opendatalab/MinerU · error · Exception

Unknown process_mode: {process_mode}

Error message

Unknown process_mode: {process_mode}

What it means

Raised by _process_output in mineru/cli/common.py when selecting the markdown 'union_make' function: the process_mode argument matched none of the accepted values. Accepted values are the literal strings 'pipeline' and 'vlm', plus any office suffix ('docx', 'pptx', 'xlsx') from mineru.cli.common.office_suffixes. This is an internal dispatch error, not a user-input error, so it indicates a caller bug or a new backend name that this function was never taught.

Source

Thrown at mineru/cli/common.py:286

        f_dump_orig_pdf,
        f_dump_md,
        f_dump_content_list,
        f_dump_middle_json,
        f_dump_model_output,
        f_make_md_mode,
        middle_json,
        model_output=None,
        process_mode="vlm",
):
    from mineru.backend.pipeline.pipeline_middle_json_mkcontent import union_make as pipeline_union_make
    if process_mode == "pipeline":
        make_func = pipeline_union_make
    elif process_mode == "vlm":
        make_func = vlm_union_make
    elif process_mode in office_suffixes:
        make_func = office_union_make
    else:
        raise Exception(f"Unknown process_mode: {process_mode}")
    """处理输出文件"""
    if f_draw_layout_bbox:
        try:
            draw_layout_bbox(pdf_info, pdf_bytes, local_md_dir, f"{pdf_file_name}_layout.pdf")
        except Exception as exc:
            logger.warning(f"Skipping layout bbox visualization for {pdf_file_name}: {exc}")

    if f_draw_span_bbox:
        try:
            draw_span_bbox(pdf_info, pdf_bytes, local_md_dir, f"{pdf_file_name}_span.pdf")
        except Exception as exc:
            logger.warning(f"Skipping span bbox visualization for {pdf_file_name}: {exc}")

    if f_dump_orig_pdf:
        if process_mode in ["pipeline", "vlm"]:
            md_writer.write(
                f"{pdf_file_name}_origin.pdf",
                pdf_bytes,

View on GitHub (pinned to 4fe4bde114)

Solutions

1. Check the exact value being passed for process_mode at the call site; it must be exactly 'pipeline', 'vlm', or one of docx/pptx/xlsx.
2. If you intended the office path, pass the file suffix ('docx'/'pptx'/'xlsx') as process_mode, not 'office'.
3. If you are using a 'hybrid' or other new backend, update to a mineru version where _process_output handles it, or route around this helper for that backend.
4. Add an explicit whitelist check before calling so an invalid mode fails with your own clearer error.

Example fix

# before
_process_output(pdf_info, bytes, name, md_dir, img_dir, middle_json, process_mode=backend)  # backend='hybrid' -> raises

# after
if backend not in ('pipeline', 'vlm') and backend not in office_suffixes:
    raise ValueError(f"Unsupported backend for markdown generation: {backend}")
_process_output(pdf_info, bytes, name, md_dir, img_dir, middle_json, process_mode=backend)

from mineru.cli.common import office_suffixes
VALID_PROCESS_MODES = {'pipeline', 'vlm', *office_suffixes}

def assert_process_mode(mode: str) -> None:
    if mode not in VALID_PROCESS_MODES:
        raise ValueError(f"process_mode must be one of {sorted(VALID_PROCESS_MODES)}, got {mode!r}")

def is_valid_process_mode(mode: str) -> bool:
    return mode in {'pipeline', 'vlm'} or mode in office_suffixes

When it happens

Trigger: Calling _process_output (or a higher-level CLI path that forwards a backend/process_mode string into it) with process_mode='hybrid', 'txt', 'auto', or any typo like 'VLM' (case-sensitive). Any newly added backend name that is not one of pipeline/vlm/docx/pptx/xlsx hits the else branch.

Common situations: Upgrading mineru to a version that adds a new backend (e.g. hybrid) while an older code path still does the dispatch; passing a user-supplied --backend string straight through without validation; mixing up the values of backend (pipeline/vlm/hybrid) and process_mode (pipeline/vlm/office suffix).

Related errors

• Unknown backend type: {backend}
• Invalid backend. Allowed values: {allowed_values}
• Invalid effort. Allowed values: {allowed_values}
• max_concurrent_requests must be a positive integer
• Unsupported office suffix: {file_suffix}

AI-assisted analysis of opendatalab/MinerU@4fe4bde114 (2026-08-14). Data as JSON: /api/errors/72a90d8c2f502ca9. Report an issue: GitHub.