Aurelis

Azure AI Integration API

Overview

The Azure AI Integration API provides seamless integration with Microsoft Azure AI services, including Azure OpenAI, Azure Cognitive Services, and Azure Machine Learning for enhanced code analysis, generation, and AI-powered development workflows.

Endpoints

Azure OpenAI Integration

POST /api/v1/azure-ai/openai/configure

Configures Azure OpenAI service integration for code generation and analysis.

Request Body

{
  "endpoint": "https://your-resource.openai.azure.com/",
  "api_key": "your_azure_openai_api_key",
  "api_version": "2024-02-15-preview",
  "deployments": [
    {
      "name": "gpt-4",
      "model": "gpt-4",
      "deployment_id": "gpt-4-deployment",
      "capabilities": ["code_generation", "code_analysis", "documentation"]
    },
    {
      "name": "gpt-35-turbo",
      "model": "gpt-35-turbo",
      "deployment_id": "gpt-35-deployment",
      "capabilities": ["code_completion", "quick_analysis"]
    }
  ],
  "default_deployment": "gpt-4",
  "rate_limits": {
    "requests_per_minute": 120,
    "tokens_per_minute": 120000
  }
}

Response

{
  "success": true,
  "configuration_id": "azure_openai_12345",
  "endpoint": "https://your-resource.openai.azure.com/",
  "deployments_configured": 2,
  "status": "active",
  "last_tested": "2025-06-17T10:30:00Z",
  "test_results": {
    "connection": "successful",
    "authentication": "valid",
    "deployments": {
      "gpt-4": "available",
      "gpt-35-turbo": "available"
    }
  }
}

Code Generation with Azure OpenAI

POST /api/v1/azure-ai/openai/generate

Generates code using Azure OpenAI models.

Request Body

{
  "deployment": "gpt-4",
  "prompt": "Create a Python function that validates email addresses",
  "language": "python",
  "context": {
    "existing_code": "import re\n\n# Existing validation functions",
    "requirements": ["RFC 5322 compliance", "return boolean"],
    "style_guide": "PEP 8"
  },
  "parameters": {
    "temperature": 0.3,
    "max_tokens": 1000,
    "top_p": 0.95,
    "frequency_penalty": 0.0,
    "presence_penalty": 0.0
  }
}

Response

{
  "success": true,
  "generation_id": "gen_67890",
  "deployment_used": "gpt-4",
  "generated_code": "def validate_email(email: str) -> bool:\n    \"\"\"\n    Validates an email address according to RFC 5322.\n    \n    Args:\n        email (str): The email address to validate\n        \n    Returns:\n        bool: True if email is valid, False otherwise\n    \"\"\"\n    pattern = r'^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}$'\n    return re.match(pattern, email) is not None",
  "metadata": {
    "tokens_used": 156,
    "processing_time_ms": 1250,
    "model_version": "gpt-4-0613",
    "finish_reason": "stop"
  },
  "quality_metrics": {
    "syntax_valid": true,
    "style_compliance": 95,
    "security_score": 100,
    "performance_rating": "good"
  }
}

Azure Cognitive Services Integration

POST /api/v1/azure-ai/cognitive/configure

Configures Azure Cognitive Services for enhanced code analysis.

Request Body

{
  "services": {
    "text_analytics": {
      "endpoint": "https://your-resource.cognitiveservices.azure.com/",
      "api_key": "your_text_analytics_key",
      "enabled": true
    },
    "translator": {
      "endpoint": "https://api.cognitive.microsofttranslator.com/",
      "api_key": "your_translator_key",
      "enabled": true
    },
    "speech": {
      "endpoint": "https://your-region.stt.speech.microsoft.com/",
      "api_key": "your_speech_key",
      "enabled": false
    }
  },
  "default_language": "en-US",
  "fallback_language": "en"
}

Response

{
  "success": true,
  "configuration_id": "azure_cognitive_54321",
  "services_configured": 2,
  "services_active": ["text_analytics", "translator"],
  "status": "active",
  "last_tested": "2025-06-17T10:30:00Z"
}

Sentiment Analysis for Code Comments

POST /api/v1/azure-ai/cognitive/sentiment

Analyzes sentiment in code comments and documentation.

