#!/usr/bin/env python3
"""
Standalone test script for Azure AI Foundry/Azure OpenAI connectivity.
Tests API key authentication and API access.
"""
import json
import sys
import requests
import re
import io
import argparse
# Fix Windows console encoding issues
if sys.platform == 'win32':
sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding='utf-8', errors='replace')
sys.stderr = io.TextIOWrapper(sys.stderr.buffer, encoding='utf-8', errors='replace')
# Default configuration
DEFAULT_API_VERSION = "2024-10-21"
DEFAULT_MODEL = "gpt-4.1"
def normalize_endpoint(endpoint):
"""
Normalize endpoint URL and auto-detect if it's Azure AI Foundry or Azure OpenAI.
Returns: (normalized_endpoint, endpoint_type, resource_name)
"""
endpoint = endpoint.strip().rstrip('/')
# Azure AI Foundry: https://[resource].services.ai.azure.com
# Or project endpoint: https://[resource].services.ai.azure.com/api/projects/[project-name]
if ".services.ai.azure.com" in endpoint:
match = re.search(r'https://([^.]+)\.services\.ai\.azure\.com', endpoint)
if match:
resource_name = match.group(1)
# Extract base resource endpoint
if "/api/projects/" in endpoint:
# Project endpoint - extract base resource
normalized = f"https://{resource_name}.services.ai.azure.com"
return normalized, "Azure AI Foundry (Project)", resource_name
else:
return endpoint, "Azure AI Foundry", resource_name
# Azure OpenAI: https://[resource].openai.azure.com or https://[resource].cognitiveservices.azure.com
elif ".openai.azure.com" in endpoint or ".cognitiveservices.azure.com" in endpoint:
# Try openai.azure.com first
match = re.search(r'https://([^.]+)\.openai\.azure\.com', endpoint)
if match:
resource_name = match.group(1)
return endpoint, "Azure OpenAI", resource_name
# Try cognitiveservices.azure.com
match = re.search(r'https://([^.]+)\.cognitiveservices\.azure\.com', endpoint)
if match:
resource_name = match.group(1)
return endpoint, "Azure OpenAI", resource_name
# Unknown format, return as-is
return endpoint, "Unknown", ""
def build_api_url(endpoint, model, api_version):
"""
Build the API URL based on endpoint type.
"""
endpoint_normalized, endpoint_type, _ = normalize_endpoint(endpoint)
# Azure AI Foundry uses: /models/chat/completions (model in request body)
# Azure OpenAI uses: /openai/deployments/{model}/chat/completions (model in URL)
if endpoint_type == "Azure AI Foundry" or endpoint_type == "Azure AI Foundry (Project)":
# Foundry format: model goes in request body, not URL
url = f"{endpoint_normalized}/models/chat/completions?api-version={api_version}"
elif endpoint_type == "Azure OpenAI":
# Azure OpenAI format: model in URL path
url = f"{endpoint_normalized}/openai/deployments/{model}/chat/completions?api-version={api_version}"
else:
# Unknown format - assume Foundry format by default
url = f"{endpoint_normalized}/models/chat/completions?api-version={api_version}"
return url
def test_azure_ai_api(api_key, endpoint, deployment_name, api_version=DEFAULT_API_VERSION):
"""Test Azure AI API with a simple request."""
