# Data Tools & Services

Supporting data tooling and services for the warehouse: dbt, the Glue Catalog, and the (decommissioned) Vocabulary Service.

## Data Build Tool (dbt)

dbt is the canonical authoring tool for everything in `CLINICOGENOMICS` — staging views, intermediate Dynamic Tables, and partner cohort marts all live as dbt models. Deployment to each environment is automated through GitLab CI/CD on merge to master.

```bash
uv tool install dbt-core==1.11.8 --with dbt-snowflake==1.11.4 --python 3.13
```

For the repository structure, materializations, and the full build pipeline, see **dbt & Build Pipeline**.

## Glue Catalog

The AWS Glue Catalog provides table metadata and schema discovery for data assets outside the Snowflake warehouse. It complements the Snowflake-native **Schema Reference** for the broader data platform.

## Vocabulary Service

> **Decommissioned**
>
> The Vocabulary Service is decommissioned — its endpoints no longer respond. Vocabulary mapping now happens inside the Clinical LLM extraction pipeline and the dbt mart models.

It was a standalone term-mapping service (fuzzy matching, common ancestors, level alignment) fronting a SciGraph graph database and a Postgres similarity store. Vocabulary and code-mapping now live in the `cgdb/clinical-llm` repo and dbt models under `cgdb/snowflake/dbt/models/`.