This directory (metadata-models/docs/entities/) contains hand-authored descriptions of DataHub
metadata entities. These files are inputs to the modelDocGen build step — they are not served
directly.
The Gradle task :metadata-ingestion:modelDocGen (runs metadata-ingestion/scripts/modeldocgen.py)
combines three sources to produce auto-generated entity reference pages:
- Entity registry —
metadata-models/src/main/resources/entity-registry.yml - Avro schemas — generated from PDL files during the build
- Entity docs — the markdown files in this directory
Output lands in docs/generated/metamodel/entities/ (gitignored).
Each file is named {entityName}.md (matching the entity name in the registry). The content is
free-form markdown that describes what the entity represents, its key aspects, and how it relates to
other entities. This prose is spliced into the auto-generated reference page.
Existing files to use as reference: dataset.md, dashboard.md, chart.md.
Regenerate and preview:
./gradlew :metadata-ingestion:modelDocGen # regenerate entity reference pages
scripts/dev/datahub-dev.sh docs # preview the docs site