There are two bots: poll.py and stream.py.

This bot polls the Twitter API for its followers. It will follow any follower and unfollow ex-followers. It does not use a stream but instead runs the function for following and unfollowing in an infinite while loop, sleeping at the end of each iteration to avoid exceeding the API rate limit.

This bot uses a stream to receive tweets from matching tracked sources in a live feed. It will ignore tweets where the API user is the author and handle other tweets accordingly;

• Reply to a mention which is not also a reply.
• Retweet and like any non-mentions from tracked sources.

The bots have been hosted and run on Heroku, on two free dynos.

You'll need a Twitter developer account to run the bots. Creating one is, unfortunately, not trivial. Be honest and say that you're trying out the API for developer and learning purposes. You'll also need a website URL to register a new app once your account is ready.

I strongly recommend using a virtual environment.

Add your Twitter API credentials to .env-sample and set any additional settings according to the sample file and config.py. Save and rename the file to .env. Once you've made the initial configurations, run each bot in a console with either python -m poll or python -m stream.

Deploying the bots on Heroku is almost as simple as following the official Heroku Python guide on web apps. The Procfile differs, and dynos won't start automatically when the process type is not web. To run the bots, you need to scale the dynos from 0 to 1. The free tier on Heroku gives you access to two free dynos. Scale-up dynos via the Heroku CLI:

$ heroku ps
Free dyno hours quota remaining this month: 550h 0m (100%)
Free dyno usage for this app: 0h 0m (0%)
For more information on dyno sleeping and how to upgrade, see:
https://devcenter.heroku.com/articles/dyno-sleeping

No dynos on ⬢ magicandcode-bot
$ heroku ps:scale stream=1
Scaling dynos... done, now running stream at 1:Free
$ heroku ps:scale poll=1
Scaling dynos... done, now running poll at 1:Free

Scale down dynos (take one or both bots offline):

$ heroku ps:scale stream=0
Scaling dynos... done, now running stream at 0:Free
$ heroku ps:scale poll=0
Scaling dynos... done, now running poll at 0:Free

Note that the instructions are tested on the macOS Terminal only. To use Windows CMD or Powershell, check the official documentation.

• python >= 3.7
• tweepy - Twitter API wrapper
• environs - Environment variables parsing

I use environs, but you can replace it with python-dotenv, os.getenv or similar. If you do, change the code in config.py to use your preferred package or module.

• Update to Twitter API v2
• Use a more modern API wrapper like twython