Feat/25 clean up workflow and schema

[radishmouse]

Closes #18 #19 #23 #27 #24 #25

This PR:

• Deletes the extraneous schema files #18
• Adds description kwargs to pydantic class #19
• Concatenates the yaml files and writes a single json/notifcations.json #27
• Uses the schema generated by the pydantic class, which properly handles missing properties #23
• Replaces Go validator with a Python #24

[radishmouse]

@aaspinwall and @MelissaAutumn - The validate_json check fails because there is no existing json/notifications.json file. However, the convert_yaml_to_json check passes, creates the json/notifications.json, and subsequently runs the validate_json check.

Since there are no notifications, we can disregard the failing standalone validate_json check.

But also, I'm wondering: when exactly should we run the validate_json action? It is implicitly run via convert_yaml_to_json when:

• a push is made to main and there are changes to files in yaml/
• there is a pull request on any branch

When should we explicitly run validate_json? Currently, it runs on any pull request.

[Sancus]

General comment: There is still some 4-space indention in generate_json_schema.py but don't 2-space indent Python, it should always be 4. We probably ought to decide on a standard linter and rules for all of Python repos, something for later.

[MelissaAutumn]

General comment: There is still some 4-space indention in generate_json_schema.py but don't 2-space indent Python, it should always be 4. We probably ought to decide on a standard linter and rules for all of Python repos, something for later.

For most of our projects we have ruff setup and some settings setup. https://github.com/thunderbird/appointment/blob/main/backend/pyproject.toml#L66

I have setup in my ide to run on save and format. Works well!