Inference server

manifests/helm/app/templates/deployment.yaml

@@ -34,6 +34,14 @@ spec:
     spec:
       containers:
       - env:
+        - name: USE_SERVING_RUNTIME
+          value: "{{ .Values.useServingRuntime }}"
+        - name: ENVIRONMENT
+{{- if .Values.isProd }}
+          value: prod
+{{- else }}
+          value: dev
+{{- end }}
         - name: MLFLOW_TRACKING_URI
           value: http://mlflow.mlflow.svc.cluster.local:5000
         - name: MLFLOW_S3_ENDPOINT_URL

manifests/helm/app/templates/rolebinding-model-controller.yaml

@@ -0,0 +1,38 @@
+kind: Role
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  name: {{ .Release.Name }}-model-controller
+  namespace: {{ .Release.Namespace }}
+rules:
+  - verbs:
+      - create
+      - delete
+      - get
+      - list
+      - patch
+      - update
+    apiGroups:
+      - serving.kserve.io
+    resources:
+      - inferenceservices
+  - verbs:
+      - create
+      - delete
+    apiGroups:
+      - batch
+    resources:
+      - jobs
+---
+kind: RoleBinding
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  name: {{ .Release.Name }}-model-controller
+  namespace: {{ .Release.Namespace }}
+subjects:
+  - kind: ServiceAccount
+    name: model-controller
+    namespace: {{ .Release.Namespace }}
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: Role
+  name: {{ .Release.Name }}-model-controller

manifests/helm/app/templates/serving-runtime.yaml

@@ -0,0 +1,66 @@
+{{- if .Values.useServingRuntime }}
+apiVersion: serving.kserve.io/v1alpha1
+kind: ServingRuntime
+metadata:
+  annotations:
+    opendatahub.io/apiProtocol: REST
+    opendatahub.io/recommended-accelerators: '["nvidia.com/gpu"]'
+    opendatahub.io/template-display-name: OpenVINO Model Server
+    opendatahub.io/template-name: ovms
+    openshift.io/display-name: {{ .Release.Name }}
+  name: {{ .Release.Name }}
+  namespace: {{ .Release.Namespace }}
+  labels:
+    opendatahub.io/dashboard: 'true'
+spec:
+  supportedModelFormats:
+    - autoSelect: true
+      name: openvino_ir
+      version: opset1
+    - autoSelect: true
+      name: onnx
+      version: '1'
+    - autoSelect: true
+      name: tensorflow
+      version: '2'
+  builtInAdapter:
+    env:
+      - name: OVMS_FORCE_TARGET_DEVICE
+        value: AUTO
+    memBufferBytes: 134217728
+    modelLoadingTimeoutMillis: 90000
+    runtimeManagementPort: 8888
+    serverType: ovms
+  multiModel: true
+  containers:
+    - args:
+        - '--port=8001'
+        - '--rest_port=8888'
+        - '--config_path=/models/model_config_list.json'
+        - '--file_system_poll_wait_seconds=0'
+        - '--grpc_bind_address=0.0.0.0'
+        - '--rest_bind_address=0.0.0.0'
+      image: 'quay.io/modh/openvino_model_server@sha256:f1140e9d987580d1aab1ccc62519b48b1d2673308b2db496e9e505e3be788d9f'
+      name: ovms
+      resources:
+        limits:
+          cpu: '2'
+          memory: 8Gi
+        requests:
+          cpu: '1'
+          memory: 4Gi
+      volumeMounts:
+        - mountPath: /dev/shm
+          name: shm
+  protocolVersions:
+    - grpc-v1
+  grpcEndpoint: 'port:8085'
+  volumes:
+    - emptyDir:
+        medium: Memory
+        sizeLimit: 2Gi
+      name: shm
+  replicas: {{ .Values.replicaCount }}
+  tolerations: []
+  grpcDataEndpoint: 'port:8001'
+{{- end }}

