fix: normalize preferred-name ranking comparisons

[SkyeAv]

Fixes preferred-name ranking so the PR boost survives the normalized v6 keys: both sides of the comparison now go through the same Rust-backed normalizer.

Ranking (src/tablassert/fullmap.py)

• Normalized comparison: filter_and_rank normalizes PREFERRED_NAME and term via the existing _normalize_terms_series/rs.normalize_terms batch helper before the level-one preferred-name equality, so reordered (Oral Aspirin) and inflected (Oral Aspirins) preferred names still earn the boost against the stored key aspirin oral.
• Historical branch preserved: the raw exact-match boost stays first and unchanged; rows without preferred names and genuinely non-matching names keep their existing ranking.

Tests (tests/test_fullmap.py)

• New cases with WHY docstrings: reordered preferred name earns the boost (PR=250), inflected preferred name earns it, genuinely non-matching name does not (PR=500); all pre-existing assertions, including test_pr_case_insensitive_preferred, pass unmodified.

Testing

• uv run pytest tests/test_fullmap.py -n 0 -q → 61 passed
• uv run pytest tests/test_nlp.py -n 0 -q → 15 passed
• uv run pyright src/tablassert/fullmap.py tests/test_fullmap.py → 0 errors; ruff check/format --check → clean

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Advanced

Run ID: 7f02cc8b-1daf-412f-88c8-3daf03f532a1

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}