literature-search-openalex

Use the filter command with an entity type like works, adding --search for full-text queries or --filter for structured expressions such as publication_year:2023. Always resolve names to OpenAlex IDs first before filtering by author or institution.

First run the resolve command with the author's name to get their OpenAlex ID, then filter works with authorships.author.id set to that ID. Sort by cited_by_count:desc to surface their most influential papers.

No, OpenAlex works without a key using the free polite pool, but rate limits are very restrictive. Providing an OPENALEX_API_KEY enables roughly 10 requests per second and unlocks PDF downloads from the premium content server.

Yes, the download-pdf command fetches PDFs for a given work ID, trying the OpenAlex content server first and falling back to pdf_url locations in the work metadata. Downloaded files are validated for the %PDF- header to catch corrupt or paywalled responses.

A 429 means you hit the rate limit, which is common without an API key since unauthenticated usage is tightly capped. Add an OPENALEX_API_KEY to your .env file to raise limits, or wait and retry with backoff.

Filtering by display name is unreliable because names are ambiguous across authors and institutions. The correct approach is to resolve the name to a canonical OpenAlex ID first, then use that ID in the filter expression for accurate results.