pubmed-database

Run the pubmed_api.py CLI with the search_pubmed function, passing a free-text or structured NCBI query. It returns a JSON list of PMIDs that you can pipe into fetch_article_abstracts to retrieve titles, authors, and abstracts.

Use find_linked_biological_data with a target database such as gene or pccompound and the appropriate linkname like pubmed_gene. It returns NCBI record IDs, which fetch_database_summary resolves into gene names or compound synonyms.

Yes, but only for articles in the PMC Open Access Subset via get_full_text_pmc through the BioC API. Paywalled or embargoed articles return an error, so fall back to fetch_article_abstracts for those papers.

NCBI cross-database links lag publication by four to eight weeks due to indexing delays. For very recent papers, search the abstract or full text directly for accession numbers and identifiers instead of relying on ELink.

Use match_raw_citations with a pipe-delimited string in the format journal|year|volume|first_page|author|key|. If ecitmatch fails, fall back to search_pubmed with the title, author, or DOI as the query.

No, the API works without a key at 3 requests per second. Setting the NCBI_API_KEY environment variable raises the limit to 10 requests per second, which is recommended for bulk queries or if you encounter 429 errors.