feat(helm): Cogstack CE to provision medcat trainer on startup (#50)

alhendrickson · web-flow · commit f63c22d843bc · 2026-03-05T09:37:51.000Z
diff --git a/deployment/kubernetes/charts/cogstack-helm-ce/Chart.lock b/deployment/kubernetes/charts/cogstack-helm-ce/Chart.lock
@@ -14,8 +14,8 @@ dependencies:
 - name: opensearch-dashboards
   repository: https://opensearch-project.github.io/helm-charts/
   version: 3.5.0
-- name: jupyterhub
+- name: cogstack-jupyterhub
   repository: file://charts/jupyterhub
   version: 0.1.0
-digest: sha256:0dbf6980653337af2efa16bb340d0a04dca9ed517bf4ab3c87044da348e8a84c
-generated: "2026-02-27T16:13:56.822297965Z"
+digest: sha256:d9fb8600b71642cbbdcfc0e9c4b512f3efedadf99544b3e81f97f0deb5a8469a
+generated: "2026-03-03T17:30:34.27108704Z"
diff --git a/deployment/kubernetes/charts/cogstack-helm-ce/README.md b/deployment/kubernetes/charts/cogstack-helm-ce/README.md
@@ -24,19 +24,14 @@ helm install cogstack oci://registry-1.docker.io/cogstacksystems/cogstack-helm-c
 ```
 
 ## Configuration
-
-Key values in `values.yaml`:
+These are some values that are likely to need customization for your deployment:
 
 | Value | Default | Description |
 |-------|---------|-------------|
-| `imagePullSecrets` | `[]` | Secrets for pulling images from a private registry. |
-| `nameOverride` | `""` | Override the chart name used in resource names. |
-| `fullnameOverride` | `""` | Override the full name used in resource names. |
 | `medcat-service.replicasCount` | `1` | Number of MedCAT service replicas. |
 | `anoncat-service.replicasCount` | `1` | Number of AnonCAT service replicas. |
-| `anoncat-service.env.APP_ENABLE_DEMO_UI` | `true` | Enable AnonCAT demo UI. |
-| `anoncat-service.env.DEID_MODE` | `true` | Run AnonCAT in de-identification mode. |
 | `anoncat-service.env.DEID_REDACT` | `false` | Redaction behaviour for de-identification. |
+| `medcat-trainer.env.CSRF_TRUSTED_ORIGINS` | `"http://localhost:8080"` | CSRF trusted origins for MedCAT Trainer frontend (set correct value for your deployment, this default works for port forwarding). |
 
 Subcharts (MedCAT service, AnonCAT service, MedCAT Trainer) support additional options; see their respective charts under `../medcat-service-helm` and `../medcat-trainer-helm`. Pass them under the same keys as in this chart’s `values.yaml` (e.g. `medcat-service.*`, `anoncat-service.*`, `medcat-trainer-helm.*`).
 
diff --git a/deployment/kubernetes/charts/cogstack-helm-ce/templates/NOTES.txt b/deployment/kubernetes/charts/cogstack-helm-ce/templates/NOTES.txt
@@ -14,7 +14,7 @@ kubectl --namespace {{ .Release.Namespace }} port-forward $MEDCAT_POD_NAME 5000:
 export TRAINER_POD_NAME=$(kubectl get pods --namespace {{ .Release.Namespace }} -l "app.kubernetes.io/name=medcat-trainer,app.kubernetes.io/instance={{ .Release.Name }}" -o jsonpath="{.items[0].metadata.name}")
 export CONTAINER_PORT=$(kubectl get pod --namespace {{ .Release.Namespace }} $TRAINER_POD_NAME -o jsonpath="{.spec.containers[?(@.name==\"nginx\")].ports[0].containerPort}")
 export CONTAINER_PORT_API=$(kubectl get pod --namespace {{ .Release.Namespace }} $TRAINER_POD_NAME -o jsonpath="{.spec.containers[?(@.name==\"medcat-trainer\")].ports[0].containerPort}")
-kubectl --namespace {{ .Release.Namespace }} port-forward $TRAINER_POD_NAME 8000:$CONTAINER_PORT 8001:$CONTAINER_PORT_API &
+kubectl --namespace {{ .Release.Namespace }} port-forward $TRAINER_POD_NAME 8080:$CONTAINER_PORT 8001:$CONTAINER_PORT_API &
 
 {{- if .Values.opensearch.enabled }}
 # 4. Opensearch
@@ -38,7 +38,7 @@ kubectl --namespace {{ .Release.Namespace }} port-forward svc/$JUPYTERHUB_PROXY_
 
 echo "Visit http://127.0.0.1:5000 to use MedCAT Service"