print("\n" + "="*60)
print("Testing Azure AI API")
print("="*60)
# Validate inputs
if not api_key or api_key.strip() == "":
print("ERROR: API Key is required")
return False
if not endpoint or endpoint.strip() == "":
print("ERROR: Resource Endpoint is required")
return False
if not deployment_name or deployment_name.strip() == "":
print("ERROR: Deployment Name is required")
return False
# Normalize endpoint - remove trailing slashes to avoid double slashes
endpoint = endpoint.strip().rstrip('/')
endpoint_normalized, endpoint_type, resource_name = normalize_endpoint(endpoint)
endpoint_normalized = endpoint_normalized.rstrip('/') # Ensure no trailing slash
print(f"\nEndpoint Information:")
print(f" Original: {endpoint}")
print(f" Normalized: {endpoint_normalized}")
print(f" Type: {endpoint_type}")
if resource_name:
print(f" Resource: {resource_name}")
# Build URL
url = build_api_url(endpoint_normalized, deployment_name, api_version)
# Ensure URL doesn't have double slashes
url = url.replace('://', '://temp').replace('//', '/').replace('://temp', '://')
print(f"\nRequest URL:")
print(f" {url}")
# Build headers
headers = {
"Content-Type": "application/json",
"api-key": api_key
}
print(f"\nRequest Headers:")
print(f" Content-Type: {headers['Content-Type']}")
print(f" api-key: {api_key[:10]}...{api_key[-5:] if len(api_key) > 15 else ''}")
print(f"\nAPI Version: {api_version}")
print(f"Deployment/Model: {deployment_name}")
print(f"\n⚠️ IMPORTANT: Deployment names are CASE-SENSITIVE!")
print(f" Make sure '{deployment_name}' matches exactly (case) as in Azure Portal")
# Build payload - strategy depends on endpoint type
# For Azure OpenAI: model is in URL, but some API versions accept it in payload too
# For Azure AI Foundry: model MUST be in payload
base_payload = {
"messages": [
{
"role": "user",
"content": "Hello, this is a test. Please respond with 'Test successful'."
}
],
"max_tokens": 100,
"temperature": 0.7
}
# Strategy: For Azure OpenAI, try without model in payload first (recommended)
# For Azure AI Foundry, model must be in payload
payloads_to_try = []
if endpoint_type == "Azure AI Foundry" or endpoint_type == "Azure AI Foundry (Project)":
# Foundry: model MUST be in payload
payload = base_payload.copy()
payload["model"] = deployment_name
payloads_to_try = [("With model in payload (required for Foundry)", payload)]
elif endpoint_type == "Azure OpenAI":
# Azure OpenAI: model is in URL, try without model first (recommended), then with model
payload_no_model = base_payload.copy()
payload_with_model = base_payload.copy()
payload_with_model["model"] = deployment_name
payloads_to_try = [
("Without model in payload (recommended for Azure OpenAI)", payload_no_model),
("With model in payload (fallback)", payload_with_model)
]
else:
# Unknown: try both strategies
payload_no_model = base_payload.copy()
payload_with_model = base_payload.copy()
payload_with_model["model"] = deployment_name
payloads_to_try = [
("Without model in payload", payload_no_model),
("With model in payload", payload_with_model)
]
print(f"\nMaking API request(s)...")
last_error = None
for attempt_name, payload in payloads_to_try:
print(f"\n{'='*60}")
print(f"Attempt: {attempt_name}")
print(f"{'='*60}")
print(f"Request Payload:")
print(json.dumps(payload, indent=2))
try:
response = requests.post(url, json=payload, headers=headers, timeout=30)
print(f"\nResponse Status Code: {response.status_code}")
print(f"Response Headers:")
for key, value in response.headers.items():
if key.lower() in ['content-type', 'content-length', 'x-request-id', 'apim-request-id']:
print(f" {key}: {value}")
print(f"\nResponse Body:")
try:
response_json = response.json()
print(json.dumps(response_json, indent=2))
except:
print(response.text[:1000]) # First 1000 chars if not JSON
if response.status_code == 200:
print("\n" + "="*60)
print("✓ SUCCESS! API request completed successfully")
print("="*60)
# Extract response text
if 'choices' in response_json and len(response_json['choices']) > 0:
result_text = response_json['choices'][0].get('message', {}).get('content', '')
print(f"\nAI Response:")
print(f" {result_text}")
# Show usage info if available
if 'usage' in response_json:
usage = response_json['usage']
print(f"\nToken Usage:")
print(f" Prompt tokens: {usage.get('prompt_tokens', 'N/A')}")
print(f" Completion tokens: {usage.get('completion_tokens', 'N/A')}")
print(f" Total tokens: {usage.get('total_tokens', 'N/A')}")
return True
else:
last_error = (response.status_code, response)
# Continue to next attempt if available
if len(payloads_to_try) > 1 and payloads_to_try.index((attempt_name, payload)) < len(payloads_to_try) - 1:
print(f"\n⚠️ Attempt failed with status {response.status_code}, trying next strategy...")
