First beta release
Pre-release
Pre-release
This is the first beta release. The released docker image is available here.
Main features
- Pipelines to fetch resources from a FHIR server and transform them into Analytics friendly formats:
- Input modes include: FHIR-search API, direct JDBC connection (for HAPI FHIR or OpenMRS), and JSON files.
- Output modes include: Parquet files, another FHIR server, or PostgreSQL JSONB based tables.
- Pipelines implemented using Apache Beam making it possible to run on different local or cluster environments such as Apache Flink, Apache Spark or Dataflow.
- Support bulk fetch of all resources and incremental updates.
- A controller to manage and monitor continuous updates of the data warehouse:
- Simplifies monitoring of pipelines, shows progress, vital pipeline execution metrics, error logs, etc.
- Automatically updates table definitions if a Spark Thrift-server is configured.
- Updated version of the Bunsen library to convert FHIR resources to SQL-on-FHIR based records.
- Added FHIR R4 support.
- High fidelity with SQL-on-FHIR on BigQuery.
- A set of docker based configurations to demonstrate sample deployment options.
- Cloud Build based end-to-end tests continuously run on each PR.
- A corpus of Wiki pages on how to use the components, connect them to external tools, and query the data-warehouse.