ziadalaa7/Fake_Image_Detection

1"""
2FastAPI routes and endpoints.
3Defines the API contract and request/response handling.
4This is the "View" layer responsible for HTTP handling.
5"""
6
7from fastapi import APIRouter, HTTPException, status
8from fastapi.responses import JSONResponse
9import time
10from models.schemas import PredictionResponse, ErrorResponse, HealthCheckResponse, ImageURLRequest
11from services.model_service import get_model_service
12from core.config import CLASS_LABELS
13
14# Create router for v1 API
15router = APIRouter(prefix="/api/v1", tags=["Predictions"])
16
17
18@router.get(
19    "/health",
20    response_model=HealthCheckResponse,
21    summary="Health Check",
22    description="Check if the service is running and model is loaded"
23)
24async def health_check() -> HealthCheckResponse:
25    """
26    Health check endpoint to verify service status.
27    
28    Returns:
29        HealthCheckResponse: Service health status and model state.
30    """
31    try:
32        model_service = get_model_service()
33        return HealthCheckResponse(
34            status="healthy",
35            model_loaded=model_service.model is not None,
36            version="1.0.0"
37        )
38    except Exception as e:
39        raise HTTPException(
40            status_code=status.HTTP_503_SERVICE_UNAVAILABLE,
41            detail=f"Service unavailable: {str(e)}"
42        )
43
44
45@router.post(
46    "/predict",
47    response_model=PredictionResponse,
48    summary="Predict Image Classification from URL",
49    description="Send an image URL to detect whether it's real or AI-generated",
50    responses={
51        200: {
52            "description": "Successful prediction",
53            "model": PredictionResponse
54        },
55        400: {
56            "description": "Invalid request or URL",
57            "model": ErrorResponse
58        },
59        500: {
60            "description": "Server error during processing",
61            "model": ErrorResponse
62        }
63    }
64)
65async def predict_image(request: ImageURLRequest) -> PredictionResponse:
66    """
67    Analyze an image from URL and get a prediction of whether it's real or AI-generated.
68    
69    Args:
70        request (ImageURLRequest): Request containing the image URL from S3.
71        
72    Returns:
73        PredictionResponse: Prediction with verdict ("real" or "fake") and confidence score (0-100).
74        
75    Raises:
76        HTTPException: If URL is invalid or prediction fails.
77        
78    Example:
79        ```bash
80        curl -X POST "http://localhost:8000/api/v1/predict" \
81          -H "Content-Type: application/json" \
82          -d '{"image_url": "https://my-bucket.s3.amazonaws.com/image.jpg"}'
83        ```
84    """
85    
86    try:
87        # Validate URL format
88        if not request.image_url or not request.image_url.startswith(('http://', 'https://')):
89            raise HTTPException(
90                status_code=status.HTTP_400_BAD_REQUEST,
91                detail="Invalid URL. Must start with http:// or https://"
92            )
93        
94        # Get model service and perform prediction
95        model_service = get_model_service()
96        prediction_result = model_service.predict(request.image_url)
97        
98        # Return formatted response
99        return PredictionResponse(
100            verdict=prediction_result["verdict"],
101            confidenceScore=prediction_result["confidenceScore"]
102        )
103    
104    except HTTPException:
105        raise
106    except ValueError as e:
107        raise HTTPException(
108            status_code=status.HTTP_400_BAD_REQUEST,
109            detail=str(e)
110        )
111    except Exception as e:
112        raise HTTPException(
113            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
114            detail=f"Prediction failed: {str(e)}"
115        )
116
117
118@router.get(
119    "/info",
120    summary="Model Information",
121    description="Get information about the loaded model"
122)
123async def model_info():
124    """
125    Get metadata about the currently loaded model.
126    
127    Returns:
128        dict: Model information including name, version, and class labels.
129    """
130    try:
131        return {
132            "model_name": "EfficientNet-Fine-Tuned",
133            "version": "1.0.0",
134            "class_labels": CLASS_LABELS,
135            "supported_formats": list(ALLOWED_EXTENSIONS),
136            "max_file_size_mb": MAX_IMAGE_SIZE_MB,
137            "input_shape": (224, 224, 3)
138        }
139    except Exception as e:
140        raise HTTPException(
141            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
142            detail=f"Failed to retrieve model info: {str(e)}"
143        )
144