fix: resolve 'bytes' object has no attribute 'decoded' for large CODEOWNERS files

[jmeridth]

Pull Request

Proposed Changes

Fixes #378

What

Normalize CODEOWNERS file contents to bytes before processing, so both the normal path (Contents object with .decoded) and the large-file path (raw bytes from blob().decode_content()) use the same type downstream.

Why

When a CODEOWNERS file is large enough that the GitHub API returns null content, the code fetches it via blob().decode_content() which returns raw bytes. The code then called .decoded on that bytes object, crashing with 'bytes' object has no attribute 'decoded' and preventing PR creation for user removal.

Notes

• The isinstance check in get_usernames_from_codeowners handles both bytes lines and str lines, since the function now accepts raw bytes or str directly
• Existing tests updated to pass bytes/str directly instead of mocking .decoded, matching the new contract

Testing

• Added regression test test_get_usernames_from_codeowners_with_raw_bytes that passes raw bytes (simulating the large file path) and verifies usernames are extracted correctly
• Updated existing tests to pass bytes/str directly to get_usernames_from_codeowners instead of using a mock with .decoded
• All 63 tests pass with 96% coverage

Readiness Checklist

Author/Contributor

• If documentation is needed for this change, has that been included in this pull request
• run make lint and fix any issues that you have introduced
• run make test and ensure you have test coverage for the lines you are introducing

[zkoppert]

Looks great - the normalization to codeowners_decoded handles both paths cleanly, and the isinstance check in get_usernames_from_codeowners is a good move for handling both bytes and str. The regression test covers the new code path well. 🚀