 echo "Visit http://127.0.0.1:5001 to use AnonCAT"
-echo "Visit http://127.0.0.1:8000 to use MedCAT Trainer"
+echo "Visit http://127.0.0.1:8080 to use MedCAT Trainer"
 {{ if .Values.opensearch.enabled }}echo "Visit https://127.0.0.1:9200 to use OpenSearch"{{- end }}
 {{ if index .Values "opensearch-dashboards" "enabled" }}echo "Visit http://127.0.0.1:5601 to use OpenSearch Dashboards"{{- end }}
 {{ if index .Values "cogstack-jupyterhub" "enabled" }}echo "Visit http://127.0.0.1:8000 to use jupyterhub"{{- end }}
diff --git a/deployment/kubernetes/charts/cogstack-helm-ce/templates/medcat-trainer-provisioining.yaml b/deployment/kubernetes/charts/cogstack-helm-ce/templates/medcat-trainer-provisioining.yaml
@@ -0,0 +1,29 @@
+{{- if and (index .Values "medcat-trainer" "provisioning" "enabled") (eq (index .Values "medcat-trainer" "provisioning" "existingConfigMap" "name") "cogstack-helm-ce-example-trainer-provisioining") }}
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: {{ index .Values "medcat-trainer" "provisioning" "existingConfigMap" "name" }}
+data:
+  provisioning.yaml: |
+    projects:
+      - dataset:
+          name: "Example Dataset - psych.csv"
+          url: "https://raw.githubusercontent.com/CogStack/cogstack-nlp/be9825f151da2d3b6faac80d49e5be91a1629c68/medcat-trainer/notebook_docs/example_data/psych.csv"
+          description: "Dataset from https://raw.githubusercontent.com/CogStack/cogstack-nlp/be9825f151da2d3b6faac80d49e5be91a1629c68/medcat-trainer/notebook_docs/example_data/psych.csv"
+        project:
+          name: "Example Project"
+          description: "Example project for medcat trainer using MedCAT Service"
+          annotationGuidelineLink: "https://docs.google.com/document/d/1xxelBOYbyVzJ7vLlztP2q1Kw9F5Vr1pRwblgrXPS7QM/edit?usp=sharing"
+          useModelService: true
+          modelServiceUrl: "http://{{ include "cogstack-helm-ce.fullname" . }}-medcat-service:5000"
+      - dataset:
+          name: "Example Dataset - psych.csv - deid"
+          url: "https://raw.githubusercontent.com/CogStack/cogstack-nlp/be9825f151da2d3b6faac80d49e5be91a1629c68/medcat-trainer/notebook_docs/example_data/psych.csv"
+          description: "Dataset from https://raw.githubusercontent.com/CogStack/cogstack-nlp/be9825f151da2d3b6faac80d49e5be91a1629c68/medcat-trainer/notebook_docs/example_data/psych.csv"
+        project:
+          name: "Example Deid project"
+          description: "Example project for medcat trainer using AnonCAT Service"
+          annotationGuidelineLink: "https://docs.google.com/document/d/1xxelBOYbyVzJ7vLlztP2q1Kw9F5Vr1pRwblgrXPS7QM/edit?usp=sharing"
+          useModelService: true
+          modelServiceUrl: "http://{{ include "cogstack-helm-ce.fullname" . }}-anoncat-service:5000"
+{{- end }}
diff --git a/deployment/kubernetes/charts/cogstack-helm-ce/values.yaml b/deployment/kubernetes/charts/cogstack-helm-ce/values.yaml
@@ -24,6 +24,17 @@ medcat-service:
     repository: cogstacksystems/medcat-service
     tag: "1.2.0"
 
+
+medcat-trainer:
+  image:
+    tag: "latest@sha256:dcbd9f7c480dcb88e072189d233a146317f210a826573fd0c47d469da5360be0"
+  provisioning:
+    enabled: true
+    existingConfigMap:
+      name: "cogstack-helm-ce-example-trainer-provisioining"
+  env:
+    CSRF_TRUSTED_ORIGINS: "http://localhost:8080"
+
 opensearch:
   enabled: true
   extraEnvs:
diff --git a/deployment/kubernetes/charts/medcat-trainer-helm/templates/medcat-trainer-configmap.yaml b/deployment/kubernetes/charts/medcat-trainer-helm/templates/medcat-trainer-configmap.yaml
@@ -39,7 +39,3 @@ data:
     stderr_logfile=/dev/stderr
     stderr_logfile_maxbytes=0
