truskovskiyk / dicom_app

DICOM records

This is WIP applications

Production url

• API - http://104.198.185.94:8080/
• DICOM storage SCP - host 35.202.37.159 port 8083

Test

TODO

Local development

TODO

Deploy [manual so far]

We use kubernetes in GCE for a deploy You will need 3 components for the deploy

NFS server

1. Create persistent disk

gcloud compute disks create --size=10GiB --zone=us-central1-a dicmo-app-storage

2. Deploy NFS to k8n

kubectl apply -f ./deployment/nfs/nfs-deployment.yaml
kubectl apply -f ./deployment/nfs/nfs-service.yaml

3. Get NFS ClusterIP

kubectl get service nfs-server

DICOM storage SCP (receiver)

more info here

1. Build docker images for dcmrecv

cd ./dcmrecv/
docker build -t dcmrecv .
docker tag dcmrecv:latest truskovskyi/dcmrecv:latest
docker push truskovskyi/dcmrecv:latest
cd ./../

2. Deploy

kubectl create -f ./deployment/dcmrecv/dcmrecv-deployment.yaml
kubectl expose deployment dcmrecv --type=LoadBalancer --name=dcmrecv-service

3. Get DICOM storage SCP IP

kubectl get service dcmrecv-service

4. Clean it

kubectl delete service dcmrecv-service
kubectl delete  deployments dcmrecv

DICOM API

1. Build docker images for DICOM API

docker build -t dcmrecv_api .
docker tag dcmrecv_api:latest truskovskyi/dcmrecv_api:latest
docker push truskovskyi/dcmrecv_api:latest

2. Deploy

kubectl create -f ./deployment/dicom_api/dicom-api-deployment.yaml
kubectl expose deployment dicom-api --type=LoadBalancer --name=dicom-api-service

3. Get DICOM API

kubectl get service dicom-api-service

4. Clean it

kubectl delete service dicom-api-service
kubectl delete  deployments dicom-api

Manual testing

1. Install dcmtk

brew install dcmtk

2. Send test image

dcmsend --verbose <host> <port> <image>
example: dcmsend --verbose 35.202.37.159 8083 ./data/patient_1/image_49.dcm

Plan

• NFS server
• DICOM storage SCP (receiver)
• DICOM API
• Manual Deploy
• Tests
• CI Deploy