What causes "{model_path} is not a file." in opendatalab/MinerU?

Point model_path at the .onnx file itself, not its containing directory.

opendatalab/MinerU · error · FileExistsError

{model_path} is not a file.

Error message

{model_path} is not a file.

What it means

The companion check in _verify_model: the path exists but is not a regular file, so it raises FileExistsError('<path> is not a file.'). Typical causes: the path is a directory (user pointed at the model folder instead of the .onnx file), a symlink whose target is gone, or a device/socket path.

Source

Thrown at mineru/model/table/rec/slanet_plus/table_structure_utils.py:109

        return meta_dict[key].splitlines()

    def have_key(self, key: str = "character") -> bool:
        meta_dict = self.session.get_modelmeta().custom_metadata_map
        if key in meta_dict.keys():
            return True
        return False

    @staticmethod
    def _verify_model(model_path: Union[str, Path, None]):
        if model_path is None:
            raise ValueError("model_path is None!")

        model_path = Path(model_path)
        if not model_path.exists():
            raise FileNotFoundError(f"{model_path} does not exists.")

        if not model_path.is_file():
            raise FileExistsError(f"{model_path} is not a file.")


class ONNXRuntimeError(Exception):
    pass


class TableLabelDecode:
    def __init__(self, dict_character, merge_no_span_structure=True, **kwargs):
        if merge_no_span_structure:
            if "<td></td>" not in dict_character:
                dict_character.append("<td></td>")
            if "<td>" in dict_character:
                dict_character.remove("<td>")

        dict_character = self.add_special_char(dict_character)
        self.dict = {}
        for i, char in enumerate(dict_character):
            self.dict[char] = i

View on GitHub (pinned to 4fe4bde114)

Solutions

Point model_path at the .onnx file itself, not its containing directory.
If using a symlink, make sure its target exists.
Standardize on Path(model_dir) / '<model>.onnx' when composing paths.

Example fix

# before
model_path = '/models/table_structure'  # directory

# after
model_path = '/models/table_structure/table_structure_slanetplus.onnx'

Defensive patterns

Strategy: validation

Validate before calling

p = Path(model_path)
if p.exists() and not p.is_file():
    candidates = sorted(p.glob('*.onnx'))
    model_path = candidates[0] if candidates else None

Type guard

def is_model_file_not_dir(p) -> bool:
    return Path(p).is_file() and not Path(p).is_dir()

Try / catch

try:
    eng = TableRecognition(model_path=p)
except FileExistsError as e:
    if 'is not a file' in str(e):
        p = next(Path(p).glob('*.onnx'))  # resolve dir -> onnx file
        eng = TableRecognition(model_path=p)
    else:
        raise

Prevention

Always compose model paths as dir / '<name>.onnx'.
Document expected file layout for manually installed models.
Add a startup assertion listing resolved model paths.

When it happens

Trigger: Passing model_path='/models/slanet_plus' (the extracted directory) instead of '/models/slanet_plus/model.onnx'; passing a broken symlink.

Common situations: Confusion between the model directory and the model file after manual download/extraction, or path-join bugs that omit the filename.

Related errors

AI-assisted analysis of opendatalab/MinerU@4fe4bde114 (2026-08-14). Data as JSON: /api/errors/6bf73196b7860af5. Report an issue: GitHub.