src/api/routes.py

from fastapi import APIRouter, File, UploadFile, Form, Depends, HTTPException, status
from fastapi.responses import JSONResponse
from src.services.image_processor import image_processor
from src.services.ocr_engine import ocr_service
from src.services.markdown_formatter import format_ocr_result
from src.api.dependencies import verify_api_key

router = APIRouter()


@router.get("/health")
def health_check():
    """Unauthenticated endpoint to check server health."""
    return {"status": "healthy"}


@router.post("/api/v1/image/extract-text")
async def extract_text(
    image: UploadFile = File(...),
    lang: str = Form("en", description="Target language code (e.g. 'en', 'hi', 'ar') or 'auto' for automatic language detection."),
    api_key: str = Depends(verify_api_key)
):
    """
    Receives an image file via multipart form, processes it with PaddleOCR,
    and returns both plain text and structured Markdown in the JSON response.

    Supported lang values: 'auto' (automatic detection), or specific codes like 'en', 'hi', 'ar', 'es', 'pt', etc.

    Response shape:
    {
        "success": true,
        "data": {
            "result":   "<plain text, newline-joined>",
            "markdown": "<structured Markdown with headings, tables, key-values>",
            "fileName": "<original filename>"
        }
    }
    """
    try:
        # Validate mime type
        if not image.content_type.startswith("image/"):
            raise HTTPException(
                status_code=status.HTTP_400_BAD_REQUEST,
                detail="Uploaded file is not a supported image format."
            )

        # Read raw bytes
        image_bytes = await image.read()

        # Preprocess: EXIF correction → resolution guard → CLAHE → unsharp mask
        img_np = image_processor.process_bytes_to_numpy(image_bytes)

        # Run OCR — returns raw bounding-box result list (not yet formatted)
        raw_result = await ocr_service.run_ocr_raw(img_np, lang)

        # Convert raw result → plain text + structured Markdown
        plain_text, markdown_text = format_ocr_result(raw_result)

        return {
            "success": True,
            "data": {
                "result":   plain_text,
                "markdown": markdown_text,
                "fileName": image.filename or "uploaded_image.jpg"
            }
        }

    except HTTPException as he:
        raise he
    except Exception as e:
        import traceback
        traceback.print_exc()
        return JSONResponse(
            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
            content={
                "success": False,
                "error": f"OCR processing failed: {str(e)}"
            }
        )