06-model-deployment.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "ee01c81b",
   "metadata": {},
   "source": [
    "# 06 - Model Deployment\n",
    "\n",
    "The purpose of this notebook is to execute a CI/CD routine to test and deploy the trained model to `Vertex AI` as an `Endpoint` for online prediction serving. The notebook covers the following steps:\n",
    "1. Run the test steps locally.\n",
    "2. Execute the model deployment `CI/CD` steps using `Cloud Build`.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0da8290c",
   "metadata": {},
   "source": [
    "## Setup"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4873f8cf",
   "metadata": {},
   "source": [
    "### Import libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "59085129",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import logging\n",
    "\n",
    "logging.getLogger().setLevel(logging.INFO)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e37fb189",
   "metadata": {},
   "source": [
    "### Setup Google Cloud project"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e45be804",
   "metadata": {},
   "outputs": [],
   "source": [
    "PROJECT = '[your-project-id]' # Change to your project id.\n",
    "REGION = 'us-central1' # Change to your region.\n",
    "\n",
    "if PROJECT == \"\" or PROJECT is None or PROJECT == \"[your-project-id]\":\n",
    "    # Get your GCP project id from gcloud\n",
    "    shell_output = !gcloud config list --format 'value(core.project)' 2>/dev/null\n",
    "    PROJECT = shell_output[0]\n",
    "\n",
    "print(\"Project ID:\", PROJECT)\n",
    "print(\"Region:\", REGION)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1574964f",
   "metadata": {},
   "source": [
    "### Set configurations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4a01278c",
   "metadata": {},
   "outputs": [],
   "source": [
    "VERSION = 'v01'\n",
    "DATASET_DISPLAY_NAME = 'chicago-taxi-tips'\n",
    "MODEL_DISPLAY_NAME = f'{DATASET_DISPLAY_NAME}-classifier-{VERSION}'\n",
    "ENDPOINT_DISPLAY_NAME = f'{DATASET_DISPLAY_NAME}-classifier'\n",
    "\n",
    "CICD_IMAGE_NAME = 'cicd:latest'\n",
    "CICD_IMAGE_URI = f\"gcr.io/{PROJECT}/{CICD_IMAGE_NAME}\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "87f6f1e0",
   "metadata": {},
   "source": [
    "## 1. Run CI/CD steps locally"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a223cdf6",
   "metadata": {},
   "outputs": [],
   "source": [
    "os.environ['PROJECT'] = PROJECT\n",
    "os.environ['REGION'] = REGION\n",
    "os.environ['MODEL_DISPLAY_NAME'] = MODEL_DISPLAY_NAME\n",
    "os.environ['ENDPOINT_DISPLAY_NAME'] = ENDPOINT_DISPLAY_NAME"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b6546ac1",
   "metadata": {},
   "source": [
    "### Run the model artifact testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "74c0f8a8",
   "metadata": {},
   "outputs": [],
   "source": [
    "!py.test src/tests/model_deployment_tests.py::test_model_artifact -s"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "77885b24",
   "metadata": {},
   "source": [
    "### Run create endpoint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0efe73b5",
   "metadata": {},
   "outputs": [],
   "source": [
    "!python build/utils.py \\\n",
    "    --mode=create-endpoint\\\n",
    "    --project={PROJECT}\\\n",
    "    --region={REGION}\\\n",
    "    --endpoint-display-name={ENDPOINT_DISPLAY_NAME}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3eb28c6f",
   "metadata": {},
   "source": [
    "### Run deploy model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9cb3f19d",
   "metadata": {},
   "outputs": [],
   "source": [
    "!python build/utils.py \\\n",
    "    --mode=deploy-model\\\n",
    "    --project={PROJECT}\\\n",
    "    --region={REGION}\\\n",
    "    --endpoint-display-name={ENDPOINT_DISPLAY_NAME}\\\n",
    "    --model-display-name={MODEL_DISPLAY_NAME}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee492355",
   "metadata": {},
   "source": [
    "### Test deployed model endpoint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3d4bce50",
   "metadata": {},
   "outputs": [],
   "source": [
    "!py.test src/tests/model_deployment_tests.py::test_model_endpoint"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "37b150c9",
   "metadata": {},
   "source": [
    "## 2. Execute the Model Deployment CI/CD routine in Cloud Build\n",
    "\n",
    "The CI/CD routine is defined in the [model-deployment.yaml](model-deployment.yaml) file, and consists of the following steps:\n",
    "1. Load and test the the trained model interface.\n",
    "2. Create and endpoint in Vertex AI if it doesn't exists.\n",
    "3. Deploy the model to the endpoint.\n",
    "4. Test the endpoint."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "839e540c",
   "metadata": {},
   "source": [
    "### Build CI/CD container Image for Cloud Build\n",
    "\n",
    "This is the runtime environment where the steps of testing and deploying model will be executed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a7f9bf4e",
   "metadata": {},
   "outputs": [],
   "source": [
    "!echo $CICD_IMAGE_URI"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3855daae",
   "metadata": {},
   "outputs": [],
   "source": [
    "!gcloud builds submit --tag $CICD_IMAGE_URI build/. --timeout=15m"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "90fbd4b9",
   "metadata": {},
   "source": [
    "### Run CI/CD from model deployment using Cloud Build"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e1aec70c",
   "metadata": {},
   "outputs": [],
   "source": [
    "REPO_URL = \"https://github.com/GoogleCloudPlatform/mlops-with-vertex-ai.git\" # Change to your github repo.\n",
    "BRANCH = \"main\" "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "01995fa5",
   "metadata": {},
   "outputs": [],
   "source": [
    "SUBSTITUTIONS=f\"\"\"\\\n",
    "_REPO_URL='{REPO_URL}',\\\n",
    "_BRANCH={BRANCH},\\\n",
    "_CICD_IMAGE_URI={CICD_IMAGE_URI},\\\n",
    "_PROJECT={PROJECT},\\\n",
    "_REGION={REGION},\\\n",
    "_MODEL_DISPLAY_NAME={MODEL_DISPLAY_NAME},\\\n",
    "_ENDPOINT_DISPLAY_NAME={ENDPOINT_DISPLAY_NAME},\\\n",
    "\"\"\"\n",
    "\n",
    "!echo $SUBSTITUTIONS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8849d3e4",
   "metadata": {},
   "outputs": [],
   "source": [
    "!gcloud builds submit --no-source --config build/model-deployment.yaml --substitutions {SUBSTITUTIONS} --timeout=30m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "01831724",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4418b01e",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "environment": {
   "name": "common-cpu.m79",
   "type": "gcloud",
   "uri": "gcr.io/deeplearning-platform-release/base-cpu:m79"
  },
  "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": "3.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}