continue
else:
# Last attempt failed, show detailed error
break
except requests.exceptions.Timeout:
last_error = ("timeout", None)
print(f"\n✗ Network Error: Request timed out after 30 seconds")
if len(payloads_to_try) > 1 and payloads_to_try.index((attempt_name, payload)) < len(payloads_to_try) - 1:
print(" Trying next strategy...")
continue
break
except requests.exceptions.ConnectionError as e:
last_error = ("connection", str(e))
print(f"\n✗ Network Error: Could not connect to endpoint")
print(f" Error: {e}")
if len(payloads_to_try) > 1 and payloads_to_try.index((attempt_name, payload)) < len(payloads_to_try) - 1:
print(" Trying next strategy...")
continue
break
except requests.exceptions.RequestException as e:
last_error = ("request", str(e))
print(f"\n✗ Network Error: {e}")
if len(payloads_to_try) > 1 and payloads_to_try.index((attempt_name, payload)) < len(payloads_to_try) - 1:
print(" Trying next strategy...")
continue
break
except Exception as e:
last_error = ("exception", str(e))
print(f"\n✗ Unexpected Error: {e}")
import traceback
traceback.print_exc()
if len(payloads_to_try) > 1 and payloads_to_try.index((attempt_name, payload)) < len(payloads_to_try) - 1:
print(" Trying next strategy...")
continue
break
# All attempts failed
print("\n" + "="*60)
print("✗ ERROR: All API request attempts failed")
print("="*60)
# Handle last error
if isinstance(last_error, tuple) and len(last_error) == 2:
status_code, response = last_error
if isinstance(status_code, int):
# HTTP error
error_details = {}
try:
error_details = response.json()
except:
error_details = {"error": response.text[:500] if response else "No response"}
if status_code == 401:
print("\n401 Unauthorized Error Analysis:")
print("Common causes:")
print(" 1. Invalid API Key")
print(" 2. API Key expired or revoked")
print(" 3. Wrong API Key for this resource")
print("\nSolution:")
print(" - Verify your API key in Azure Portal")
print(" - Ensure you're using the correct key for this resource")
if 'error' in error_details:
print(f"\nError details from API:")
print(f" {json.dumps(error_details.get('error', {}), indent=2)}")
elif status_code == 404:
print("\n404 Not Found Error Analysis:")
print("="*60)
print("Based on common issues found in Azure OpenAI documentation:")
print("\nMost Common Causes (in order of likelihood):")
print(" 1. ✗ Deployment name mismatch (CASE-SENSITIVE!)")
print(f" Current deployment name: '{deployment_name}'")
print(" → Check Azure Portal and ensure EXACT match (case matters!)")
print(" → Common mistake: 'gpt-4.1' vs 'GPT-4.1' vs 'gpt-41'")
print("\n 2. ✗ Deployment doesn't exist in this resource/region")
print(f" Resource: {resource_name if resource_name else endpoint_normalized}")
print(" → Verify deployment exists in Azure Portal > Your Resource > Deployments")
print(" → Ensure deployment is in the same resource/region as your endpoint")
print("\n 3. ✗ API version incompatible with deployment/model")
print(f" Current API version: {api_version}")
print(" → Try these API versions (newest first):")
print(" • 2025-01-01-preview (latest)")
print(" • 2024-10-21")
print(" • 2024-02-15-preview")
print(" • 2023-12-01-preview")
print(" → For GPT-4.1 and GPT-4o, use: 2024-02-15-preview or later")
print("\n 4. ✗ URL format issue")
print(f" Current URL: {url}")
if endpoint_type == "Azure OpenAI":
print(" → Should be: .../openai/deployments/{deployment-name}/chat/completions?api-version=...")
