Merge pull request #87 from Azure-Samples/template-lab-improvs

Improvements in the backend pool load balancing lab to segregate apim…

Author: vieiraae

.github/workflows/nbchecks.py

@@ -0,0 +1,23 @@
+# to run locally, use the following command in the root folder of the repository:python .github/workflows/nbchecks.py
+
+import json, sys, os
+from pathlib import Path
+
+def has_outputs_stored(file):
+    with open(file, 'r', encoding='utf-8') as f:
+        notebook_content = json.load(f)
+    
+    for cell in notebook_content.get('cells', []):
+        if cell.get('cell_type') == 'code' and cell.get('outputs'):
+            print(f"The notebook {filename} has outputs stored.", file=sys.stderr)
+            return True
+    return False
+
+exit_code = 0
+for file in Path(".").glob('**/*.ipynb'):
+    filename = os.fsdecode(file)
+    if has_outputs_stored(file):        
+        exit_code = 1
+if exit_code == 0:
+    print("All good. No stored output found.")
+sys.exit(exit_code)

.github/workflows/nbchecks.yaml

@@ -0,0 +1,28 @@
+name: notebook-checks
+
+on:
+  pull_request:
+    paths:
+      - "**/*.ipynb"
+    branches: [main]
+  push:
+    paths:
+      - "**/*.ipynb"
+    branches: [main]
+  
+permissions:
+  contents: read
+  pull-requests: write
+
+jobs:
+  nbchecks:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Set up Python 3.12
+        uses: actions/setup-python@v1
+        with:
+          python-version: "3.12"
+      - uses: actions/checkout@v1
+      - name: nbchecks
+        run: |
+          python ./.github/workflows/nbchecks.py

README.md

@@ -6,7 +6,7 @@
 
 ## What's new ✨
 
-<!-- ➕ the [**AI Foundry SDK**](labs/ai-foundry-sdk/ai-foundry-sdk.ipynb) lab. -->
+➕ the [**AI Foundry SDK**](labs/ai-foundry-sdk/ai-foundry-sdk.ipynb) lab.  
 ➕ the [**Content filtering**](labs/content-filtering/content-filtering.ipynb) and [**Prompt shielding**](labs/content-filtering/prompt-shielding.ipynb) labs.  
 ➕ the [**Model routing**](labs/model-routing/model-routing.ipynb) lab with OpenAI model based routing.  
 ➕ the [**Prompt flow**](labs/prompt-flow/prompt-flow.ipynb) lab to try the [Azure AI Studio Prompt Flow](https://learn.microsoft.com/azure/ai-studio/how-to/prompt-flow) with Azure API Management.  

