SCinet Telegram Bot

Small utility bot that keeps the SCinet student volunteers coordinated inside Telegram. It publishes uniform reminders, exposes the help-desk schedule, and provides booth lookup tools via inline queries.

What It Does

• /hello – quick sanity check to confirm the bot is alive.
• Inline query (type @<YourBotName> [query] in any chat):
  • With no query text, Telegram displays a stack of default cards that include “What to wear today/tomorrow”, “Current/next shift”, and “Today’s schedule”.
  • When text is provided, the bot switches to exhibitor booth search (fuzzy booth name matching and reverse lookup by booth number).

Project Layout

• tg.py – entry point; registers Telegram handlers and runs the polling loop.
• schedule.py & schedule_config.json – returns the last/current/next shifts and the full schedule for a given day.
• wear.py & wear_config.json – returns the wardrobe plan for today/tomorrow.
• booth.py & booth.json – fuzzy booth name search and reverse lookup by booth number.
• config.json – stores {"botToken": "123:ABC"} for authentication.

Requirements

• Python 3.11+ (needed for zoneinfo, EC2 timezone may different from local).
• Packages: python-telegram-bot>=21, fuzzywuzzy[speedup], requests.

  pip install "python-telegram-bot>=21" "fuzzywuzzy[speedup]" requests

  If you are more famliar with word embedding and vector data/ vector database, I highly recommend you use it. For example, Bert with Faiss instead of using fuzzywuzzy

How to Run

1. Add your bot token to config.json.
2. Populate the JSON data files (schedule_config.json, wear_config.json, booth.json) with the latest information.
3. Start the bot locally (or on any host where Python is available):

   python tg.py

4. In Telegram:
   • Type @<YourBotName> to see the default inline cards (wear info + shift info).
   • Type @<YourBotName> Dell (or any exhibitor/booth number) to search the booth list.

Hosting Options

• Local workstation: great for testing changes quickly; just run python tg.py in a terminal and keep it alive during development.
• AWS EC2 instance: deploy the repo to a small instance, install the dependencies, and keep the bot online 24/7.
  • I use a simple systemd service file (e.g., /etc/systemd/system/scinet-bot.service) that runs python /opt/scinet/tg.py, restarts the bot on failure, and hooks into EC2’s boot process so the bot starts automatically.
  • After placing the service file, run sudo systemctl daemon-reload, then sudo systemctl enable --now scinet-bot to keep it managed by systemd.
  • Update the JSON config files on the instance and sudo systemctl restart scinet-bot whenever data changes.

CI/CD Workflow

• GitHub Actions workflow (.github/workflows/main.yml) triggers on every push to main.
• Steps:
  1. Check out the repository.
  2. Use rsync (via burnett01/rsync-deployments) to mirror the repo to /home/ec2-user/SCinet/SCinetBot on the EC2 host specified by EC2_HOST/EC2_SSH_KEY secrets.
  3. SSH in (via appleboy/ssh-action) to stop the running tg.service, reload systemd, enable tg_bot.service, and start it again—ensuring the latest code is live.
• Keep the EC2_HOST and EC2_SSH_KEY secrets up to date in the repo settings so deployments keep working.

Updating Data

• Edit the JSON files whenever the schedule, wardrobe plan, or exhibitor list changes.
• Restart tg.py to reload the files.
• For timezone changes, set the timezone argument in schedule.Schedule.

Troubleshooting

• If the bot does not start, confirm the token and that inline mode is enabled in BotFather (/setinline).
• Inline queries require the bot to be added to the chat and may take a second to return results due to fuzzy searching.