Skip to content

Knowledge-graph schema

Source

The schema lives in schema/*.sql and schema/deep/*.sql.

The graph is a property graph:

  • kg.node(id, slug, kind, label, ...) — entities.
  • kg.edge(id, src_id, dst_id, kind, ...) — relationships.
  • kg.property(entity_id, key, value_text, value_num, value_json, ...) — facts.

572 nodes / 791 edges / 1027 properties at the time of archive.

Conventions worth knowing:

  • Every node has a stable dot-namespaced slug (pathogen.west_nile, county.maricopa). Renames cascade through agents/, dashboards, and the map/graph viewers.
  • Each schema/deep/*.sql seed owns a contiguous edge-ID range (counties 10000-, tribes 11000-, pathogens 12000-, outbreaks 13000-, standards 14000-, datasets_apis 15000-, mcp_servers 16000-, application 17000-, followups 30000-).
  • Tribal data is suppressed by default. Opt-in lives in consent_profile rows, consulted by ValidationAgent at write time.

See Seed load order and Example queries.