Request Body

{
  "texts": [
    "# TODO: This is a terrible hack that needs to be fixed",
    "# Excellent implementation of the algorithm",
    "# This function works but could be optimized"
  ],
  "language": "en",
  "include_confidence": true
}

Response

{
  "success": true,
  "results": [
    {
      "text": "# TODO: This is a terrible hack that needs to be fixed",
      "sentiment": "negative",
      "confidence_scores": {
        "positive": 0.05,
        "neutral": 0.10,
        "negative": 0.85
      },
      "analysis": {
        "issues_detected": ["technical_debt", "urgent_todo"],
        "priority": "high",
        "suggestion": "Consider prioritizing this refactoring task"
      }
    },
    {
      "text": "# Excellent implementation of the algorithm",
      "sentiment": "positive",
      "confidence_scores": {
        "positive": 0.92,
        "neutral": 0.06,
        "negative": 0.02
      },
      "analysis": {
        "issues_detected": [],
        "priority": "low",
        "suggestion": "Well-documented code"
      }
    }
  ]
}

Code Translation

POST /api/v1/azure-ai/cognitive/translate

Translates code comments and documentation between languages.

Request Body

{
  "texts": [
    "# Cette fonction valide les adresses email",
    "# 这个函数验证电子邮件地址"
  ],
  "source_language": "auto",
  "target_language": "en",
  "preserve_formatting": true
}

Response

{
  "success": true,
  "translations": [
    {
      "original_text": "# Cette fonction valide les adresses email",
      "translated_text": "# This function validates email addresses",
      "detected_language": "fr",
      "confidence": 0.98
    },
    {
      "original_text": "# 这个函数验证电子邮件地址",
      "translated_text": "# This function validates email addresses",
      "detected_language": "zh-Hans",
      "confidence": 0.95
    }
  ]
}

Azure Machine Learning Integration

POST /api/v1/azure-ai/ml/configure

Configures Azure Machine Learning workspace for custom model deployment.

Request Body

{
  "workspace_name": "aurelis-ml-workspace",
  "resource_group": "aurelis-resources",
  "subscription_id": "your_subscription_id",
  "region": "eastus",
  "authentication": {
    "type": "service_principal",
    "tenant_id": "your_tenant_id",
    "client_id": "your_client_id",
    "client_secret": "your_client_secret"
  },
  "compute_targets": [
    {
      "name": "code-analysis-cluster",
      "type": "AmlCompute",
      "vm_size": "Standard_D3_v2",
      "min_nodes": 0,
      "max_nodes": 4
    }
  ]
}

Response

{
  "success": true,
  "workspace_id": "aml_workspace_99999",
  "workspace_url": "https://ml.azure.com/workspaces/workspace-id",
  "status": "configured",
  "compute_targets_created": 1,
  "available_models": []
}

Custom Model Deployment

POST /api/v1/azure-ai/ml/deploy

Deploys a custom AI model for specialized code analysis.

Request Body

{
  "model_name": "code-quality-analyzer",
  "model_version": "1.0.0",
  "model_file": "model.pkl",
  "environment": {
    "name": "aurelis-env",
    "python_version": "3.8",
    "dependencies": [
      "scikit-learn==1.0.2",
      "pandas==1.3.0",
      "numpy==1.21.0"
    ]
  },
  "deployment_config": {
    "instance_type": "Standard_DS2_v2",
    "instance_count": 1,
    "endpoint_name": "code-quality-endpoint"
  }
}

Response

{
  "success": true,
  "deployment_id": "deploy_88888",
  "endpoint_url": "https://code-quality-endpoint.eastus.inference.ml.azure.com/score",
  "status": "deploying",
  "estimated_completion": "2025-06-17T10:45:00Z",
  "swagger_url": "https://code-quality-endpoint.eastus.inference.ml.azure.com/swagger.json"
}

Model Inference

POST /api/v1/azure-ai/ml/predict

Makes predictions using deployed Azure ML models.

