Replace drf-yasg with drf_specatcular and make swagger enable-able in production via environment variable

[KevinMind]

Relates to: mozilla/addons#15019

Description

Replace drf_yasg with drf_spectacular as the more maintained and v3 supporting library. The new lib has better control over documenting serializers and supports the openapi spec v3 which natively has better type support.

https://drf-spectacular.readthedocs.io/en/latest/drf_yasg.html

Testing

Dev

Run the project

make up DOCKER_TARGET=development

Open the swagger urls:

• schema: http://olympia.test/api/v5/schema?format=json
• swagger http://olympia.test/api/v5/swagger
• redoc: http://olympia.test/api/v5/redoc

These URLs should work, load the appropriate content and there should be no failing asset requests.

Run the generate command

make shell
./manage.py generate_js_swagger_files

Expect files in ./static-build/js/swagger/{version}.js to be created for each supported API version (v3,v4,v5)

Prod

Run the project

make up DOCKER_TARGET=production

Expect the same as dev but the files should already exist without running the generate command as it is run during the build.

Note

Expect the files in both static-build and site-static directories in prod mode
The files will be served from the nginx server and not the olympia server

Checklist

• Add #ISSUENUM at the top of your PR to an existing open issue in the mozilla/addons repository.
• Successfully verified the change locally.
• The change is covered by automated tests, or otherwise indicated why doing so is unnecessary/impossible.
• Add before and after screenshots (Only for changes that impact the UI).
• Add or update relevant docs reflecting the changes made.

[diox]

Should I wait until #22663 is merged before reviewing this ? Or could you remove that commit (6b9dbdc) ? It makes the diff a lot bigger and harder to follow.

[KevinMind]

@diox Reviving this since it has been sitting here and there is nothing stopping its merge. It is bottom of the list in priority but enables some of the work I'm doing on the data retention policy since I want to add better documentation of some of the models.