labs/backend-pool-load-balancing/backend-pool-load-balancing.ipynb

@@ -21,7 +21,7 @@
     "\n",
     "### TOC\n",
     "- [0️⃣ Initialize notebook variables](#0)\n",
-    "- [1️⃣ Create the Azure Resource Group](#1)\n",
+    "- [1️⃣ Verify the Azure CLI and the connected Azure subscription](#1)\n",
     "- [2️⃣ Create deployment using 🦾 Bicep](#2)\n",
     "- [3️⃣ Get the deployment outputs](#3)\n",
     "- [🧪 Test the API using a direct HTTP call](#requests)\n",
@@ -53,35 +53,43 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
-    "import os\n",
+    "import os, sys, json\n",
+    "sys.path.insert(1, '../../shared')  # add the shared directory to the Python path\n",
+    "import utils\n",
     "\n",
     "deployment_name = os.path.basename(os.path.dirname(globals()['__vsc_ipynb_file__']))\n",
     "resource_group_name = f\"lab-{deployment_name}\" # change the name to match your naming style\n",
     "resource_group_location = \"westeurope\"\n",
     "\n",
+    "apim_sku = 'Basicv2'\n",
+    "\n",
     "openai_resources = [\n",
     "    {\"name\": \"openai1\", \"location\": \"uksouth\", \"priority\": 1, \"weight\": 80},\n",
     "    {\"name\": \"openai2\", \"location\": \"swedencentral\", \"priority\": 1, \"weight\": 10},\n",
     "    {\"name\": \"openai3\", \"location\": \"francecentral\", \"priority\": 1, \"weight\": 10}\n",
     "]\n",
     "\n",
+    "openai_deployment_name = \"gpt-35-turbo\"\n",
     "openai_model_name = \"gpt-35-turbo\"\n",
     "openai_model_version = \"0613\"\n",
-    "openai_deployment_name = \"gpt-35-turbo\"\n",
-    "openai_api_version = \"2024-02-01\"\n"
+    "openai_model_capacity = 8\n",
+    "openai_api_version = \"2024-02-01\"\n",
+    "\n",
+    "utils.print_ok('Notebook initiaized')"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
     "<a id='1'></a>\n",
-    "### 1️⃣ Create the Azure Resource Group\n",
-    "All resources deployed in this lab will be created in the specified resource group. Skip this step if you want to use an existing resource group."
+    "### 1️⃣ Verify the Azure CLI and the connected Azure subscription\n",
+    "\n",
+    "The following commands ensure that you have the latest version of the Azure CLI and that the Azure CLI is connected to your Azure subscription."
    ]
   },
   {
@@ -90,18 +98,11 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# %load ../../shared/snippets/create-az-resource-group.py\n",
-    "\n",
-    "# type: ignore\n",
-    "\n",
-    "import datetime\n",
-    "\n",
-    "resource_group_stdout = ! az group create --name {resource_group_name} --location {resource_group_location}\n",
-    "\n",
-    "if resource_group_stdout.n.startswith(\"ERROR\"):\n",
-    "    print(resource_group_stdout)\n",
-    "else:\n",
-    "    print(f\"✅ Azure Resource Group {resource_group_name} created ⌚ {datetime.datetime.now().time()}\")\n"
+    "output = utils.run(\"az account show\", \"Retrieved az account\", \"Failed to get the current az account\")\n",
+    "if output.success and output.json_data:\n",
+    "    current_user = output.json_data['user']['name']\n",
+    "    subscription_id = output.json_data['id']\n",
+    "    tenant_id = output.json_data['tenantId']"
    ]
   },
   {
@@ -111,9 +112,9 @@
     "<a id='2'></a>\n",
     "### 2️⃣ Create deployment using 🦾 Bicep\n",
     "\n",
-    "This lab uses [Bicep](https://learn.microsoft.com/azure/azure-resource-manager/bicep/overview?tabs=bicep) to declarative define all the resources that will be deployed. Change the parameters or the [main.bicep](main.bicep) directly to try different configurations. \n",
+    "This lab uses [Bicep](https://learn.microsoft.com/azure/azure-resource-manager/bicep/overview?tabs=bicep) to declarative define all the resources that will be deployed in the specified resource group. Change the parameters or the [main.bicep](main.bicep) directly to try different configurations. \n",
     "\n",
-    "`openAIModelCapacity` is set intentionally low to `8` (8k tokens per minute) in _main.bicep_ to showcase the retry logic in the load balancer."
+    "`openAIModelCapacity` is set intentionally low to `8` (8k tokens per minute) to showcase the retry logic in the load balancer."
    ]
   },
   {
@@ -122,45 +123,43 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# %load ../../shared/snippets/create-az-deployment.py\n",
-    "\n",
-    "# type: ignore\n",
-    "\n",
-    "import json\n",
-    "\n",
-    "backend_id = \"openai-backend-pool\" if len(openai_resources) > 1 else openai_resources[0].get(\"name\")\n",
+    "# create the resource group if doesn't exist\n",
+    "utils.create_resource_group(True, resource_group_name, resource_group_location)\n",
     "\n",
+    "# update the APIM policy file before the deployment\n",
+    "policy_xml = None\n",
     "with open(\"policy.xml\", 'r') as policy_xml_file:\n",
-    "    policy_xml = policy_xml_file.read()\n",
-    "\n",
-    "    if \"{backend-id}\" in policy_xml:\n",
