doc_processer/app/api/v1/endpoints/image.py

"""Image OCR endpoint."""

import time
import uuid

from fastapi import APIRouter, Depends, HTTPException, Request, Response

from app.core.dependencies import (
    get_image_processor,
    get_glmocr_endtoend_service,
)
from app.core.logging_config import get_logger, RequestIDAdapter
from app.schemas.image import ImageOCRRequest, ImageOCRResponse
from app.services.image_processor import ImageProcessor
from app.services.ocr_service import GLMOCREndToEndService

router = APIRouter()
logger = get_logger()


@router.post("/ocr", response_model=ImageOCRResponse)
async def process_image_ocr(
    request: ImageOCRRequest,
    http_request: Request,
    response: Response,
    image_processor: ImageProcessor = Depends(get_image_processor),
    glmocr_service: GLMOCREndToEndService = Depends(get_glmocr_endtoend_service),
) -> ImageOCRResponse:
    """Process an image and extract content as LaTeX, Markdown, and MathML.

    The processing pipeline:
    1. Load and preprocess image
    2. Detect layout regions using PP-DocLayoutV3
    3. Crop each region and recognize with GLM-OCR via vLLM (task-specific prompts)
    4. Aggregate region results into Markdown
    5. Convert to LaTeX, Markdown, and MathML formats

    Note: OMML conversion is not included due to performance overhead.
    Use the /convert/latex-to-omml endpoint to convert LaTeX to OMML separately.
    """
    request_id = http_request.headers.get("x-request-id", str(uuid.uuid4()))
    response.headers["x-request-id"] = request_id

    log = RequestIDAdapter(logger, {"request_id": request_id})
    log.request_id = request_id

    try:
        log.info("Starting image OCR processing")
        start = time.time()

        image = image_processor.preprocess(
            image_url=request.image_url,
            image_base64=request.image_base64,
        )

        ocr_result = glmocr_service.recognize(image)

        log.info(f"OCR completed in {time.time() - start:.3f}s")

    except RuntimeError as e:
        log.error(f"OCR processing failed: {str(e)}", exc_info=True)
        raise HTTPException(status_code=503, detail=str(e))
    except Exception as e:
        log.error(f"Unexpected error during OCR processing: {str(e)}", exc_info=True)
        raise HTTPException(status_code=500, detail="Internal server error")

    return ImageOCRResponse(
        latex=ocr_result.get("latex", ""),
        markdown=ocr_result.get("markdown", ""),
        mathml=ocr_result.get("mathml", ""),
        mml=ocr_result.get("mml", ""),
    )