Merge branch 'main' into feature/consolidate-bicep

pyrightconfig.json

@@ -0,0 +1,5 @@
+{
+    "ignore": [
+        "shared"
+    ]
+}

shared/functions.py

@@ -1,5 +1,3 @@
-# type: ignore
-
 import datetime
 import json
 import subprocess

shared/snippets/api-http-requests.py

@@ -0,0 +1,45 @@
+import requests
+import time
+
+runs = 10
+sleep_time_ms = 100
+url = f"{apim_resource_gateway_url}/openai/deployments/{openai_deployment_name}/chat/completions?api-version={openai_api_version}"
+api_runs = []
+
+for i in range(runs):
+    print(f"▶️ Run {i+1}/{runs}:")
+
+    messages = {"messages": [
+        {"role": "system", "content": "You are a sarcastic, unhelpful assistant."},
+        {"role": "user", "content": "Can you tell me the time, please?"}
+    ]}
+
+    start_time = time.time()
+    response = requests.post(url, headers = {'api-key':apim_subscription_key}, json = messages)
+    response_time = time.time() - start_time
+    print(f"⌚ {response_time:.2f} seconds")
+
+    # Check the response status code and apply formatting
+    if 200 <= response.status_code < 300:
+        status_code_str = f"\x1b[1;32m{response.status_code} - {response.reason}\x1b[0m" # Bold and green
+    elif response.status_code >= 400:
+        status_code_str = f"\x1b[1;31m{response.status_code} - {response.reason}\x1b[0m" # Bold and red
+    else:
+        status_code_str = str(response.status_code)  # No formatting
+
+    # Print the response status with the appropriate formatting
+    print(f"Response status: {status_code_str}")
+    print(f"Response headers: {response.headers}")
+
+    if "x-ms-region" in response.headers:
+        print(f"x-ms-region: \x1b[1;31m{response.headers.get("x-ms-region")}\x1b[0m") # this header is useful to determine the region of the backend that served the request
+        api_runs.append((response_time, response.headers.get("x-ms-region")))
+
+    if (response.status_code == 200):
+        data = json.loads(response.text)
+        print(f"Token usage: {data.get("usage")}\n")
+        print(f"💬 {data.get("choices")[0].get("message").get("content")}\n")
+    else:
+        print(response.text)
+
+    time.sleep(sleep_time_ms/1000)

shared/snippets/create-az-deployment.py

@@ -0,0 +1,25 @@
+import json
+
+backend_id = "openai-backend-pool" if len(openai_resources) > 1 else openai_resources[0].get("name")
+
+with open("policy.xml", 'r') as policy_xml_file:
+    policy_template_xml = policy_xml_file.read()
+    policy_xml = policy_template_xml.replace("{backend-id}", backend_id)
+    policy_xml_file.close()
+open("policy-updated.xml", 'w').write(policy_xml)
+
+bicep_parameters = {
+    "$schema": "https://schema.management.azure.com/schemas/2019-04-01/deploymentParameters.json#",
+    "contentVersion": "1.0.0.0",
+    "parameters": {
+        "openAIConfig": { "value": openai_resources },
+        "openAIDeploymentName": { "value": openai_deployment_name },
+        "openAIModelName": { "value": openai_model_name },
+        "openAIModelVersion": { "value": openai_model_version },
+    }
+}
+
+with open('params.json', 'w') as bicep_parameters_file:
+    bicep_parameters_file.write(json.dumps(bicep_parameters))
+
+! az deployment group create --name {deployment_name} --resource-group {resource_group_name} --template-file "main.bicep" --parameters "params.json"

shared/snippets/create-az-resource-group.py

@@ -0,0 +1,6 @@
+resource_group_stdout = ! az group create --name {resource_group_name} --location {resource_group_location}
+
+if resource_group_stdout.n.startswith("ERROR"):
+    print(resource_group_stdout)
+else:
+    print(f"✅ Azure Resource Group {resource_group_name} created ⌚ {datetime.datetime.now().time()}")

shared/snippets/deployment-outputs.py

@@ -0,0 +1,18 @@
+stdout = ! az deployment group show --name {deployment_name} -g {resource_group_name} --query properties.outputs.apimServiceId.value -o tsv
+apim_service_id = stdout.n
+print(f"👉🏻 APIM Service Id: {apim_service_id}")
+
+stdout = ! az deployment group show --name {deployment_name} -g {resource_group_name} --query properties.outputs.apimSubscriptionKey.value -o tsv
+print(f"👉🏻 APIM Subscription Key (masked): ***...{apim_subscription_key[-4:]}")
+
+stdout = ! az deployment group show --name {deployment_name} -g {resource_group_name} --query properties.outputs.apimResourceGatewayURL.value -o tsv
+apim_resource_gateway_url = stdout.n
+print(f"👉🏻 APIM API Gateway URL: {apim_resource_gateway_url}")
+
+stdout = ! az deployment group show --name {deployment_name} -g {resource_group_name} --query properties.outputs.logAnalyticsWorkspaceId.value -o tsv
+workspace_id = stdout.n
+print(f"👉🏻 Workspace ID: {workspace_id}")
+
+stdout = ! az deployment group show --name {deployment_name} -g {resource_group_name} --query properties.outputs.applicationInsightsAppId.value -o tsv
+app_id = stdout.n
+print(f"👉🏻 App ID: {app_id}")

shared/snippets/openai-api-requests.py

@@ -0,0 +1,27 @@
+import time
+from openai import AzureOpenAI
+
+runs = 10
+sleep_time_ms = 100
+
+client = AzureOpenAI(
+    azure_endpoint = apim_resource_gateway_url,
+    api_key = apim_subscription_key,
+    api_version = openai_api_version
+)
+
+for i in range(runs):
+    print(f"▶️ Run {i+1}/{runs}:")
+
+    messages = {"messages": [
+        {"role": "system", "content": "You are a sarcastic, unhelpful assistant."},
+        {"role": "user", "content": "Can you tell me the time, please?"}
+    ]}
+
+    start_time = time.time()
+    response = client.chat.completions.create(model = openai_model_name, messages = messages)
+    response_time = time.time() - start_time
+    print(f"⌚ {response_time:.2f} seconds")
+    print(f"💬 {response.choices[0].message.content}\n")
+
+    time.sleep(sleep_time_ms/1000)