Fix Cartesian product in contributor metrics

[EngCaioFonseca]

Description

• Please include a summary of the change.

The contributor queries compare commit_comment_ref.cmt_id to itself instead of joining comments to their commits. This matches commits with unrelated comments, inflating contribution counts and causing excessive query work on large repositories.

This PR fixes #
Replace the condition with commits.cmt_id = commit_comment_ref.cmt_id in both the repository and repository-group queries, matching the existing approach in contributors_new().

Notes for Reviewers
Both were verified affecting subqueries using in-memory SQLite fixtures covering exact counts, repository/date filtering, and empty results. The fixtures reproduce the original bug.

Signed commits

• Yes, I signed my commits.

Generative AI disclosure

Please select one option:

• This contribution was NOT assisted or created by Generative AI tools.
• This contribution was assisted or created by Generative AI tools.

If AI tools were used, please provide details below:
- What tools were used? GPT 6 Astra
- How were these tools used? Running an "audit" (for my own knowledge, deepening my knowledge on collectoss), exploratory investigation into CollectOSS and potential for optimizations.
- Did you review these outputs before submitting this PR? Yes.

[MoralCode]

it would be nice to see this issue demonstrated via a live API response if possible.

in-memory SQLite fixtures

CollectOSS doesnt use sqlite. Can you provide more details on how you tested this so the test can be replicated?