-    "        policy_xml = policy_xml.replace(\"{backend-id}\", backend_id)\n",
-    "\n",
-    "    if \"{aad-client-application-id}\" in policy_xml:\n",
-    "        policy_xml = policy_xml.replace(\"{aad-client-application-id}\", client_id)\n",
-    "\n",
-    "    if \"{aad-tenant-id}\" in policy_xml:\n",
-    "        policy_xml = policy_xml.replace(\"{aad-tenant-id}\", tenant_id)\n",
-    "\n",
+    "    policy_template_xml = policy_xml_file.read()\n",
+    "    if \"{backend-id}\" in policy_template_xml:\n",
+    "        policy_xml = policy_template_xml.replace(\"{backend-id}\", str(\"openai-backend-pool\" if len(openai_resources) > 1 else openai_resources[0].get(\"name\")))                \n",
     "    policy_xml_file.close()\n",
-    "open(\"policy-updated.xml\", 'w').write(policy_xml)\n",
+    "if policy_xml is not None:\n",
+    "    open(\"policy.xml\", 'w').write(policy_xml)\n",
     "\n",
+    "# define the BICEP parameters\n",
     "bicep_parameters = {\n",
     "    \"$schema\": \"https://schema.management.azure.com/schemas/2019-04-01/deploymentParameters.json#\",\n",
     "    \"contentVersion\": \"1.0.0.0\",\n",
     "    \"parameters\": {\n",
+    "        \"apimSku\": { \"value\": apim_sku },\n",
     "        \"openAIConfig\": { \"value\": openai_resources },\n",
     "        \"openAIDeploymentName\": { \"value\": openai_deployment_name },\n",
     "        \"openAIModelName\": { \"value\": openai_model_name },\n",
     "        \"openAIModelVersion\": { \"value\": openai_model_version },\n",
+    "        \"openAIModelCapacity\": { \"value\": openai_model_capacity },\n",
     "        \"openAIAPIVersion\": { \"value\": openai_api_version }\n",
     "    }\n",
     "}\n",
     "\n",
+    "# write the parameters to a file \n",
     "with open('params.json', 'w') as bicep_parameters_file:\n",
     "    bicep_parameters_file.write(json.dumps(bicep_parameters))\n",
     "\n",
-    "! az deployment group create --name {deployment_name} --resource-group {resource_group_name} --template-file \"main.bicep\" --parameters \"params.json\"\n"
+    "# run the deployment\n",
+    "output = utils.run(f\"az deployment group create --name {deployment_name} --resource-group {resource_group_name} --template-file main.bicep --parameters params.json\", \n",
+    "                f\"Deployment '{deployment_name}' succeeded\", f\"Deployment '{deployment_name}' failed\")\n",
+    "open(\"policy.xml\", 'w').write(policy_template_xml)\n",
+    "\n"
    ]
   },
   {
@@ -170,64 +169,22 @@
     "<a id='3'></a>\n",
     "### 3️⃣ Get the deployment outputs\n",
     "\n",
-    "We are now at the stage where we only need to retrieve the gateway URL and the subscription before we are ready for testing."
+    "Retrieve the required outputs from the Bicep deployment."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
-    "# %load ../../shared/snippets/deployment-outputs.py\n",
-    "# type: ignore\n",
-    "\n",
     "# Obtain all of the outputs from the deployment\n",
-    "stdout = ! az deployment group show --name {deployment_name} -g {resource_group_name} --query properties.outputs -o json\n",
-    "outputs = json.loads(stdout.n)\n",
-    "\n",
-    "# Extract the individual properties\n",
-    "apim_service_id = outputs.get('apimServiceId', {}).get('value', '')\n",
-    "apim_subscription_key = outputs.get('apimSubscriptionKey', {}).get('value', '')\n",
-    "apim_subscription1_key = outputs.get('apimSubscription1Key', {}).get('value', '')\n",
-    "apim_subscription2_key = outputs.get('apimSubscription2Key', {}).get('value', '')\n",
-    "apim_subscription3_key = outputs.get('apimSubscription3Key', {}).get('value', '')\n",
-    "apim_resource_gateway_url = outputs.get('apimResourceGatewayURL', {}).get('value', '')\n",
-    "workspace_id = outputs.get('logAnalyticsWorkspaceId', {}).get('value', '')\n",
-    "app_id = outputs.get('applicationInsightsAppId', {}).get('value', '')\n",
-    "function_app_resource_name = outputs.get('functionAppResourceName', {}).get('value', '')\n",
-    "cosmosdb_connection_string = outputs.get('cosmosDBConnectionString', {}).get('value', '')\n",
-    "\n",
-    "# Print the extracted properties if they are not empty\n",
-    "if apim_service_id:\n",
-    "    print(f\"👉🏻 APIM Service Id: {apim_service_id}\")\n",
-    "\n",
-    "if apim_subscription_key:\n",
-    "    print(f\"👉🏻 APIM Subscription Key (masked): ****{apim_subscription_key[-4:]}\")\n",
-    "\n",
-    "if apim_subscription1_key:\n",
-    "    print(f\"👉🏻 APIM Subscription Key 1 (masked): ****{apim_subscription1_key[-4:]}\")\n",
-    "\n",
-    "if apim_subscription2_key:\n",
-    "    print(f\"👉🏻 APIM Subscription Key 2 (masked): ****{apim_subscription2_key[-4:]}\")\n",
-    "\n",
-    "if apim_subscription3_key:\n",
-    "    print(f\"👉🏻 APIM Subscription Key 3 (masked): ****{apim_subscription3_key[-4:]}\")\n",
-    "\n",
