Food Pics

This is a script that connects to Reddit, uses a configured delimited list of subreddit names, scrapes the subs for "hot" posts, and submits a random one to a configured Discord webhook.

The original idea for this came from https://github.com/SaxyPandaBear/discord-food-bot

Configurations

Configurations for the script are expected to be in environment variables. Redis variables are injected by Railway (docs)

Key	Description
REDISHOST	hostname to connect to Railway Redis instance
REDISUSER	Redis user
REDISPORT	Port to connect to on Redis instance
`REDISPASSWORD	Redis password
WEBHOOK_URL	URL for Discord channel webhook
REDDIT_CLIENT_ID	client ID for Reddit API access
REDDIT_CLIENT_SECRET	client secret for Reddit API access
SUBREDDITS	subreddits to scrape split by +, e.g.: foo+bar+Baz
LIMIT	(optional) batch size for scraping from Reddit

Dependency management

This project uses pipenv, which has it's challenges.

Railway requires a requirements.txt file to detect a Python app, so run:

pipenv requirements > requirements.txt

after updating dependencies in order to propagate them on deployments.

How does it know what webhook to push to?

Currently it's only pushing to one webhook, defined in the WEBHOOK_URL environment variable (in Heroku), but can split this out into separate variables as necessary in the future.

How does this prevent duplicate posts?

There are a couple methods to try to deduplicate Reddit submissions.

1. Perform exact match on submission IDs
2. Compute image hash and compare with other image hashes from the same author
3. Fuzzy match submission title with same author, from the past week

Testing

Run Python unit tests as follows:

# From the root of the repository
pipenv run python -m unittest  # will find all files prefixed 'test_' and execute them

An example output of this:

$ python3 -m unittest
......
----------------------------------------------------------------------
Ran 6 tests in 0.000s

OK

Linting and Formatting

Uses black and pylint for formatting and linting.

pipenv shell
(pipenv) black .
(pipenv) pylint --rcfile ../.pylintrc .

Troubleshooting

Clear Redis data

TODO: update this since not using Heroku anymore

1. Log in to the Heroku CLI
2. heroku redis:cli -a the-app-name
3. flushall