Simplify request parsing and validation

From @mtlynch - We’re going to fix this for existing parsers, but in the meantime, we’re not going to add new validators. We’ll just do parsing and validating in the same place.

---

The request_parsers package contains logic that’s mostly concerned with extracting data from a JSON payload. The validation of the data is split off into a sub-package validators.

• Both the implementation and the tests of the request parsers are quite repetitive, while the parsing logic in itself is rather trivial.
• The fact that we keep parsing and validating separately, “forces” us to write redundant tests. See e.g. the video settings, where the same logic is tested in both the parser and the validator.
• The request parsing tests all have their own definition of make_mock_request, which we have 7 identical copies of in Pro.
• We are not consistent with the code style: in some cases, we just call it parse, whereas in others we call it parse_XYZ.

I think it would be worth to re-evaluate the current pattern and see whether we can simplify and streamline it.

cc @jdeanwallace

[mtlynch]

Nice! Great work, @rw4nn!

[jotaen]

🥳 Awesome @rw4nn, I think we’ve reached a simple and clean state now!