Keep Unicode letters in benchmark name searches

[rudycelekli]

Problem

The website removes all non-ASCII letters when folding names and queries. Searches for τ, π or Φ become empty even though the committed Claire Radar snapshot contains τ-Rec, π-SUB and Φ-Bench.

Change

Retain Unicode letters and numbers while keeping punctuation folding. Do not add fields, source preferences or a new ranking policy.

Verification

Three production-search regressions consume every record name in the committed Claire Radar CSV. All three fail before the fix and pass afterward. Tests were added to the existing score-filter module.

All six required checks passed in a clean detached checkout at 27a0d380f21d33c0ed9b9254344322a53771fc71, with its own editable Python 3.12 environment and initialized paper submodule:

• ruff check . — passed
• ruff format --check . — passed
• benchmark-radar normalize-catalog — passed
• benchmark-radar classify — passed
• benchmark-radar build-data-release — passed
• pytest -q — passed

Pytest: 1643 passed in 88.32s (0:01:28). Node 24.19.0 was used for the existing JavaScript consumer tests; upstream CI uses Node 22.

Notes

This is independent of the existing description-search PR: that patch adds a searched field but does not preserve these letters.

No source-data or generated-data edits are included. The clean rebuild preserved 3,212 benchmark records across 5 sources. Index keys and shard slugs exactly matched the generated CLI archive; no source records were merged or removed.

Prepared with Codex assistance. The commit is signed and includes a Signed-off-by trailer.

[rudycelekli]

330226 gpt-6.1-sol

[Copilot]

Copilot review overview

🟢 Approval recommended

The focused fix preserves existing search semantics and has representative full-snapshot regression coverage.

Review effort: Balanced
Findings: None

What changed in this PR

Preserves Unicode letters and numbers in website benchmark searches without changing ranking behavior.

Changes:

• Updates name folding to retain Unicode letters and numbers.
• Adds regressions for τ-Rec, π-SUB, and Φ-Bench using the complete Claire snapshot.

File	Description
site/​assets/​app.js	Uses Unicode-aware search folding.
tests/​test_score_filters.py	Tests production search against Claire records.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[ktwu01]

Closing as not planned. One author now has a cap of 10 open issues + PRs, and this account had 56 open PRs and 11 open issues, mostly paired with each other. We merged the three changes that fix silent data loss in the corpus (#723, #774, and #773 once rebased) and are closing the rest so review stays possible.

This is not a verdict that the change is wrong. If you think this one matters, reopen it later when you are under the cap, one at a time, with a reproduction from a real user or caller rather than a hypothetical edge case.

不是否定这个改动本身。为保证评审能跟上，同一作者最多同时保留 10 个未关闭的 issue/PR；该账号曾有 56 个未关闭 PR 和 11 个 issue。已合并修复语料静默丢失的三个改动，其余先关闭。若认为本条重要，请在低于上限后逐个重新打开，并附真实用户或调用方的复现。

---

Generated by Claude Code