fix: skip non-added mirror files with missing base content in token scoring

[MkDev11]

Summary

calculate_token_score_from_file_changes had no guard for non-added files whose base-side blob is unavailable. When a modified or renamed file arrives with old_content=None and valid new_content, score_tree_diff receives None as the old tree, treats old_signatures as an empty Counter, and counts every node in the head file as a net addition — inflating the miner's token score and downstream reward.

The das-github-mirror fetcher (extractBlobText) returns null for any base blob that is binary, over 1 MB, or unreachable from git. The legacy PAT fetcher in github_api_tools.py applies the same three conditions. Both paths feed into the shared calculate_token_score_from_file_changes, so both OSS and mirror PR scoring are affected.

Fix: one elif guard after skipped-unsupported that skips non-added files with old_content=None (scoring_method='skipped-missing-base', score=0.0). Added files with old_content=None are legitimate new-file contributions and continue to reach tree-diff unchanged.

Adjacent reference: PR #977 added the skipped-large guard for oversized old_content but left the None-for-non-added case unguarded.

Related Issues

Closes #1192

Type of Change

• Bug fix

Testing

• Tests added/updated

test_modified_file_missing_base_content_is_skipped — regression: asserts scoring_method='skipped-missing-base' and score=0.0 for a modified file with old_content=None; was tree-diff before the fix.

test_renamed_file_missing_base_content_is_skipped — regression: same guard for renamed files.

test_added_file_null_old_content_scores_as_new_file — control: added file with old_content=None still reaches tree-diff with score>0.

Full suite: uv run pytest tests/ -v → 1491 passed.

Checklist

• Code follows project style guidelines
• Self-review completed
• Changes are documented (if applicable)