There is a binary DICOM to static file generator, which provides easily served binary files. The files are all compressed in order to reduce space significantly, and are pre-computed for the files required for OHIF, so that the performance of serving the files is just the read from disk/write to http stream time, without any extra processing time.
The project for the static wado files is located here: [static-wado]: https://github.com/wayfarer3130/static-wado
It can be compiled with Java and Gradle, and then run against a set of dicom, in the example located in /dicom/study1 outputting to /dicomweb, and then a server run against that data, like this:
git clone https://github.com/wayfarer3130/static-wado.gitcd static-wado./gradlew installDistStaticWado/build/install/StaticWado/bin/StaticWado -d /dicomweb /dicom/study1cd /dicomwebnpx http-server -p 5000 --cors -g
There is then a dev environment in the platform/viewer directory which can be run against those files, like this:
cd platform/vieweryarn dev:static
Additional studies can be added to the dicomweb by re-running the StaticWado command. It will create a single studies.gz index file (JSON DICOM file, compressed) containing an index of all studies created. There is then a small extension to OHIF which performs client side indexing.
The StaticWado command also knows how to deploy a client and dicomweb directory to Amazon s3, which can then server files up directly. There is another build setup build:aws in the viewer package.json to create such a deployment.