Goal
1,602 wiki pages have empty refs_json. For each of 40 wiki pages: (1) query PubMed for papers relevant to the page title/entity, (2) select the 3-5 most relevant papers by relevance to the entity's neurodegenerative context, (3) populate refs_json with DOI/PMID/title/authors/year fields. Focus on gene, protein, and mechanism entity pages.
Acceptance Criteria
☐ 40 wiki pages updated with refs_json containing at least 3 citations each
☐ Each updated page has at least 3 inline [PMID:...] markers in content_md
☐ Each updated page has a ## References section listing cited papers
☐ Updates persisted via scidex.core.db_writes.save_wiki_page()
☐ Verification confirms all 40 target slugs meet the inline citation threshold
Approach
Select 40 high-word-count pages from genes/proteins/mechanisms categories still missing refs_json
Implement a batch script modeled after scripts/add_inline_pmid_citations_batch.py
Use PubMed E-utilities to find relevant papers for each page
Inject citations into content_md and populate refs_json
Run the batch and verify resultsDependencies
- Sibling task b13ea762: Added inline citations to 20 pages (same pattern)
Dependents
Work Log
2026-04-26 PT — Slot
- Read AGENTS.md and verified task is still necessary: 1,602 pages still have empty refs_json
- Selected 40 target pages from genes/proteins/mechanisms categories with word_count > 300
- Implemented
scripts/add_inline_pmid_citations_batch_2235a3e1.py modeled after the existing batch script
- Ran script and verified all 40 pages updated with 3+ inline PMID citations