-
released this
2026-11-12 14:27:27 +01:00 | 0 commits to main since this releaseRelease Notes - MeDaX Pipeline v1.1.1 (Feature Release)
Overview
This feature release modernizes the build and deployment setup, adds optional authentication and a local FHIR test server, and makes batched imports more reliable for large datasets.
New Features
Deployment & Infrastructure
- Switched from Poetry to
uvwith a locked dependency set; Docker image now uses Python 3.11 - Replaced the bundled BioCypher copy with the official
biocypherpackage (>= 0.16) - New Docker Compose profiles:
blaze(local FHIR server preloaded with POLAR test data),keycloakandproxy(login in front of Neo4j),server(production proxy setup) - Optional TLS for the Neo4j Bolt connection (
ENABLE_BOLT_TLS)
Import Options
- Import a fixed list of patients from a file (
SELECTED_PATIENTS_FILE) DISABLE_IMPORTmodes: run the full pipeline, only load existing import files, or only start Neo4j- Synthea reference handling can now be switched on via
IS_SYNTHEA - Beta support for ODM/MDM data alongside FHIR (
multiple_adapters.py)
Bug Fixes
- References to resources in other batches are now kept with the correct edge label
- The Neo4j import script is generated once after all batches instead of with the last batch only
- Leftover files from earlier runs are cleared before each import, so data is no longer imported twice
- The pipeline no longer hangs indefinitely while waiting for Neo4j or the FHIR server
Documentation & Testing
- README extended with setup guides for each Compose profile and post-processing queries (patient timelines, shared concept nodes)
- New unit test suite covering graph creation, batching and import script generation
Compatibility
- Not compatible with v1.0.1 setups without changes
- Update your
.envusing the new.env.example - Python 3.10 or newer is required for local runs
- Neo4j now uses different mount points, so existing databases from v1.0.1 must be re-imported
Upgrade Recommendation
All users are recommended to upgrade, especially those importing large FHIR datasets in batches. Review your
.envanddocker-compose.ymlbefore the first run.
© 2026 MeDaX Project Team | Released under MIT License
Downloads
- Switched from Poetry to