"""
OCR API Route — Accepts document images and returns extracted structured data.
Uses the multi-provider OCR router (Groq → Gemini → OpenAI → HuggingFace).
"""

from fastapi import APIRouter, UploadFile, File, Form, HTTPException
from app.schemas import OCRResponse, DocumentType
from app.services.ocr.ocr_router import ocr_router
from PIL import Image
import asyncio
import io
import logging

router = APIRouter()
logger = logging.getLogger(__name__)

MAX_FILE_SIZE_MB = 10
MAX_FILE_SIZE_BYTES = MAX_FILE_SIZE_MB * 1024 * 1024


@router.post("/ocr", response_model=OCRResponse)
async def process_ocr(
    file: UploadFile = File(..., description="Document image (PNG, JPG, WEBP)"),
    document_type: DocumentType = Form(..., description="Type of the document being scanned"),
):
    """
    Process an Arabic document image and extract structured data as JSON.

    - **file**: The document image to scan (max 10 MB).
    - **document_type**: One of the supported document types (id_card, income_proof, etc.).

    The router automatically tries providers in order (Groq → Gemini → OpenAI → HuggingFace)
    and falls back if any provider hits a rate limit or error.
    """
    # ── Validate file type ──
    if not file.content_type or not file.content_type.startswith("image/"):
        raise HTTPException(status_code=400, detail="File provided is not an image.")

    # ── Validate file size (safe: only read limit + 1 byte) ──
    contents = await file.read(MAX_FILE_SIZE_BYTES + 1)
    if len(contents) > MAX_FILE_SIZE_BYTES:
        raise HTTPException(
            status_code=400,
            detail=f"File size exceeds the {MAX_FILE_SIZE_MB} MB limit.",
        )

    try:
        image = await asyncio.to_thread(Image.open, io.BytesIO(contents))
        result_data = await ocr_router.run_ocr(image, document_type=document_type)

        # Extract '_provider' from result (injected by the router)
        provider_name = result_data.pop("_provider", "unknown")

        return OCRResponse(data=result_data, provider=provider_name)

    except RuntimeError as e:
        # All providers exhausted
        logger.error("All OCR providers failed: %s", e)
        raise HTTPException(status_code=503, detail="All OCR providers are currently unavailable. Please try again later.")
    except Exception as e:
        logger.error("Error processing OCR: %s", e, exc_info=True)
        raise HTTPException(status_code=500, detail="Internal server error during OCR processing.")