skip malformed entries
• 🧹 Graceful shutdown — SIGTERM/SIGINT drain ticks and close handles cleanly
• 🐛 Sentry integration — optional error tracking
• 🐳 Docker ready — multi-arch GHCR + Docker Hub images, runs as non-root
• 🧩 Protocol-driven — custom listeners/transports/caches via simple protocols

📦 Installation

Requires Python 3.10+.

pip install feedforbot -U

For the full CLI (Click, structlog, YAML config, Sentry, Redis):

pip install "feedforbot[cli]" -U

🚀 Quick start

🐍 As a library

from feedforbot import Scheduler, TelegramBotTransport, RSSListener

scheduler = Scheduler(
    '*/5 * * * *',
    listener=RSSListener('https://www.debian.org/News/news'),
    transport=TelegramBotTransport(
        token='123456789:AAAAAAAAAA-BBBB-CCCCCCCCCCCC-DDDDDD',
        to='@channel',
    ),
)
scheduler.run()  # blocks, checks the feed every 5 minutes

⚡ Async — multiple schedulers

import asyncio

from feedforbot import Scheduler, TelegramBotTransport, RSSListener

schedulers = [
    Scheduler(
        '*/5 * * * *',
        listener=RSSListener('https://www.debian.org/News/news'),
        transport=TelegramBotTransport(
            token='123456789:AAAAAAAAAA-BBBB-CCCCCCCCCCCC-DDDDDD',
            to='@channel',
        ),
    ),
    Scheduler(
        '0 * * * *',
        listener=RSSListener('https://habr.com/ru/rss/all/all/'),
        transport=TelegramBotTransport(
            token='123456789:AAAAAAAAAA-BBBB-CCCCCCCCCCCC-DDDDDD',
            to='@another_channel',
        ),
    ),
]

async def main() -> None:
    await asyncio.gather(*(s.arun() for s in schedulers))

asyncio.run(main())

📝 CLI with YAML config

Create a schedulers.yml — a top-level list of listener+transport entries:

---
- rule: '*/5 * * * *'
  listener:
    type: 'rss'
    params:
      url: 'https://habr.com/ru/rss/all/all/?fl=ru'
  transport:
    type: 'telegram_bot'
    params:
      token: '123456789:AAAAAAAAAA-BBBB-CCCCCCCCCCCC-DDDDDD'
      to: '@tmfeed'
      template: |-
        <b>{{ TITLE }}</b> #habr
        {{ ID }}
        <b>Tags</b>: {% for category in CATEGORIES %}{{ category }}{{ ", " if not loop.last else "" }}{% endfor %}
        <b>Author</b>: <a href="https://habr.com/users/{{ AUTHORS[0] }}">{{ AUTHORS[0] }}</a>
- listener:
    type: 'rss'
    params:
      url: 'http://www.opennet.ru/opennews/opennews_all.rss'
  transport:
    type: 'telegram_bot'
    params:
      token: '123456789:AAAAAAAAAA-BBBB-CCCCCCCCCCCC-DDDDDD'
      to: '@tmfeed'
      disable_web_page_preview: yes
      template: |-
        <b>{{ TITLE }}</b> #opennet
        {{ URL }}

        {{ TEXT }}

Run:

feedforbot --verbose schedulers.yml

On the first run the cache is populated without sending messages, so existing feed entries won't flood the channel. 🤫

📋 Config reference

Each list entry supports:

An empty top-level list is rejected at parse time.

💾 Cache CLI flag

Cache backend selection is a CLI concern (not part of the schedulers file). --cache-dsn takes a URL-style DSN; scheme selects the backend:

🧩 Template variables

All fields from ArticleModel are available in uppercase:

Templates render in a Jinja2 sandbox with HTML autoescape enabled (messages use parse_mode=HTML). Article fields containing &, <, > are escaped automatically — you don't need to guard against Telegram rejecting the message for invalid HTML.

🎛️ CLI options

🐳 Docker

Images are published to both GHCR and Docker Hub on every release. Tags follow semver: latest, 5, 5.0, 5.0.0.

docker run -v $(pwd)/schedulers.yml:/schedulers.yml \
  ghcr.io/shpaker/feedforbot --verbose /schedulers.yml

docker run -v $(pwd)/schedulers.yml:/schedulers.yml \
  shpaker/feedforbot --verbose /schedulers.yml

The container runs as a non-root user (appuser).

🩺 Healthcheck

The CLI includes a built-in HTTP healthcheck server. Pass --healthcheck-port to expose a lightweight endpoint that responds with 200 OK at /health (other paths return 404 Not Found):

feedforbot --healthcheck-port 8080 --verbose schedulers.yml

The server binds to 127.0.0.1 by default. Pass --healthcheck-host 0.0.0.0 when the port must be reachable from outside the container (e.g. for an external probe); the in-container Docker HEALTHCHECK below works either way since it queries localhost from the same network namespace.

Useful for container orchestrators (Docker HEALTHCHECK, Kubernetes liveness probes, etc.):

# docker-compose.yml
services:
  feedforbot:
    image: ghcr.io/shpaker/feedforbot
    command: ["--healthcheck-port", "8080", "--verbose", "/schedulers.yml"]
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8080/health"]
      interval: 30s
      timeout: 5s
      retries: 3

🧱 Docker Compose with Redis

For multi-instance deployments or when you want cache state to survive container rebuilds without a host volume for JSON files, run feedforbot alongside a Redis container and point --cache-dsn at it:

# docker-compose.yml
services:
  redis:
    image: redis:7-alpine
    restart: always
    volumes:
      - ./redis_data:/data
    command: ["redis-server", "--save", "60", "1", "--appendonly", "no"]
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 10s
      timeout: 3s
      retries: 3
      start_period: 5s

  feedforbot:
    image: ghcr.io/shpaker/feedforbot
    restart: always
    depends_on:
      redis:
        condition: service_healthy
    volumes:
      - ./schedulers.yml:/schedulers.yml:ro
    command: >
      -v
      --cache-dsn redis://redis:6379/0
      --healthcheck-port 8080
      /schedulers.yml
    healthcheck:
      test: ["CMD", "python", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:8080/health')"]
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 10s

A working Ansible playbook that deploys this layout is in tmfeed/deploy.playbook.yml.

📄 License

MIT