The scenario reads PDFs from Google Drive, rates each paper against your criteria, sorts the fits by theme into Postgres PGVector Store and Affinity, and sets the rejects aside.
Screening papers by strict criteria
Rated without reading each in full
Sorted into sub-themes
Search by meaning in Postgres PGVector Store
Google DriveAffinityPostgres PGVector StoreDefault Data LoaderEmbeddings Google GeminiQdrant Vector Store