Kescience Pubmed

Why This Collection Matters

kescience_pubmed is a live BERDL database in the KE Science tenant. This scaffold page makes the collection visible to humans and agents even before a domain-specific curator has written a richer interpretation.

What It Contains

The live Spark inventory records 4 tables. Key tables include: pubmed_article_chemical_fact, pubmed_article_id_fact, pubmed_article_mesh_fact, pubmed_article_wide

Best Uses

Use this page as an entry point for deciding whether the database should be promoted into a curated collection, linked to a data-type page, or used as a source for a derived product.

High-Value Joins

Start by inspecting stable identifiers in the table schemas, then connect through the linked data-type page. For user, staging, or imported namespaces, write a join recipe before using the collection as evidence for a claim.

Reuse And Gaps

High-value reuse would turn this raw collection into a documented mapping table, benchmark, feature matrix, annotation layer, or caveat label. Missing complementary data should be recorded as a data_gap page once the blocker is clear.

Caveats

• The live Phase 1 refresh used Spark SQL table inventory and intentionally skipped per-table schemas to avoid long-running DESCRIBE calls across user and staging namespaces.