print(" → Check: No double slashes, deployment name matches exactly")
else:
print(" → Should be: .../models/chat/completions?api-version=...")
print(" → Model should be in request body, not URL")
print("\n 5. ✗ Missing /chat in URL path")
print(" → URL must include '/chat/completions', not just '/completions'")
print(" → This is a common mistake from documentation examples")
if 'error' in error_details:
error_info = error_details.get('error', {})
print(f"\nError details from API:")
print(f" Message: {error_info.get('message', 'N/A')}")
print(f" Code: {error_info.get('code', 'N/A')}")
# Check for specific error messages
error_msg = str(error_info.get('message', '')).lower()
if 'deployment' in error_msg and 'not found' in error_msg:
print("\n⚠️ SPECIFIC ISSUE: Deployment not found")
print(" → Double-check deployment name spelling and case in Azure Portal")
print(" → Verify the deployment exists and is active")
elif 'api deployment' in error_msg:
print("\n⚠️ SPECIFIC ISSUE: API deployment issue")
print(" → The deployment may not be active or may be in a different region")
elif 'resource' in error_msg and 'not found' in error_msg:
print("\n⚠️ SPECIFIC ISSUE: Resource not found")
print(" → Verify your endpoint URL is correct")
print(" → Check that the resource exists in Azure Portal")
print("\n" + "="*60)
print("Recommended Fixes (try in this order):")
print("="*60)
print("1. Verify deployment name in Azure Portal:")
print(" - Go to Azure Portal > Your Resource > Deployments")
print(" - Copy the EXACT deployment name (including case)")
print(" - Common issue: Case sensitivity - 'gpt-4.1' ≠ 'GPT-4.1'")
print()
print("2. Verify deployment exists in the same resource:")
print(f" - Resource: {resource_name if resource_name else endpoint_normalized}")
print(" - Deployments are resource/region specific")
print()
print("3. Try a different API version:")
print(" python test_azure_ai.py --api-key YOUR_KEY --endpoint YOUR_ENDPOINT \\")
print(" --deployment YOUR_DEPLOYMENT --api-version 2025-01-01-preview")
print()
print("4. Verify URL format:")
print(f" Current: {url}")
print(" - No trailing slashes before path")
print(" - No double slashes")
print(" - Includes '/chat/completions'")
print("="*60)
elif status_code == 400:
print("\n400 Bad Request Error Analysis:")
print("Possible issues:")
print(" 1. API version is not supported")
print(f" 2. Model/deployment name format issue")
print(" 3. Request payload format issue")
if 'error' in error_details:
error_info = error_details.get('error', {})
print(f"\nError details from API:")
print(f" Message: {error_info.get('message', 'N/A')}")
print(f" Code: {error_info.get('code', 'N/A')}")
elif status_code == 429:
print("\n429 Too Many Requests Error Analysis:")
print("You've exceeded the rate limit for this resource.")
print("Solution:")
print(" - Wait a few moments and try again")
print(" - Check your rate limits in Azure Portal")
else:
print(f"\n{status_code} Error:")
if 'error' in error_details:
error_info = error_details.get('error', {})
print(f" Message: {error_info.get('message', 'N/A')}")
print(f" Code: {error_info.get('code', 'N/A')}")
elif status_code == "timeout":
print("\n✗ Network Error: Request timed out after 30 seconds")
print(" - Check your internet connection")
print(" - Verify the endpoint URL is accessible")
elif status_code == "connection":
print("\n✗ Network Error: Could not connect to endpoint")
print(f" Error: {last_error[1]}")
print(" - Verify the endpoint URL is correct")
print(" - Check your network connection")
print(" - Ensure the endpoint is accessible from your network")
elif status_code == "request":
print(f"\n✗ Network Error: {last_error[1]}")
elif status_code == "exception":
print(f"\n✗ Unexpected Error: {last_error[1]}")
return False
def get_user_input(prompt, default=None, password=False):
"""Get user input with optional default value."""
