[wanjinhao1]

Closes #14142

Fix. Search fields were checked case-insensitively but sent to Solr with the case the user typed. As a result:

• /search?q=ISBN:…, LCC:…, DDC:…, ID_GOODREADS:…, Edition.ISBN:… and Work.Title:… returned 500 (ValueError: Unknown field in convert_work_field_to_edition_field).
• /search.json returned no results for those same queries.
• Author, subject and edition search silently returned nothing. For example, on production /search/authors.json?q=NAME:Twain finds 0 authors, while name:Twain finds 128.

Technical

• New SearchScheme.canonical_field_name(field). It matches a field case-insensitively, resolves unprefixed aliases from field_name_map, and applies the scheme's own is_search_field. It returns None for unknown fields.
• process_user_query uses canonical_field_name both to decide which fields are valid and to rename them after parsing. Every scheme (works, editions, authors, subjects, lists) now sends Solr canonical, lowercase names.
• Alias resolution moves out of WorkSearchScheme.transform_user_query. It now only does the ISBN/LCC/DDC transforms, on names that are already canonical. This replaces the first commit's works-only normalization.
• Behavior change on /search/editions.json: the edition scheme's publishers → publisher alias now applies. publishers: used to match nothing; now it matches.
• Unchanged: lowercase queries produce identical output. Unknown fields are still escaped to plain text. Prefixed aliases (work.author) are still not resolved, as on master.

Testing

• New rows in QUERY_PARSER_TESTS cover uppercase ISBN/LCC/DDC, aliases, negated and grouped fields, LCC ranges, id_*, edition./work. prefixes, unknown fields and *:*. The new schemes/tests/test_scheme.py covers canonical_field_name and every scheme. The new rows fail on master.
• A differential run of about 4,000 generated queries, in lowercase, upper and title case, through all five schemes: lowercase output matches master except the edition alias above, and every cased variant now matches its lowercase form.
• Local stack, master vs this branch:

query	master	this branch
/search?q=ISBN:9780060751043	500	303 to the work
/search?q=LCC:PS1303*	500	200
/search?q=DDC:818.402	500	200
/search?q=ID_GOODREADS:123	500	200
/search?q=Edition.ISBN:9780060751043	500	200
/search/authors.json?q=NAME:…	0 results	1
/search/subjects.json?q=NAME:…	0 results	1
/search/editions.json?q=publishers:…	0 results	1

Lowercase, alias, plain-text and malformed queries are unchanged.

Screenshot

No UI changes.

Stakeholders

@RayBB @cdrini

---

Attribution disclaimer: substantial parts of this change were developed with AI coding assistance; the approach follows the analysis in issue #14142.

[openlibrary-bot]

Thanks for the PR, @wanjinhao1!

The linked issue hasn't been triaged yet — triage happens on Mondays and Fridays. There are currently 120 open non-draft PRs ahead of yours.

PR triage checklist (maintainers / Pierre)

• PR description — not empty; explains what the change does and how to verify it
• References an issue — PR body contains a #NNN reference
  • Linked issue is triaged — has a Priority: * label (not just Needs: Triage)
  • Linked issue is assigned — has at least one assignee
• Commit history clean — no WIP/fixup/conflict noise; commit messages are meaningful
• CI passing — no failing check-runs
• Test cases present — if the change touches substantive logic, test coverage exists or is explained
• Proof of testing — PR body includes a description of what was tested, a screenshot, or a video

---

Note

This comment was automatically generated by Pierre, Open Library's PR first-touch assistant, on behalf of @mekarpeles. This is NOT a code review or sign-off.

[RayBB]

Thanks @wanjinhao1, great find, and the clear write-up made this easy to dig into!

Your fix worked for ISBN:/LCC:/DDC: in works search. While testing it, I found the same bug one level up: SearchScheme.process_user_query checks field names case-insensitively but never rewrites them. That meant author, subject, list and edition search had it too: /search/authors.json?q=NAME:Twain finds 0 authors on production, while name:Twain finds 128.

I pushed two commits on top of yours:

• a merge of master;
• a new SearchScheme.canonical_field_name(), which both validates and renames fields for every search type. It replaces your elif in works.py. I kept your test rows and added tests for the other schemes.

I've also updated the title and description to match. Verified on testing: the cased /search pages that 500 on production now load, and NAME:/ISBN: style queries return the same results as lowercase across works, authors, subjects, lists and editions. Thanks again!