Add image pull secret sample.

samples/notebooks/imagepullsecret.ipynb

@@ -0,0 +1,176 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Sample of using image pull secrets\n",
+    "ImagePullSecrets is a way provided by kubernetes to access hidden/protected images. It is a list of references to secrets. Please refer to [here](https://kubernetes.io/docs/concepts/containers/images/#specifying-imagepullsecrets-on-a-pod) and [Argo reference](https://github.com/argoproj/argo/blob/47eba519107c229edf61dbe024a6a5e0f1618a8d/pkg/apis/workflow/v1alpha1/types.go#L115) for more details.\n",
+    "\n",
+    "In KFP, we can simply set the ImagePullSecrets for each pipeline to grant it the ability to access the images in other registries when needed. Please see the example below for a simple demonstration."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!pip3 install kfp --upgrade"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import kfp\n",
+    "import kfp.dsl as dsl\n",
+    "from kfp import compiler\n",
+    "from kubernetes import client as k8s_client"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "class GetFrequentWordOp(dsl.ContainerOp):\n",
+    "  \"\"\"A get frequent word class representing a component in ML Pipelines.\n",
+    "  \"\"\"\n",
+    "  def __init__(self, name, message):\n",
+    "    \"\"\"Args:\n",
+    "         name: An identifier of the step which needs to be unique within a pipeline.\n",
+    "         message: a dsl.PipelineParam object representing an input message.\n",
+    "    \"\"\"\n",
+    "    super(GetFrequentWordOp, self).__init__(\n",
+    "        name=name,\n",
+    "        image='python:3.5-jessie',\n",
+    "        command=['sh', '-c'],\n",
+    "        arguments=['python -c \"from collections import Counter; '\n",
+    "                   'words = Counter(\\'%s\\'.split()); print(max(words, key=words.get))\" '\n",
+    "                   '| tee /tmp/message.txt' % message],\n",
+    "        file_outputs={'word': '/tmp/message.txt'})"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "In this example, we defined a KFP component by directly specifying the base image and the python command it runs. Note that we're using a public image so actually there is no need to use imagepullsecrets. However, if the image lives within a private/protected container registry, an imagepullsecret might be necessary to authenticate the pipeline in order to pull the image."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "@dsl.pipeline(\n",
+    "  name='Save Most Frequent',\n",
+    "  description='Get Most Frequent Word and Save to GCS'\n",
+    ")\n",
+    "def save_most_frequent_word(message: str):\n",
+    "  \"\"\"A pipeline function describing the orchestration of the workflow.\"\"\"\n",
+    "\n",
+    "  counter = GetFrequentWordOp(\n",
+    "          name='get-Frequent',\n",
+    "          message=message)\n",
+    "  dsl.get_pipeline_conf().set_image_pull_secrets([k8s_client.V1ObjectReference(name=\"secretA\")])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Compile the pipeline.\n",
+    "compiler.Compiler().compile(save_most_frequent_word,  'save_most_frequent.yaml')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "The above code should generate following `save_most_frequent.yaml` pipeline definition.\n",
+    "```yaml\n",
+    "apiVersion: argoproj.io/v1alpha1\n",
+    "kind: Workflow\n",
+    "metadata:\n",
+    "  generateName: save-most-frequent-\n",
+    "spec:\n",
+    "  arguments:\n",
+    "    parameters:\n",
+    "    - name: message\n",
+    "  entrypoint: save-most-frequent\n",
+    "  imagePullSecrets:\n",
+    "  - name: secretA\n",
+    "  serviceAccountName: pipeline-runner\n",
+    "  templates:\n",
+    "  - container:\n",
+    "      args:\n",
+    "      - python -c \"from collections import Counter; words = Counter('{{inputs.parameters.message}}'.split());\n",
+    "        print(max(words, key=words.get))\" | tee /tmp/message.txt\n",
+    "      command:\n",
+    "      - sh\n",
+    "      - -c\n",
+    "      image: python:3.5-jessie\n",
+    "    inputs:\n",
+    "      parameters:\n",
+    "      - name: message\n",
+    "    name: get-frequent\n",
+    "    outputs:\n",
+    "      artifacts:\n",
+    "      - name: mlpipeline-ui-metadata\n",
+    "        optional: true\n",
+    "        path: /mlpipeline-ui-metadata.json\n",
+    "      - name: mlpipeline-metrics\n",
+    "        optional: true\n",
+    "        path: /mlpipeline-metrics.json\n",
+    "      parameters:\n",
+    "      - name: get-frequent-word\n",
+    "        valueFrom:\n",
+    "          path: /tmp/message.txt\n",
+    "  - dag:\n",
+    "      tasks:\n",
+    "      - arguments:\n",
+    "          parameters:\n",
+    "          - name: message\n",
+    "            value: '{{inputs.parameters.message}}'\n",
+    "        name: get-frequent\n",
+    "        template: get-frequent\n",
+    "    inputs:\n",
+    "      parameters:\n",
+    "      - name: message\n",
+    "    name: save-most-frequent\n",
+    "```\n",
+    "\n",
+    "When looking into the `save_most_frequent.yaml` generated, one can find that the generated workflow is associated with an imagePullSecrets specified via its name."
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.6.7"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}