if default:
prompt_text = f"{prompt} [{default}]: "
else:
prompt_text = f"{prompt}: "
if password:
import getpass
value = getpass.getpass(prompt_text)
else:
value = input(prompt_text).strip()
return value if value else default
def main():
parser = argparse.ArgumentParser(
description="Test Azure AI Foundry/Azure OpenAI connectivity",
formatter_class=argparse.RawDescriptionHelpFormatter,
epilog="""
Examples:
# Interactive mode
python test_azure_ai.py
# Command line arguments
python test_azure_ai.py --api-key YOUR_KEY --endpoint https://your-resource.services.ai.azure.com --deployment gpt-4.1
# Azure OpenAI
python test_azure_ai.py --api-key YOUR_KEY --endpoint https://your-resource.openai.azure.com --deployment gpt-4
# Azure OpenAI (cognitiveservices.azure.com)
python test_azure_ai.py --api-key YOUR_KEY --endpoint https://your-resource.cognitiveservices.azure.com --deployment gpt-4.1
# Custom API version
python test_azure_ai.py --api-key YOUR_KEY --endpoint https://your-resource.services.ai.azure.com --deployment gpt-4.1 --api-version 2024-02-15-preview
"""
)
parser.add_argument('--api-key', '-k', help='API Key')
parser.add_argument('--endpoint', '-e', help='Resource Endpoint URL')
parser.add_argument('--deployment', '-d', help='Deployment Name (Model)')
parser.add_argument('--api-version', '-v', default=DEFAULT_API_VERSION, help=f'API Version (default: {DEFAULT_API_VERSION})')
args = parser.parse_args()
print("="*60)
print("Azure AI API Test Script")
print("="*60)
print()
# Get inputs (command line args or interactive)
api_key = args.api_key or get_user_input("Enter API Key", password=True)
endpoint = args.endpoint or get_user_input("Enter Resource Endpoint", default="https://your-resource.services.ai.azure.com")
deployment_name = args.deployment or get_user_input("Enter Deployment Name (Model)", default=DEFAULT_MODEL)
api_version = args.api_version
# Test API
success = test_azure_ai_api(api_key, endpoint, deployment_name, api_version)
if not success:
print("\n" + "="*60)
print("Troubleshooting Steps:")
print("="*60)
print("1. Verify your API Key:")
print(" - Azure AI Foundry: https://learn.microsoft.com/en-us/azure/ai-foundry/foundry-models/how-to/quickstart-ai-project")
print(" - Azure OpenAI: https://portal.azure.com > Your Resource > Keys and Endpoint")
print()
print("2. Verify your Resource Endpoint:")
print(" - Should be: https://[resource-name].services.ai.azure.com (Foundry)")
print(" - Or: https://[resource-name].openai.azure.com (Azure OpenAI)")
print(" - Or: https://[resource-name].cognitiveservices.azure.com (Azure OpenAI)")
print(" - Or: https://[resource-name].services.ai.azure.com/api/projects/[project-name] (Foundry Project)")
print()
print("3. Verify your Deployment Name:")
print(" - Check in Azure Portal that the deployment exists")
print(" - Ensure the deployment name matches exactly (case-sensitive)")
print()
print("4. Check API Version:")
print(f" - Current: {api_version}")
print(" - Try: 2024-10-21, 2024-02-15-preview, or 2023-12-01-preview")
print()
print("5. Verify Network Access:")
print(" - Ensure you can reach the endpoint from your network")
print(" - Check firewall/proxy settings if applicable")
print("="*60)
sys.exit(1)
else:
print("\n✓ All tests passed! Azure AI is properly configured.")
print("\nYou can now use these settings in Pomera AI Commander:")
print(f" API Key: {api_key[:10]}...{api_key[-5:] if len(api_key) > 15 else ''}")
print(f" Resource Endpoint: {endpoint}")
print(f" Deployment/Model: {deployment_name}")
print(f" API Version: {api_version}")
sys.exit(0)
if __name__ == "__main__":
main()
