87 lines
3.3 KiB
Python
87 lines
3.3 KiB
Python
"""Image OCR endpoint."""
|
|
|
|
from fastapi import APIRouter, Depends, HTTPException
|
|
|
|
from app.core.dependencies import get_image_processor, get_layout_detector, get_ocr_service, get_mineru_ocr_service, get_converter
|
|
from app.schemas.image import ImageOCRRequest, ImageOCRResponse, LatexToOmmlRequest, LatexToOmmlResponse
|
|
from app.services.image_processor import ImageProcessor
|
|
from app.services.layout_detector import LayoutDetector
|
|
from app.services.ocr_service import OCRService, MineruOCRService
|
|
from app.services.converter import Converter
|
|
|
|
router = APIRouter()
|
|
|
|
|
|
@router.post("/ocr", response_model=ImageOCRResponse)
|
|
async def process_image_ocr(
|
|
request: ImageOCRRequest,
|
|
image_processor: ImageProcessor = Depends(get_image_processor),
|
|
layout_detector: LayoutDetector = Depends(get_layout_detector),
|
|
mineru_service: MineruOCRService = Depends(get_mineru_ocr_service),
|
|
paddle_service: OCRService = Depends(get_ocr_service),
|
|
) -> ImageOCRResponse:
|
|
"""Process an image and extract content as LaTeX, Markdown, and MathML.
|
|
|
|
The processing pipeline:
|
|
1. Load and preprocess image (add 30% whitespace padding)
|
|
2. Detect layout using DocLayout-YOLO
|
|
3. Based on layout:
|
|
- If plain text exists: use PP-DocLayoutV2 for mixed recognition
|
|
- Otherwise: use PaddleOCR-VL with formula prompt
|
|
4. Convert output to LaTeX, Markdown, and MathML formats
|
|
|
|
Note: OMML conversion is not included due to performance overhead.
|
|
Use the /latex-to-omml endpoint to convert LaTeX to OMML separately.
|
|
"""
|
|
|
|
image = image_processor.preprocess(
|
|
image_url=request.image_url,
|
|
image_base64=request.image_base64,
|
|
)
|
|
|
|
try:
|
|
if request.model_name == "mineru":
|
|
ocr_result = mineru_service.recognize(image)
|
|
elif request.model_name == "paddle":
|
|
ocr_result = paddle_service.recognize(image)
|
|
else:
|
|
raise HTTPException(status_code=400, detail="Invalid model name")
|
|
except RuntimeError as e:
|
|
raise HTTPException(status_code=503, detail=str(e))
|
|
|
|
return ImageOCRResponse(
|
|
latex=ocr_result.get("latex", ""),
|
|
markdown=ocr_result.get("markdown", ""),
|
|
mathml=ocr_result.get("mathml", ""),
|
|
mml=ocr_result.get("mml", ""),
|
|
)
|
|
|
|
|
|
@router.post("/latex-to-omml", response_model=LatexToOmmlResponse)
|
|
async def convert_latex_to_omml(
|
|
request: LatexToOmmlRequest,
|
|
converter: Converter = Depends(get_converter),
|
|
) -> LatexToOmmlResponse:
|
|
"""Convert LaTeX formula to OMML (Office Math Markup Language).
|
|
|
|
OMML is the math format used by Microsoft Word and other Office applications.
|
|
This endpoint is separate from the main OCR endpoint due to the performance
|
|
overhead of OMML conversion (requires creating a temporary DOCX file).
|
|
|
|
Args:
|
|
request: Contains the LaTeX formula to convert (without $ or $$ delimiters).
|
|
|
|
Returns:
|
|
OMML representation of the formula.
|
|
"""
|
|
if not request.latex or not request.latex.strip():
|
|
raise HTTPException(status_code=400, detail="LaTeX formula cannot be empty")
|
|
|
|
try:
|
|
omml = converter.convert_to_omml(request.latex)
|
|
return LatexToOmmlResponse(omml=omml)
|
|
except ValueError as e:
|
|
raise HTTPException(status_code=400, detail=str(e))
|
|
except RuntimeError as e:
|
|
raise HTTPException(status_code=503, detail=str(e))
|