manifests/helm/app/templates/workbench.yaml

@@ -87,6 +87,14 @@ spec:
             secretKeyRef:
               name: aws-connection-my-storage
               key: AWS_S3_ENDPOINT
+        - name: USE_SERVING_RUNTIME
+          value: "{{ .Values.useServingRuntime }}"
+        - name: ENVIRONMENT
+{{- if .Values.isProd }}
+          value: prod
+{{- else }}
+          value: dev
+{{- end }}
         - name: PIP_CERT
           value: /etc/pki/tls/custom-certs/ca-bundle.crt
         - name: REQUESTS_CA_BUNDLE
@@ -158,7 +166,7 @@ spec:
           valueFrom:
             fieldRef:
               fieldPath: metadata.namespace
-        image: registry.redhat.io/openshift4/ose-oauth-proxy@sha256:4bef31eb993feb6f1096b51b4876c65a6fb1f4401fee97fa4f4542b6b7c9bc46
+        image: registry.redhat.io/openshift4/ose-oauth-proxy@sha256:4f8d66597feeb32bb18699326029f9a71a5aca4a57679d636b876377c2e95695
         #image: registry.redhat.io/openshift4/ose-oauth-proxy:v4.14
         imagePullPolicy: Always
         livenessProbe:

manifests/helm/app/values.yaml

@@ -14,6 +14,7 @@ image:
   # Overrides the image tag whose default is the chart appVersion.
 
 isProd: True
+useServingRuntime: ${{ values.use_serving_runtime }}
 
 route:
   host: ""

manifests/helm/build/templates/pipeline-training.yaml

@@ -11,6 +11,8 @@ spec:
     - image: {{ include "image.dev-url" . }}-training:latest
       name: train
       env:
+      - name: USE_SERVING_RUNTIME
+        value: "{{ .Values.useServingRuntime }}"
       - name: MLFLOW_TRACKING_URI
         value: 'http://mlflow.mlflow.svc.cluster.local:5000'
       - name: MLFLOW_S3_ENDPOINT_URL
@@ -36,6 +38,10 @@ spec:
     - image: {{ include "image.dev-url" . }}-evaluation:latest
       name: evaluate
       env:
+      - name: USE_SERVING_RUNTIME
+        value: "{{ .Values.useServingRuntime }}"
+      - name: ENVIRONMENT
+        value: dev
       - name: MLFLOW_TRACKING_URI
         value: 'http://mlflow.mlflow.svc.cluster.local:5000'
       - name: MLFLOW_S3_ENDPOINT_URL

manifests/helm/build/templates/serviceaccount-model-controller.yaml

@@ -0,0 +1,4 @@
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  name: model-controller

manifests/helm/build/templates/triggertemplate-training.yaml

@@ -11,5 +11,8 @@ spec:
         labels:
           tekton.dev/pipeline: cron-job
       spec:
+        taskRunSpecs:
+        - pipelineTaskName: evaluate
+          taskServiceAccountName: model-controller
         pipelineRef:
           name: train-and-evaluate

manifests/helm/build/values.yaml

@@ -3,6 +3,8 @@ app:
   namespace: ${{values.component_id}}-dev
   cluster: ${{ values.cluster }}
 
+useServingRuntime: ${{ values.use_serving_runtime }}
+
 git:
   branch: main
   repo: ${{ values.source_repository }}

skeleton/.che/che-editor.yaml

@@ -0,0 +1,99 @@
+inline:
+  schemaVersion: 2.1.0
+  metadata:
+    name: {{ cookiecutter.project_name }}
+  commands:
+    - id: on-init-command
+      apply:
+        component: che-pycharm-injector
+  events:
+    preStart:
+      - on-init-command
+  attributes:
+    projects:
+      - name: {{ cookiecutter.project_name }}
+        git:
+          remotes:
+            origin: "{{ cookiecutter.git_repo }}"
+          checkoutFrom:
+            revision: main
+  components:
+    - name: che-pycharm-runtime-description
+      container:
+        image: 'quay.io/devfile/universal-developer-image:ubi8-9436df2'
+        command:
+          - /projector/entrypoint-volume.sh
+        env:
+          - name: PROJECTOR_ASSEMBLY_DIR
+            value: /projector
+          - name: PROJECTOR_CONFIG_DIR
+            value: /home/user/.jetbrains
+        volumeMounts:
+          - name: projector-volume
+            path: /projector
+          - name: projector-configuration
+            path: /home/user/.jetbrains
+          - name: projector-java-configuration
+            path: /home/user/.java
+        memoryLimit: 6144Mi
+        memoryRequest: 2048Mi
+        cpuLimit: 2000m
+        cpuRequest: 1500m
+        endpoints:
+          - name: intellij
+            attributes:
+              type: main
+              cookiesAuthEnabled: true
+              discoverable: false
+              urlRewriteSupported: true
+            targetPort: 8887
+            exposure: public
+            path: '/?backgroundColor=434343&wss'
+            secure: false
+            protocol: http
+          - name: intellij-redirect-1
+            targetPort: 13131
+            exposure: public
+            protocol: http
+            attributes:
+              discoverable: false
+              urlRewriteSupported: true
+          - name: intellij-redirect-2
+            targetPort: 13132
+            exposure: public
+            protocol: http
+            attributes:
+              discoverable: false
+              urlRewriteSupported: true
+          - name: intellij-redirect-3
+            targetPort: 13133
+            exposure: public
+            protocol: http
+            attributes:
+              discoverable: false
+              urlRewriteSupported: true
+      attributes:
+        app.kubernetes.io/component: che-pycharm-injector
+        app.kubernetes.io/part-of: che-pycharm.eclipse.org
+    - name: projector-volume
+      volume: { }
+    - name: projector-configuration
+      volume: { }
+    - name: projector-java-configuration
+      volume: { }
+    - name: che-pycharm-injector
+      container:
+        image: 'quay.io/che-incubator/che-pycharm:2022.1-next'
+        command: ["/projector/entrypoint-init-container.sh"]
+        env:
+          - name: PROJECTOR_VOLUME_MOUNT
+            value: /projector-volume
+          - name: PROJECTOR_ASSEMBLY_DIR
+            value: /projector
+        volumeMounts:
+          - name: projector-volume
+            path: /projector-volume
+        memoryLimit: 2G
+        cpuLimit: 500m
+      mountSources: true
+      sourceMapping: /projects

