[beanbean9339]

Closes #68

Independently from PR #64, scan up to 100 recent commits without a token or 1,000 with a token for usable human author names. Exclude known GitHub logins, handles and bot identities; add history authors after contributor/co-author results and preserve deduplication. Includes focused pagination tests and updates importer fixtures to expect 100 commits on the initial request.

Validation: npm run test (106 passed), npm run build.

[Copilot]

🟡 Changes recommended

Commit-history author handling can admit excluded bot identities and aliases, omit legitimate hyphenated names, and change existing metadata author ordering.

4 open issues

• Keep history authors out of metadata ranking · New
• Exclude known bot identities from history authors · New
• Allow full hyphenated human names · New
• Reject handles and emails as unlinked author names · New

What changed in this PR

This PR extends OpenCite’s GitHub importer to discover authors from recent commit history when repository metadata or contributor profiles are incomplete, addressing #68 separately from contributor fallback limits.

Changes:

• Adds bounded history scans of up to 100 commits without a token or 1,000 with a token.
• Filters discovered identities and merges history authors after co-author and contributor results with deduplication.
• Adds focused pagination/filtering tests, updates importer fixtures, and documents scan limits.

File	Description
tests/​services/​githubImporter.test.js	Updates fixtures for initial 100-commit requests.
tests/​services/​commitAuthors.test.js	Tests pagination, filtering, and unauthenticated scan limits.
src/​services/​githubImporterContributors.js	Exposes contributor logins for identity filtering.
src/​services/​githubImporterCommitAuthors.js	Implements bounded commit-author discovery.
src/​services/​githubImporter.js	Integrates history authors into imported author lists.
src/​services/​githubApi.js	Adds commit-list pagination support.
README.md	Documents history discovery and scan limits.

_{Review effort: Balanced}

[Copilot]

🟡 Changes recommended

Commit-author filtering can import GitHub handles as citation names, and the merge test does not verify that history-only authors are actually added.

1 open issue

• Test importing a distinct history-only author exactly once · New

1 previously missed issue

In code that hasn't changed since last review

Filter commit author names against logins from all scanned commits

src/​services/​githubImporterCommitAuthors.js:34

This set contains only contributor API logins, not logins exposed by the scanned commits. If the contributor API omits an account, an anonymous commit named Jane-Doe is accepted even when another scanned commit identifies Jane-Doe as its GitHub login, importing a handle as a citation name. Collect logins from all scanned commit author records and filter candidate names against the completed set, including logins discovered on later pages.

4 resolved since last review

• Reject handles and emails as unlinked author names
• Allow full hyphenated human names
• Exclude known bot identities from history authors
• Keep history authors out of metadata ranking

_{Review effort: Balanced}