Fix corroborated series titles in downloaded-file review (NOR-87)

[logabell]

Downloaded audiobooks tagged only with a series and position, such as Robert Langdon, Book 1, were rejected for the requested book even when the download and catalog identified it. A matching edition identifier also reintroduced a title conflict for otherwise accepted series annotations.

Use one request-scoped title decision for download review, candidate matching, automatic import, edition creation, and plan confirmation. Series-only titles require a complete single-book transfer, matching source title/author/medium, matching file authors, and fresh, unique, published catalog membership. Wrong positions, conflicting membership, collections, partial files, and exact-edition requests retain their existing holds. Full-title annotations now accept equivalent &/and spelling. File-created editions retain the canonical book title.

Validation through the bounded workstation runner:

• 139 backend integration checks passed across series review/import, matching, collections, AudiobookBay, shared packs, and single-file acquisition. The 23 series cases passed again after final review.
• 176 focused parser/title/linked-download unit checks passed.
• 11 mocked browser download-review journeys passed, including series titles with and without edition identifiers; production frontend build passed.
• Ruff, formatting, and diff checks passed; OpenAPI and TypeScript contracts regenerated.

Integration coverage uses a disposable PostgreSQL database, real media files and filesystem publication, and fixture download/library services. It verifies preserved bytes, one publication on repeated confirmation, and rejection through both review and direct confirmation APIs. Live external services were not exercised.

Fixes NOR-87

[linear-code]

NOR-87

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-15T04:59:51.790392Z	46b35a3	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 46b35a3493

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Reuse request identity across file groups

For a ready inspection containing many groups, this awaits file_match sequentially for every group. Each call unconditionally reloads the same user, catalog series, and metadata through identity, and a nonmatching title can trigger several additional queries in corroborated_titles; grouping permits inspections with thousands of files/groups. The previous implementation loaded identity once outside this loop, so opening a large download review now produces thousands of serial database round trips and can time out. Load the request-scoped identity once and reuse it, invoking the expensive corroboration only for groups that can qualify.

Useful? React with 👍 / 👎.