Unify HTTP 400 responses

[fridex]

Describe the bug

User API responds with 400 HTTP status codes for invalid requests. These requests can be invalid on multiple layers though. They can be invalid based on checks that are done in implemented endpoint handlers but also based on schema validation. The schema validation issue reports are handled by connexion and differ from the ones that are implemented by us in the Python code. They do not conform to HTTP 400 schema that is declared on Python endpoints.

To Reproduce
Steps to reproduce the behavior:

1. As an example, use a CLI tool to interact with API and ask for results on an endpoint but do not provide required input:

curl -X 'GET' \
  'http://stage.thoth-station.ninja/api/v1/python/package/version/metadata?name=flask&version=0.12&index=https%3A%2F%2Fpypi.org%2Fsimple&os_version=8&python_version=3.8' \
  -H 'accept: application/json'
{
  "detail": "Missing query parameter 'os_name'",
  "status": 400,
  "title": "Bad Request",
  "type": "about:blank"
}

2. See the response is not conforming to the OpenAPI spec

Expected behavior

The response should be conforming to the OpenAPI spec.

Additional context

I've tried to register a custom flask for 400 or providing a custom connexion validator. None of these worked though - I did not invest much time into this, but it might be nice to have.

[goern]

/priority important-soon
/triage accepted

[sesheta]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

/lifecycle stale

[harshad16]

/lifecycle frozen

[harshad16]

seems like: #1766