skeleton/.gitignore

@@ -163,5 +163,3 @@ cython_debug/
 #  and can be added to the global gitignore or merged into this file.  For a more nuclear
 #  option (not recommended) you can uncomment the following to ignore the entire idea folder.
 .idea/
-
-.vscode/

skeleton/.vscode/extensions.json

@@ -0,0 +1,7 @@
+{
+    // See https://go.microsoft.com/fwlink/?LinkId=827846
+    // for the documentation about the extensions.json format
+    "recommendations": [
+      "ms-python.python"
+    ]
+}

skeleton/catalog-info.yaml

@@ -17,10 +17,10 @@ metadata:
     - python
     - ai
   links:
-    - url: https://devspaces{{cookiecutter.cluster}}/#https://{{cookiecutter.host}}/{{cookiecutter.destination}}
+    - url: https://devspaces{{cookiecutter.cluster}}/#https://{{cookiecutter.host}}/{{cookiecutter.destination}}?che-editor=che-incubator/che-code/latest
       title: OpenShift Dev Spaces (VS Code)
       icon: web
-    - url: https://devspaces{{cookiecutter.cluster}}/#https://{{cookiecutter.host}}/{{cookiecutter.destination}}?che-editor=https://raw.githubusercontent.com/che-incubator/jetbrains-editor-images/main/devfiles/next/che-pycharm/2022.1-next.yaml
+    - url: https://devspaces{{cookiecutter.cluster}}/#https://{{cookiecutter.host}}/{{cookiecutter.destination}}
       title: OpenShift Dev Spaces (JetBrains PyCharm)
       icon: web
 spec:
@@ -42,7 +42,7 @@ spec:
   lifecycle: production
   owner: {{cookiecutter.owner}}
   definition:
-    $text: https://{{cookiecutter.component_id}}-prod-app-{{cookiecutter.component_id}}-prod/{{cookiecutter.cluster}}/openapi.json
+    $text: https://{{cookiecutter.component_id}}-prod-app-{{cookiecutter.component_id}}-prod{{cookiecutter.cluster}}/openapi.json
 ---
 apiVersion: backstage.io/v1alpha1
 kind: API

skeleton/common/__init__.py

@@ -1,2 +1,25 @@
+from os import getenv
+
+def _strtobool (val):
+    """
+    Convert a string representation of truth to true (1) or false (0).
+    True values are 'y', 'yes', 't', 'true', 'on', and '1'; false values
+    are 'n', 'no', 'f', 'false', 'off', and '0'.  Raises ValueError if
+    'val' is anything else.
+
+    This is taken from distutils, which is deprecated in Python 3.12
+    """
+    val = val.lower()
+    if val in ('y', 'yes', 't', 'true', 'on', '1'):
+        return 1
+    elif val in ('n', 'no', 'f', 'false', 'off', '0'):
+        return 0
+    else:
+        raise ValueError("invalid truth value %r" % (val,))
+
+
 MODEL_NAME = "{{ cookiecutter.project_name }}"
-MODEL_VERSION = "0.0.1"
+MODEL_VERSION = "0.0.1"
+USE_SERVING_RUNTIME = _strtobool(getenv("USE_SERVING_RUNTIME") or "False")
+
+CACHE_TTL = 600