A census (and other details) of the Early Detection Research Network's usage of DICOM tags throughout its flagship data search and archive product, LabCAS.
TBD.
A Unix-like (macOS included) environment is expected. The following are required in order to run locally:
- PostgreSQL 17
- Taskfile 3
- Python 3.11 (later versions not supported)
Use PostgreSQL 17. Run
createdb dicominator
For the suggested schema, see: https://chatgpt.com/share/6807cd23-5f48-8003-8f25-7d7efe7d7dc3
Make any necessary database migrations with
task manage command=makemigrations
Then apply them with
task manage command=migrate
Finally populate the server with
task manage command=dicominator_bloom
task manage command=autopopulate_main_menus
And load DICOM data with
task manage command=dicominator_load -- FOLDER [SLUG]
where FOLDER is a top-level folder containing DICOM files. The extension .dcm of the file doesn't matter; it will try to read every file as a DICOM file. The SLUG is the slug of the PatientIndex that should contain everything. By default it's patients.
Add DICOM tag frequency data with a similar command:
task manage command=scan_tag_frequencies -- FOLDER
To start the server, run
task run
This will build the Python virtual environment if it doesn't already exist and start things up. You'll be able to visit it at http://localhost:6468
If you ever need to rebuild the Python environment, you can destory the old one with
task clean
The next time it's needed it'll be rebuilt.
Just type
task --list
to see what's possible.
The variables are:
| Variable | Purpose | Default |
|---|---|---|
DICOMINATOR_VERSION |
Image version | latest |
CERT_CN |
Common name in generated TLS certificate | edrn-docker.jpl.nasa.gov |
HTTPS_PORT |
Listening port for https |
2371 |
LDAP_URI |
Where to find the LDAP server | ldaps://edrn-ds.jpl.nasa.gov |
LDAP_BIND_DN |
Distinguished name of service account | uid=service,dc=edrn,… |
LDAP_BIND_PASSWORD |
Credential for service account | (unset) |
POSTGRES_PASSWORD |
Password to Postgres DB | dicominator on CLI or unset |
SIGNING_KEY |
Secures Django data | s3cr3t or set via .env |
DATA_DIR |
Persistent Docker data | See docker-compose.yaml |
Secrets may be set in ~/.secrets/passwords.sh or as environment variables.
| Secret | Purpose | Place |
|---|---|---|
DOCKERHUB_USERNAME |
User ID in DockerHub | jpl-labcas organization secrets |
DOCKERHUB_TOKEN |
Token for DOCKERHUB_USERNAME |
jpl-labcas organization secrets |
GitHub Actions takes care of automatically creating the Docker image for this.
To do it by hand:
support/build-wheels.sh
docker image build --tag edrndocker/dicominator:latest --file docker/Dockerfile .
Spot check:
docker container run --env LDAP_BIND_PASSWORD=x --env SIGNING_KEY=x --rm --publish 8000:8000 edrndocker/dicominator:latest
And visiting http://localhost:8000/ should give you "Bad Request (400)".
You can then launch the composition with
task comp-up
Populate with
docker compose --project-name dicominator --file docker/docker-compose.yaml exec db createdb --username=postgres --encoding=UTF8 --owner=postgres dicominator
task comp-app-exec -- /app/bin/django-admin migrate
task comp-app-exec -- /app/bin/django-admin dicominator_bloom
task comp-app-exec -- /app/bin/django-admin autopopulate_main_menus
task comp-app-exec -- /app/bin/django-admin scan_tag_frequencies /mnt/data
task comp-app-exec -- /app/bin/django-admin dicominator_load /mnt/data
You can then visit https://localhost:9999/dicominator/ (ignore the certificate warning).
The following details pertinent to this project are described in the remaining sections.
TBD.
We use the SemVer philosophy for versioning this software. More TBD.
You can contact
The project is licensed under the Apache version 2 license.
None at this time.