Request Body

{
  "endpoint_name": "code-quality-endpoint",
  "input_data": {
    "code_snippet": "def calculate_average(numbers):\n    return sum(numbers) / len(numbers)",
    "language": "python",
    "context": {
      "file_type": "function",
      "complexity": "low",
      "line_count": 2
    }
  }
}

Response

{
  "success": true,
  "prediction_id": "pred_77777",
  "predictions": {
    "quality_score": 8.5,
    "maintainability": 9.0,
    "readability": 8.8,
    "performance": 7.5,
    "security": 9.2,
    "issues": [
      {
        "type": "potential_bug",
        "severity": "medium",
        "description": "Division by zero possible if empty list provided",
        "suggestion": "Add input validation for empty lists"
      }
    ]
  },
  "confidence": 0.92,
  "processing_time_ms": 150
}

Azure Content Safety

POST /api/v1/azure-ai/safety/analyze

Analyzes code and comments for potentially harmful content.

Request Body

{
  "content": [
    "# This password validation is terrible",
    "def hack_system():",
    "# Regular validation function"
  ],
  "categories": ["hate", "violence", "sexual", "self_harm"],
  "severity_threshold": "low"
}

Response

{
  "success": true,
  "results": [
    {
      "content": "# This password validation is terrible",
      "flagged": false,
      "categories": {
        "hate": {"severity": 0, "flagged": false},
        "violence": {"severity": 0, "flagged": false},
        "sexual": {"severity": 0, "flagged": false},
        "self_harm": {"severity": 0, "flagged": false}
      }
    },
    {
      "content": "def hack_system():",
      "flagged": true,
      "categories": {
        "hate": {"severity": 0, "flagged": false},
        "violence": {"severity": 2, "flagged": true},
        "sexual": {"severity": 0, "flagged": false},
        "self_harm": {"severity": 0, "flagged": false}
      },
      "recommendation": "Review function name for potential security implications"
    }
  ]
}

Usage Analytics

GET /api/v1/azure-ai/analytics

Retrieves usage analytics for Azure AI services.

Response

{
  "success": true,
  "period": "last_30_days",
  "services": {
    "azure_openai": {
      "requests": 15420,
      "tokens_consumed": 2450000,
      "average_response_time_ms": 1250,
      "cost_usd": 245.50,
      "deployments": {
        "gpt-4": {
          "requests": 8520,
          "tokens": 1800000,
          "cost_usd": 180.00
        },
        "gpt-35-turbo": {
          "requests": 6900,
          "tokens": 650000,
          "cost_usd": 65.50
        }
      }
    },
    "cognitive_services": {
      "text_analytics": {
        "requests": 5420,
        "cost_usd": 54.20
      },
      "translator": {
        "characters_translated": 125000,
        "cost_usd": 12.50
      }
    },
    "machine_learning": {
      "inference_requests": 2840,
      "compute_hours": 45.5,
      "cost_usd": 68.25
    }
  },
  "total_cost_usd": 380.45,
  "rate_limit_hits": 12,
  "error_rate": 0.02
}

Authentication

Service Principal Authentication

Recommended for production environments:

{
  "tenant_id": "your_tenant_id",
  "client_id": "your_client_id",
  "client_secret": "your_client_secret"
}

API Key Authentication

For Azure OpenAI and Cognitive Services:

{
  "api_key": "your_service_api_key",
  "endpoint": "https://your-resource.cognitiveservices.azure.com/"
}

Error Handling

Common Error Codes

400: Invalid request parameters or malformed data
401: Authentication failed or invalid API key
403: Insufficient permissions or quota exceeded
404: Resource or deployment not found
429: Rate limit exceeded
500: Azure service error or internal server error

