HKUDS/DeepTutor · error · MinerUError

MinerU archive exceeds the size limit.

Error message

MinerU archive exceeds the size limit.

What it means

Raised by _extract_archive when the cumulative uncompressed size of members extracted from a MinerU-result zip exceeds _MAX_TOTAL_BYTES. It is a defensive guard against zip bombs served by the MinerU cloud API. Extraction is aborted mid-stream.

Source

Thrown at deeptutor/services/parsing/engines/mineru/cloud.py:347

    total = 0
    try:
        with zipfile.ZipFile(io.BytesIO(archive_bytes)) as archive:
            members = [m for m in archive.infolist() if not m.is_dir()]
            if len(members) > _MAX_ENTRIES:
                raise MinerUError(f"MinerU archive has too many entries ({len(members)}).")
            for member in members:
                # Collapse to a POSIX-relative path and reject traversal.
                rel = Path(member.filename.replace("\\", "/"))
                if rel.is_absolute() or ".." in rel.parts:
                    logger.warning("Skipping unsafe zip member: %s", member.filename)
                    continue
                dest = (target_root / rel).resolve()
                if target_root not in dest.parents and dest != target_root:
                    logger.warning("Skipping zip member escaping root: %s", member.filename)
                    continue
                total += member.file_size
                if total > _MAX_TOTAL_BYTES:
                    raise MinerUError("MinerU archive exceeds the size limit.")
                dest.parent.mkdir(parents=True, exist_ok=True)
                with archive.open(member) as src, open(dest, "wb") as out:
                    out.write(src.read())
    except zipfile.BadZipFile as exc:
        raise MinerUError(f"MinerU returned an invalid archive: {exc}") from exc


__all__ = ["parse_cloud", "verify_credentials"]

View on GitHub (pinned to 3e82f13042)

Solutions

1. Increase _MAX_TOTAL_BYTES if your documents legitimately produce larger archives
2. Check the MinerU job output size (e.g. disable image extraction / lower DPI) before downloading
3. Inspect the archive manually (unzip -l) to confirm whether the size is legitimate or a bomb
4. If malicious input is suspected, reject the document and re-run the MinerU job

Example fix

// before
_MAX_TOTAL_BYTES = 256 * 1024 * 1024

// after (if large scans are expected)
_MAX_TOTAL_BYTES = 512 * 1024 * 1024

import zipfile

def archive_total_size_ok(path, limit=_MAX_TOTAL_BYTES) -> bool:
    with zipfile.ZipFile(path) as z:
        return sum(i.file_size for i in z.infolist()) <= limit

try:
    parse_cloud(...)
except MinerUError as e:
    if 'size limit' in str(e): reduce_image_quality_and_retry()
    else: raise

When it happens

Trigger: Calling parse_cloud on a document whose returned archive's summed member.file_size values exceed _MAX_TOTAL_BYTES (accumulated in `total` during iteration); also exercised directly by test_extract_archive_rejects_zip_slip-style tests with oversized fixtures.

Common situations: A malicious or corrupted result archive from the MinerU cloud service (zip bomb), very large scanned documents with many high-resolution images, or a stale _MAX_TOTAL_BYTES constant after a MinerU API change.

Related errors

• Rejected archive '{sanitized_filename}': {exc}
• MinerU archive has too many entries ({len(members)}).
• Rendered {suffix} artifact not found.
• Generated code does not define a renderable Manim Scene clas
• Math animator config must be an object.

AI-assisted analysis of HKUDS/DeepTutor@3e82f13042 (2026-08-27). Data as JSON: /api/errors/da79f5280b72a63c. Report an issue: GitHub.