Add the CRUD endpoints to manage code references against a given feature

[matthewelwell]

Task list

• Define the data model to store code references in the database.
  Recommendation:

  FeatureFlagCodeReferencesScan:
      project: Project
      created_at: datetime
      repository_url: str
      revision: str
      code_references:
        - feature_name: <str>
          file_path: <str>
          line_number: <int>
        - ...
      ...
  

• Implement view to submit code references (from CI).

  # /api/v1/projects/{project_pk}/feature-code-references/
  repository_url: https://github.com/Flagsmith/flagsmith
  revision: <commit_hash>
  code_references:
    - feature_name: segment_cloning
      file_path: api/something/foo/bar.py
      line_number: 32
    - feature_name: <str>
      file_path: <str>
      line_number: <int>
    - ...

• Fetch code references for one feature flag (for the feature detail UI).

  # /api/v1/projects/{project_pk}/features/{feature_pk}/code-references/
  - repository_url: <str>
    last_successful_repository_scanned_at: <timestamp>
    last_feature_found_at: <timestamp>
    code_references:
      - file_path: <str>
        line_number: <int>
        permalink: <str>  # For now only GitHub
        vcs_revision: <str>
      - ...
  - 

• Add code reference counts to the feature list endpoint (for the list UI).

  # /api/v1/projects/{project_pk}/features/
  ...
  results:
    - id: <int>
      name: <str>
      ...
      code_references_counts:
        - repository_url: <str>
          count: <int>
          last_successful_repository_scanned_at: <timestamp>
          last_feature_found_at: <timestamp>
        - ...
    - ...

• POST to the new endpoint from the CI job (Build the github action logic to run on a crontab to identify code references for each feature in a given project #5731).

Note

• URLs and schemas above were aligned between @Zaimwa9 and @emyller in effort to keep things simple though flexible for extension after the PoC.
• JSON is suggested as to avoid flooding the database with a multi-table spamming of tuples.
• We might need indexes to help filtering in JSON blobs by "feature_name", "created_at".