Introduce pytest-based testing and CI via GitHub Actions

[VamatoHD]

The beginning of the implementation of tests. I've made some to test Result and Printer, but they need improving.
Also, I have no clue how to work with GitHub workflows, so @kaifcodec you must do it.

To Do:

• Add GitHub workflows
• More tests for the updating logic
• Tests for orchestrator.py
  • Test if the output is valid JSON and CSV
• Tests for utils.py
• Tests for version.py
• Some automatic way to test each module

[kaifcodec]

@VamatoHD Ok, no worries, I will look at it when i am free, maybe tomorrow.

[kaifcodec]

@VamatoHD What are the changes in the last commit (d168a04)?

[VamatoHD]

@kaifcodec Removed AnyResult type, since Result is fully integrated. Changed monkeytype and telegram return type to be Result. In orchestrator.py, the URL was still being set and is_last had type int when it's a bool.
This way, mypy doesn't error if it gets used to check the code before any PR.

[kaifcodec]

Oh, got it.

Anyways I added many tests in the recent commits and they are still not ready so I will wait for it and will try to add a CI workflow next.

[VamatoHD]

Also, in the new test_orchestrator.py, the line assert any(p.startswith("user") and len(p) > len("user") for p in perms) shouldn't be assert all(p == "user" or (p.startswith("user") and len(p) > len("user")) for p in perms)? To check if all permutations have the prefix of the initial username.

[kaifcodec]

You're right, all() better matches the function’s contract.
The test enforces the invariant instead of just detecting one valid case.

I wrote all the codes fast and didn't test it for edge cases, that's why it needs more refinements.

[kaifcodec]

@VamatoHD I merged this PR earlier to get a better sense of control, but everything looks fine now. I had to do a bit of trial and error since our codebase wasn’t very ruff-friendly, so I made the necessary fixes and it’s working well now.