     autorestart=true
-{{- if and .Values.provisioning .Values.provisioning.enabled }}
-  provisioning.yaml: |
-{{ toYaml .Values.provisioning.config | indent 4 }}
-{{- end }}
diff --git a/deployment/kubernetes/charts/medcat-trainer-helm/templates/medcat-trainer-deployment.yaml b/deployment/kubernetes/charts/medcat-trainer-helm/templates/medcat-trainer-deployment.yaml
@@ -87,8 +87,8 @@ spec:
               name: api-static
             {{- if and .Values.provisioning .Values.provisioning.enabled }}
             - mountPath: /home/configs/provisioning.yaml
-              name: medcat-trainer-config
-              subPath: provisioning.yaml
+              name: medcat-trainer-provisioning-config
+              subPath: {{ .Values.provisioning.existingConfigMap.subPath | default "provisioning.yaml" }}
             {{- end }}
             {{- if eq .Values.DB_ENGINE "sqlite3" }}  
             - mountPath: /home/api/db
@@ -140,6 +140,15 @@ spec:
         - name: medcat-trainer-config
           configMap:
             name: {{ include "medcat-trainer-helm.fullname" . }}-config
+        {{- if and .Values.provisioning .Values.provisioning.enabled }}
+        - name: medcat-trainer-provisioning-config
+          configMap:
+          {{- if .Values.provisioning.existingConfigMap.name }}
+            name: {{ .Values.provisioning.existingConfigMap.name }}
+          {{- else }}
+            name: {{ include "medcat-trainer-helm.fullname" . }}-provisioning
+          {{- end }}
+        {{- end }}
         - name: api-media
           persistentVolumeClaim:
             claimName: {{ include "medcat-trainer-helm.fullname" . }}-api-media
diff --git a/deployment/kubernetes/charts/medcat-trainer-helm/templates/medcat-trainer-provisioning-configmap.yaml b/deployment/kubernetes/charts/medcat-trainer-helm/templates/medcat-trainer-provisioning-configmap.yaml
@@ -0,0 +1,11 @@
+{{- if (and .Values.provisioning .Values.provisioning.enabled) }}
+{{- if not .Values.provisioning.existingConfigMap }}
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: {{ include "medcat-trainer-helm.fullname" . }}-provisioning
+data:
+  provisioning.yaml: |
+{{ tpl (toYaml .Values.provisioning.config) . | indent 4 }}
+{{- end }}
+{{- end }}
diff --git a/deployment/kubernetes/charts/medcat-trainer-helm/values.yaml b/deployment/kubernetes/charts/medcat-trainer-helm/values.yaml
@@ -42,7 +42,11 @@ env:
 provisioning:
   # -- Set to true to enable provisioning of projects and models on startup..
   enabled: false
-  # -- Put yaml file contents here.See https://docs.cogstack.org/projects/medcat-trainer/en/latest/provisioning/
+  # -- Optional: Reference an existing configmap for the provisioning config.
+  existingConfigMap: {}
+    # -- The name of an existing configmap. Contains a yaml definition under the key "provisioining.yaml"
+    # name: ""
+  # -- Put yaml file contents here. Can be templated See https://docs.cogstack.org/projects/medcat-trainer/en/latest/provisioning/
   config:
     projects:
       - dataset:
diff --git a/deployment/kubernetes/local_dev_startup.sh b/deployment/kubernetes/local_dev_startup.sh
@@ -25,4 +25,9 @@ helm test medcat-service --logs
 
 ## helm install trainer-registry oci://registry-1.docker.io/cogstacksystems/medcat-trainer-helm --wait --timeout 5m0s
 
-helm upgrade cogstack-helm-ce ./cogstack-helm-ce --install --render-subchart-notes --wait --timeout 5m0s
+helm upgrade cogstack-helm-ce ./cogstack-helm-ce --install
+
+# Stop any running port forwards
+# ps -ef | grep '[p]ort-forward' | awk '{print $2}' | xargs -r kill
+# Port forward all services in  cogstack-helm-ce
+# helm get notes cogstack-helm-ce | bash