Error Response Format

{
  "success": false,
  "error": {
    "code": "AZURE_QUOTA_EXCEEDED",
    "message": "Azure OpenAI quota exceeded for this deployment",
    "details": {
      "service": "azure_openai",
      "deployment": "gpt-4",
      "quota_type": "tokens_per_minute",
      "limit": 120000,
      "used": 120000,
      "reset_time": "2025-06-17T10:35:00Z"
    }
  }
}

Rate Limiting

Rate limits vary by Azure service:

Azure OpenAI: Based on deployment configuration (TPM/RPM)
Cognitive Services: 1000 requests per minute per key
Azure ML: Based on compute resources and model complexity
Content Safety: 1000 requests per minute per resource

Examples

Configure Azure OpenAI

curl -X POST "https://api.aurelis.dev/v1/azure-ai/openai/configure" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "endpoint": "https://your-resource.openai.azure.com/",
    "api_key": "your_azure_openai_key",
    "deployments": [
      {
        "name": "gpt-4",
        "model": "gpt-4",
        "deployment_id": "gpt-4-deployment"
      }
    ]
  }'

Generate Code

curl -X POST "https://api.aurelis.dev/v1/azure-ai/openai/generate" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "deployment": "gpt-4",
    "prompt": "Create a Python function for email validation",
    "language": "python",
    "parameters": {"temperature": 0.3}
  }'

Analyze Sentiment

curl -X POST "https://api.aurelis.dev/v1/azure-ai/cognitive/sentiment" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "texts": ["# This code needs improvement"],
    "language": "en"
  }'

Integration

Python SDK

from aurelis import AzureAIIntegration

azure_ai = AzureAIIntegration(api_key="your_api_key")

# Configure Azure OpenAI
azure_ai.configure_openai(
    endpoint="https://your-resource.openai.azure.com/",
    api_key="azure_openai_key",
    deployments=[
        {
            "name": "gpt-4",
            "model": "gpt-4",
            "deployment_id": "gpt-4-deployment"
        }
    ]
)

# Generate code
result = azure_ai.generate_code(
    deployment="gpt-4",
    prompt="Create a function for email validation",
    language="python"
)

# Analyze sentiment
sentiment = azure_ai.analyze_sentiment([
    "# This code is excellent",
    "# TODO: Fix this terrible bug"
])

CLI Integration

# Configure Azure OpenAI
aurelis azure-ai configure openai \
  --endpoint https://your-resource.openai.azure.com/ \
  --api-key your_key

# Generate code
aurelis azure-ai generate \
  --deployment gpt-4 \
  --prompt "Create email validation function" \
  --language python

# Analyze sentiment
aurelis azure-ai sentiment "This code needs improvement"

Best Practices

Security: Store Azure credentials securely and rotate regularly
Cost Management: Monitor usage and set appropriate quotas
Performance: Cache results when appropriate to reduce API calls
Error Handling: Implement robust retry logic for transient failures
Monitoring: Track usage, costs, and performance metrics
Compliance: Ensure data handling complies with your organization’s policies

Azure Regions

Supported Azure regions for optimal performance:

North America: East US, West US 2, Central US
Europe: West Europe, North Europe, UK South
Asia Pacific: Southeast Asia, East Asia, Australia East

Limitations

Token Limits: Vary by model and deployment
Request Size: Maximum 4MB per request
Concurrent Requests: Based on deployment tier
Data Residency: Depends on selected Azure region
Model Availability: Not all models available in all regions

Compliance

Azure AI services support various compliance standards:

SOC 1, 2, 3: System and Organization Controls
ISO 27001: Information Security Management
GDPR: General Data Protection Regulation
HIPAA: Health Insurance Portability and Accountability Act
FedRAMP: Federal Risk and Authorization Management Program

Support

For Azure AI integration issues:

Check Azure service health and status
Verify API keys and endpoint URLs
Review quota limits and usage
Contact Azure support for service-specific issues
Contact Aurelis support for integration problems