Files
OCR/backend/api/documents.py

55 lines
2.2 KiB
Python

import json
from fastapi import APIRouter, Depends, HTTPException, status, UploadFile, File
from sqlalchemy.orm import Session
from typing import List
from database import get_db
from models.template_models import TemplateDocument, DocumentLayout
from schemas.template_schemas import DocumentSchema, DocumentLayoutSchema, TemplateRecognitionResult
from engine.ocr.document_processor import DocumentProcessor
from engine.recognition.template_engine import TemplateRecognitionEngine
router = APIRouter(prefix="/api/documents", tags=["Documents"])
@router.post("/upload", response_model=DocumentSchema, status_code=status.HTTP_201_CREATED)
async def upload_document(file: UploadFile = File(...), db: Session = Depends(get_db)):
# 1. Create document record
content = await file.read()
file_hash = hash(content) # Simple hash for demo
db_doc = TemplateDocument(
document_name=file.filename,
document_type=file.content_type,
file_name=file.filename,
file_hash=str(file_hash),
status="processing"
)
db.add(db_doc)
db.commit()
db.refresh(db_doc)
# 2. Process File (OCR & Layout)
processor = DocumentProcessor(db)
processor.process_file(db_doc.pk_document_id, file.filename, content)
db.refresh(db_doc)
return db_doc
@router.get("/{document_id}", response_model=DocumentSchema)
def get_document(document_id: int, db: Session = Depends(get_db)):
doc = db.query(TemplateDocument).filter(TemplateDocument.pk_document_id == document_id).first()
if not doc:
raise HTTPException(status_code=404, detail="Document not found")
return doc
@router.get("/{document_id}/layout", response_model=List[DocumentLayoutSchema])
def get_document_layout(document_id: int, db: Session = Depends(get_db)):
layouts = db.query(DocumentLayout).filter(DocumentLayout.fk_document_id == document_id).order_by(DocumentLayout.sequence_no).all()
return layouts
@router.post("/{document_id}/recognize", response_model=TemplateRecognitionResult)
def recognize_template(document_id: int, db: Session = Depends(get_db)):
engine = TemplateRecognitionEngine(db)
return engine.recognize_template(document_id)