Appendix A: Source inventory - Terpedia White Paper

The table below is condensed from Terpedia’s operational coverage matrix as of September 2026. It lists the external and licensed sources in the knowledge base, the serving tier each is promoted to, and its ingestion status. Terpedia-internal derived tables, staging tiers, and infrastructure rows are omitted. Status meanings follow the coverage matrix: validated means the raw snapshot and a query-shaped promotion are both working; partial means a bounded or derived subset is serving; raw/document means bytes are mirrored but no validated serving adapter is enabled yet; planned means the adapter is not yet implemented.

Source	Domain	Serving tier	Status (September 2026)
PubChem	Chemical identity, structures, RDF	Fuseki + BigQuery	Partial; repaired descriptor shard loaded, remaining RDF shards staged
PubChem BioAssay	Assay records	Fuseki + BigQuery projections	Partial; XML mirrors complete, RDF promotion pending
ChEBI	Chemical ontology	Fuseki	Validated
Rhea	Biochemical reactions	Fuseki	Validated
UniProt	Proteins	BigQuery + Fuseki (planned)	Planned; adapter not yet implemented
MeSH	Biomedical vocabulary	Fuseki	Raw/manifest validated; Fuseki promotion queued
MONDO	Disease ontology	Fuseki	Raw/manifest validated; Fuseki promotion queued
COCONUT	Open natural products	BigQuery	Validated
LOTUS	Compound–organism occurrences	BigQuery + raw	Structure serving validated; occurrence metadata promotion pending
BRENDA	Enzyme data	BigQuery + raw	Validated; license-scoped
Dr. Duke	Phytochemical and ethnobotanical	BigQuery + document index	Validated
EssoilDB	Essential-oil composition	BigQuery	Validated with quality gaps (39 errors in latest run)
KNApSAcK	Metabolite mass spectra	BigQuery	Validated
TCMID	Traditional Chinese medicine herbs, ingredients, prescriptions	BigQuery + raw	Validated TSV exports
FDA	Regulatory documents and tables	BigQuery / document index	Validated subset
UNII	FDA substance registry	BigQuery	Validated; 159,376 records, 986,598 names
SCOGS	GRAS evaluations	BigQuery	Validated
Cannabis laboratory profiles	Terpene measurements	BigQuery	Validated; 43,018 rows, 409,655 measurements
CannabisDatabase.ca	Compounds, cultivars, proteins, relations	Firestore + BigQuery + raw	Validated; 12,039 records, 54,251 relations (CC BY-NC 4.0)
SAIR	Terpene–protein interactions	BigQuery	Validated with quality gaps (row/reference errors recorded)
SuperNatural II	Natural products	BigQuery	Validated
TeroKit	Terpenome database	BigQuery + Fuseki projections	Validated typed load with classification view
Cellosaurus	Cell lines	BigQuery / Fuseki projection	Validated
NAEB	Native American ethnobotany	BigQuery	Validated
Patents	Patent indexes and documents	BigQuery + document index	Indexes validated; PDFs partial
PubMed (terpene collection)	Literature	Document index / BigQuery	Automated nightly collection
Eden, Ayurvedic, Bach, HPUS, WHO monographs	Pharmacopoeial and traditional-use documents	Document index	Validated subsets
OpenTerps	Community terpene data	Raw	Raw/document
Terport TTL bundles	Terpedia-generated RDF	Fuseki	Validated selected manifest (18 loaded)

Additional acquisition candidates (for example ChEMBL, BindingDB, Open Targets) are tracked in the catalog as candidates and are explicitly not counted as current coverage. The authoritative machine-readable catalog and the full coverage matrix are maintained in the public kb-source repository at github.com/Terpedia/kb-source.

Appendix B: Knowledge API summary¶

Endpoint	Auth	Purpose
`GET /health`	none	Service health
`GET /v1/cyc`	none	Paged list of resolved entities
`GET /v1/cyc/{keyword}`	none	Resolve a compound, cultivar, or protein name; returns entry, related entities, statements with sources, citations, and evidence context
`POST /v1/tabular/search`	key	Search one source table or the unified source union
`POST /v1/tabular/count`	key	Scoped, classified counts with identity keys
`POST /v1/semantic/search`	key	Semantic search with citations and graph context
`GET /v1/entities/{id}`	key	Entity record by identifier
`GET /v1/entities/{id}/neighbors`	key	Graph neighborhood of an entity

Base URL: https://terpedia-knowledge-nanrsdlaoa-uc.a.run.app. Keyed endpoints take the key in the x-knowledge-key header. Cross-origin requests are permitted.