-    "if apim_resource_gateway_url:\n",
-    "    print(f\"👉🏻 APIM API Gateway URL: {apim_resource_gateway_url}\")\n",
-    "\n",
-    "if workspace_id:\n",
-    "    print(f\"👉🏻 Workspace ID: {workspace_id}\")\n",
-    "\n",
-    "if app_id:\n",
-    "    print(f\"👉🏻 App ID: {app_id}\")\n",
-    "\n",
-    "if function_app_resource_name:\n",
-    "    print(f\"👉🏻 Function Name: {function_app_resource_name}\")\n",
-    "\n",
-    "if cosmosdb_connection_string:\n",
-    "    print(f\"👉🏻 Cosmos DB Connection String: {cosmosdb_connection_string}\")\n"
+    "output = utils.run(f\"az deployment group show --name {deployment_name} -g {resource_group_name}\", f\"Retrieved deployment: {deployment_name}\", f\"Failed to retrieve deployment: {deployment_name}\")\n",
+    "if output.success and output.json_data:\n",
+    "    apim_service_id = utils.get_deployment_output(output, 'apimServiceId', 'APIM Service Id')\n",
+    "    apim_subscription_key = utils.get_deployment_output(output, 'apimSubscriptionKey', 'APIM Subscription Key (masked)', True)\n",
+    "    apim_resource_gateway_url = utils.get_deployment_output(output, 'apimResourceGatewayURL', 'APIM API Gateway URL')\n",
+    "\n"
    ]
   },
   {
@@ -249,11 +206,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# %load ../../shared/snippets/api-http-requests.py\n",
-    "\n",
-    "import json\n",
-    "import requests\n",
-    "import time\n",
+    "import requests, time\n",
     "\n",
     "runs = 10\n",
     "sleep_time_ms = 100\n",
@@ -267,7 +220,7 @@
     "# Initialize a session for connection pooling\n",
     "session = requests.Session()\n",
     "# Set default headers\n",
-    "session.headers.update({'api-key': apim_subscription_key})\n",
+    "session.headers.update({'api-key': apim_subscription_key})  # type: ignore\n",
     "\n",
     "try:\n",
     "    for i in range(runs):\n",
@@ -288,7 +241,7 @@
     "\n",
     "        # Print the response status with the appropriate formatting\n",
     "        print(f\"Response status: {status_code_str}\")\n",
-    "        print(f\"Response headers: {json.dumps(dict(response.headers), indent = 4)}\")\n",
+    "        # print(f\"Response headers: {json.dumps(dict(response.headers), indent = 4)}\")\n",
     "\n",
     "        if \"x-ms-region\" in response.headers:\n",
     "            print(f\"x-ms-region: \\x1b[1;32m{response.headers.get(\"x-ms-region\")}\\x1b[0m\") # this header is useful to determine the region of the backend that served the request\n",
@@ -367,16 +320,14 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# %load ../../shared/snippets/openai-api-requests.py\n",
-    "\n",
     "import time\n",
     "from openai import AzureOpenAI\n",
     "\n",
     "runs = 10\n",
     "sleep_time_ms = 100\n",
     "\n",
     "client = AzureOpenAI(\n",
-    "    azure_endpoint = apim_resource_gateway_url,\n",
+    "    azure_endpoint = apim_resource_gateway_url,  # type: ignore\n",
     "    api_key = apim_subscription_key,\n",
     "    api_version = openai_api_version\n",
     ")\n",
@@ -390,9 +341,11 @@
     "    print(f\"▶️ Run {i+1}/{runs}:\")\n",
     "\n",
     "    start_time = time.time()\n",
-    "    response = client.chat.completions.create(model = openai_model_name, messages = messages) # type: ignore\n",
+    "    raw_response = client.chat.completions.with_raw_response.create(model = openai_model_name, messages = messages) # type: ignore\n",
     "    response_time = time.time() - start_time\n",
     "    print(f\"⌚ {response_time:.2f} seconds\")\n",
+    "    print(f\"x-ms-region: \\x1b[1;32m{raw_response.headers.get(\"x-ms-region\")}\\x1b[0m\") # this header is useful to determine the region of the backend that served the request\n",
+    "    response = raw_response.parse()\n",
     "    if response.usage:\n",
     "        print(f\"Token usage: Total tokens: {response.usage.total_tokens} (Prompt tokens: {response.usage.prompt_tokens} & Completion tokens: {response.usage.completion_tokens})\")\n",
     "    print(f\"💬 {response.choices[0].message.content}\\n\")\n",
@@ -414,7 +367,7 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": ".venv",
+   "display_name": "Python 3",
    "language": "python",
    "name": "python3"
   },

labs/backend-pool-load-balancing/clean-up-resources.ipynb

@@ -15,15 +15,12 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# type: ignore\n",
-    "\n",
-    "import os\n",
-    "import sys\n",
+    "import os, sys\n",
     "sys.path.insert(1, '../../shared')  # add the shared directory to the Python path\n",
-    "from functions import cleanUpResources\n",
+    "import utils\n",
     "\n",
     "deployment_name = os.path.basename(os.path.dirname(globals()['__vsc_ipynb_file__']))\n",
-    "cleanUpResources(deployment_name)"
+    "utils.cleanup_resources(deployment_name)"
    ]
   }
  ],
@@ -43,7 +40,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.13.1"
+   "version": "3.12.8"
   }
  },
  "nbformat": 4,