Automatically test endpoints

[Mateusz-Dobrzynski]

Since our code is going to be used in real tournaments, it would be nice to know, if it works (obviously). Rust is very bug-resistant, but it cannot detect business-logic-related issues (I've already managed to create some). Therefore we should complement out project with automatic tests. We should test endpoints, i.e. send requests to the server and check, whether the responses are correct.

There are a few issues to tackle here:

• Figuring out how to write these tests in the first place. I tried to research that on my own, but with little success.
• DB issues – running the tests would require either mocking the database (which would substantially increase the workload and make the tests less credible) or having a deployed database run the tests on (which might cause problems with GitHub Actions).
• Test coverage – the bare minimum should be testing a happy path for each endpoint. On top of that, we should write a bunch of tests for entire flows (ones described on FigJam). This way we'll be able to guarantee that the app meets requirements we set for it.

[Mateusz-Dobrzynski]

TO-DO:

• Write a script which:
  • Initiates an empty database,
  • Runs test scenarios. A test scenario here means a series of HTTP requests performed one by one in order to simulate particular operations defined in the business logic. These scenarios can be implemented as Pytest tests. The assumption here is that each scenario is independent, meaning that each one starts when the database is empty.
• Create a simple test scenario (e.g. consisting of logging in as an infrastructure admin, creating a tournament, and performing some operations on it).
• Create a template pull request with a checkbox "I have covered my code with tests, run a testing script found at [script path] and ensured that the tests pass.