Function ocr_recognition_file

ddddocr/api/app.py:409–495

OCR文字识别 - 接收上传的图片文件

(
    background_tasks: BackgroundTasks,
    file: UploadFile = File(...),
    probability: Union[bool, str] = Form(False),
    colors: str = Form("[]"),
    custom_color_ranges: str = Form("null"),
    ocr: bool = Query(default_ocr, description="是否启用OCR功能"),
    det: bool = Query(default_det, description="是否启用目标检测功能"),
    old: bool = Query(default_old, description="是否使用旧版OCR模型"),
    beta: bool = Query(default_beta, description="是否使用Beta版OCR模型"),
    use_gpu: bool = Query(default_use_gpu, description="是否使用GPU加速"),
    device_id: int = Query(default_device_id, description="GPU设备ID"),
    show_ad: bool = Query(default_show_ad, description="是否显示广告")
)

# OCR识别端点 - 文件上传
@app.post("/ocr/file", response_model=OCRResponse)
async def ocr_recognition_file(
    background_tasks: BackgroundTasks,
    file: UploadFile = File(...),
    probability: Union[bool, str] = Form(False),
    colors: str = Form("[]"),
    custom_color_ranges: str = Form("null"),
    ocr: bool = Query(default_ocr, description="是否启用OCR功能"),
    det: bool = Query(default_det, description="是否启用目标检测功能"),
    old: bool = Query(default_old, description="是否使用旧版OCR模型"),
    beta: bool = Query(default_beta, description="是否使用Beta版OCR模型"),
    use_gpu: bool = Query(default_use_gpu, description="是否使用GPU加速"),
    device_id: int = Query(default_device_id, description="GPU设备ID"),
    show_ad: bool = Query(default_show_ad, description="是否显示广告")
):
    """
    OCR文字识别 - 接收上传的图片文件
    """
    image = None
    try:
        contents = await file.read()
    except Exception as exc:
        raise HTTPException(status_code=400, detail="无法读取上传文件") from exc

    if not contents:
        raise HTTPException(status_code=400, detail="上传文件为空")
    if len(contents) > CORE_MAX_IMAGE_BYTES:
        raise HTTPException(
            status_code=400,
            detail=f"图片大小超过 {CORE_MAX_IMAGE_BYTES // 1024}KB 限制"
        )
    try:
        image = Image.open(io.BytesIO(contents))
    except Exception as exc:
        raise HTTPException(status_code=400, detail="无法解析上传的图片") from exc

    try:
        colors_data = json.loads(colors) if colors else []
    except json.JSONDecodeError as exc:
        raise HTTPException(status_code=400, detail="colors JSON 解析失败") from exc
    colors_list = _ensure_colors_list(colors_data)
    custom_ranges = _ensure_custom_ranges(custom_color_ranges)
    probability_flag = _coerce_bool_param(probability, 'probability')

    config_key = f"ocr={ocr}-det={det}-old={old}-beta={beta}-gpu={use_gpu}-dev={device_id}"

    ocr_instance = get_ocr_instance(
        config_key, ocr, det, old, beta, use_gpu, device_id, show_ad,
        default_import_onnx_path, default_charsets_path
    )

    start_time = time.time()
    try:
        if probability_flag:
            result = ocr_instance.classification(
                image,
                probability=True,
                colors=colors_list,
                custom_color_ranges=custom_ranges