DICOMWeb Server w/ Test Data

THIS REPOSITORY IS NOT FOR PRODUCTION USAGE. It's a very simple way to provide tightly controlled test data to applications that integrate with DICOMWeb.

This repository is used as a version controlled collection of DICOM studies that are made available with lightweight implementation of DICOMWeb or the static wado service. We're leveraging the following projects:

• dicomweb-server: Lightweight DICOMweb Server with CouchDB
• CouchDB: NoSQL document-oriented database solution
• dicomweb-client: Python client for DICOMweb services.
• [dicomp10-to-dicomweb][dicomp10-to-dicomweb]: JavaScript version of a static wado generator for dicomweb

Installing Test Files

Run the two commands below to install mkdicomweb first, and then to run it on the test data, installing the test data to ~/viewer-testdata. The viewer testdata is automatically served by running the yarn start command in OHIF.

npm install -g dicomp10-to-dicomweb
mkdicomweb -d ~/viewer-testdata dcm

None of the remaining options are required.

Requirements

For Running Container:

• Docker: Windows | Mac
• Share your drive w/ docker (on mac your drive will be shared by default)

For Uploading New Studies:

• Windows Subsystem for Linux (or Linux)
  • This is not strictly true, but I find it easier
• Install Python 3 & pip (Linux)
  • apt-get -y install python3 python3-pip
• Install Python 3 & pip (Mac)
  • brew install python3 python3-pip
• Install Dicomweb-client and pydicom (Linux/Mac)
  • pip3 install dicomweb-client pydicom

Usage

Start:

• docker-compose build
• docker-compose up

Add Studies:

• cd dcm
• python3 seed-db.py

Request Data: (QIDO-RS)

• Studies:
  • http://localhost:5985/studies
• Series in Study:
  • http://localhost:5985/studies/<studyInstanceUid>/series
• Instances in Series:
  • http://localhost:5985/studies/<studyInstanceUid>/series/<seriesInstanceUid>/instances

Where UIDs often have a format similar to: 1.2.276.0.74.3.1167540280.200511.112514.1

OHIF Server Config:

By default dicomweb-server has a wide open CORS policy set in it's fastify config. This allows for little additional configuration.

window.config = {
  routerBasename: "/",
  extensions: [],
  showStudyList: true,
  servers: {
    dicomWeb: [
      {
        name: "dicomweb_server",
        wadoUriRoot: "http://localhost:5985",
        qidoRoot: "http://localhost:5985",
        wadoRoot: "http://localhost:5985",
        qidoSupportsIncludeField: true,
        imageRendering: "wadouri",
        thumbnailRendering: "wadors",
        enableStudyLazyLoad: true
      }
    ]
  }
};

Developing

• config/ contains files used by docker images or projects.
• couchdb-data/ is our volume mount point so we can persist our data
  • If you wanted multiple data sets, you could submodule this or choose a different mount point.
• dcm/ contains our seed DICOM files. By default, the seed-db.py file will upload all .dcm files in this directory.

👋 If couchdb-data grows too large; we can simply re-upload dcm files when the volume is spun up.

Tools helpful for creating test data

• MicroDicom (win only)
• img2dcm (command line)
• DicomBrowser (win, mac, CLI)

Resources

• Python DICOMWeb Client Docs