barbot

Barbot is a Telegram bot that automates decision making for our weekly bar night.

Running the Bot Locally

0. Ensure the aws and sam CLI tools are install.

1. Create a file at the root of the repo named env.json.

This is the environment file used by AWS SAM for setting the function's environment variables. The file will look something like this:

{
  "WebhookFunction": {
    "MAIN_CHAT_ID": "<XXXXX>",
    "TELEGRAM_BOT_TOKEN": "<XXXXX>:<XXXXX>",
    "BAR_SPREADSHEET": "https://docs.google.com/spreadsheets/d/<XXXXX>/",
    "SELENIUM_SERVER_URL": "http://<XXXXX>:4444"
  }
}

Make sure the telegram bot and main chat id in env.json are different than the live bot's. Local development will nuke the webhook information for that bot.

2. Set the NGROK_AUTHTOKEN enviornment variable.

You need a free ngrok token for development - this is so telegram can talk to your local webhook handler.

3. Run ./run-local.py

This takes care of setting up the helper services, building the application, running it, and pointing telegram to the local instance.

Invoking Scheduled Events Locally

1. Duplicate the "WebhookFunction" object in env.json under the key SequenceFunction.

This is an unfortunate limitation of SAM:

{
  "WebhookFunction": {"a": "b", "c": "d"},
  "SequenceFunction": {"a": "b", "c": "d"},
}

2. Run sam local invoke

WARNING: Make sure you've run ./run-local.py before invoking sam, because ./run-local.py is what builds the image that sam uses (otherwise, you'll be using an outdated barbot).

An invocation should look something like this:

echo '{"barnight_event_type": "CreatePoll"}' | sam local invoke SequenceFunction --docker-network barbot --env-vars env.json -e -

Building & Deploying

1. Install the version of python currently being targeted by the lambda runtime (See lambda.tf. At the time of writing, this is Python 3.12)

2. Create a virtual environment using that python version. python3.12 -m venv venv

3. Activate the virtual environment and run ./build.py

4. In the terraform directory, Copy a template .tfvars file to terraform.tfvars. Set the empty values.

5. Use terraform workspace show and terraform workspace select to choose the correct terraform state to use. default workspace is used for prod, and test is used for the 'testing' bot and channel.