# CommentAutomation

Python project initialized with uv.

## Quick start

1. Install uv.
2. Sync dependencies:

   uv sync --group dev

3. Run the app:

   uv run uvicorn comment_automation.main:app --app-dir src --host 0.0.0.0 --port 8000

4. Run tests:

   uv run pytest

## Webhooks

Instagram comments webhook endpoint:

- POST /webhooks/instagram/comments

The route validates incoming payloads and enqueues a Celery task.

## LLM task configuration

The Celery task that handles webhook payload forwarding reads these environment variables:

- LLM_ENDPOINT_URL (required): Full HTTP(S) URL for the downstream LLM endpoint.
- LLM_ENDPOINT_API_KEY (optional): Bearer token sent as Authorization header.
- LLM_ENDPOINT_TIMEOUT_SECONDS (optional): Request timeout in seconds. Defaults to 10.

Retry behavior is built into the task:

- Retries transient failures (network/timeouts and HTTP 429/500/502/503/504).
- Exponential backoff enabled.
- Retry delay is jittered.
- Maximum retries is 7.

## Running a Celery worker

The Celery entrypoint is in src/comment_automation/celery_app.py.

It reads these environment variables:

- CELERY_BROKER_URL (optional): Message broker URL. Defaults to redis://localhost:6379/0.
- CELERY_RESULT_BACKEND (optional): Result backend URL. Defaults to redis://localhost:6379/0.

With uv:

uv run celery -A comment_automation.celery_app:celery_app worker --loglevel=info

Without uv (venv already created):

./.venv/bin/celery -A comment_automation.celery_app:celery_app worker --loglevel=info