Skip to content

Contributing#

All contributions are welcome!

Setup for local development#

From the /src directory:

docker compose -f hack/compose.yaml --profile=omop up

This will start all development dependencies:

  • OHDSI WebAPI
  • OHDSI ATLAS
  • Broadsea Atlasdb - a pre-filled OMOP database
  • Traefik
  • HAPI FHIR JPA Server
  • Jaeger
  • MailDev
  • Keycloak

If you want to start any of the recruIT modules as containers, you can specify the corresponding --profile switch. For example, when working on the query module, it might be useful to run the screening list and the notify module for debugging. The following will start all development dependencies as well as build and run the list and notify containers:

docker compose -f hack/compose.yaml --profile=omop --profile=notify --profile=list build
docker compose -f hack/compose.yaml --profile=omop --profile=notify --profile=list up

You can then start the query module via gradle by running

./gradlew :query:bootRun

Setup for the Trino SQL-based query module#

Use the trino profile to start the dependencies for using the Trino-based query module:

  • Traefik
  • HAPI FHIR JPA Server
  • Jaeger
  • MailDev
  • Keycloak
  • MinIO
  • Pathling
  • Hive Metastore
  • Warehousekeeper (VACUUMs the Delta Lake tables and registers them in the Hive Metastore)
  • Trino
docker compose -f hack/compose.yaml --profile=trino up

By default, this will also upload sample FHIR resources to the FHIR server and import the same resources via Pathling into Delta Lake tables.

For development, you will also need to upload two FHIR SearchParameter resources required by the query module:

curl --fail-with-body -X POST -H "Content-Type: application/fhir+json" --data @hack/fhir/search-parameters-transaction.json "http://recruit-fhir-server.127.0.0.1.nip.io/fhir"

Afterwards you can upload a sample study with the associated SQL-encoded criteria:

curl --fail-with-body -X POST -H "Content-Type: application/fhir+json" --data @hack/fhir/trino-sql-study-transaction.json "http://recruit-fhir-server.127.0.0.1.nip.io/fhir"

Now, running

./gradlew :query-fhir-trino:bootRun

curl --fail-with-body "http://recruit-fhir-server.127.0.0.1.nip.io/fhir/List"

should create all appropriated resources to appear in the screening list.

Building Container Images#

notify & query#

Both the notify and query module are Java applications that can be build using Gradle. Therefore, they also share the same Dockerfile for building the container and can optionally also be built using jib.

Using Dockerfile#

From the /src directory, run

export MODULE_NAME=query
docker build -t "ghcr.io/miracum/recruit/${MODULE_NAME}:local" --build-arg=MODULE_NAME=${MODULE_NAME} .

The --build-arg MODULE_NAME can be either notify, query-fhir-trino, or query (default).

Using jib#

You can also build the container images using jib via gradle:

export MODULE_NAME=query
./gradlew :${MODULE_NAME}:jibDockerBuild --image="ghcr.io/miracum/recruit/${MODULE_NAME}:local"

Using jibDockerBuild will build the image against the local docker daemon. Running just ./gradlew :query:jib will attempt to push the image to the remote registry.

list#

The list module is a NodeJS app with a Vue frontend and can be built via Dockerfile. From the /src/list directory, run:

docker build -t ghcr.io/miracum/recruit/list:local .

Skaffold#

You can also directly build and deploy to a Kubernetes cluster for development:

Create a KinD cluster and install NGINX Ingress

kind create cluster --config=hack/k8s/kind-with-ingress-config.yaml
kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/static/provider/kind/deploy.yaml

Run skaffold dev:

helm dep up ../charts/recruit/

skaffold dev