def is_extractable()

in src/datasets/utils/extract.py [0:0]
32 lines of code
12 McCabe index (conditional complexity)

    def is_extractable(cls, path: Union[Path, str], magic_number: bytes = b"") -> bool:
        if super().is_extractable(path, magic_number=magic_number):
            return True
        try:
            # Alternative version of zipfile.is_zipfile that has less false positives, but misses executable zip archives.
            # From: https://github.com/python/cpython/pull/5053
            from zipfile import (
                _CD_SIGNATURE,
                _ECD_DISK_NUMBER,
                _ECD_DISK_START,
                _ECD_ENTRIES_TOTAL,
                _ECD_OFFSET,
                _ECD_SIZE,
                _EndRecData,
                sizeCentralDir,
                stringCentralDir,
                structCentralDir,
            )

            with open(path, "rb") as fp:
                endrec = _EndRecData(fp)
                if endrec:
                    if endrec[_ECD_ENTRIES_TOTAL] == 0 and endrec[_ECD_SIZE] == 0 and endrec[_ECD_OFFSET] == 0:
                        return True  # Empty zipfiles are still zipfiles
                    elif endrec[_ECD_DISK_NUMBER] == endrec[_ECD_DISK_START]:
                        fp.seek(endrec[_ECD_OFFSET])  # Central directory is on the same disk
                        if fp.tell() == endrec[_ECD_OFFSET] and endrec[_ECD_SIZE] >= sizeCentralDir:
                            data = fp.read(sizeCentralDir)  # CD is where we expect it to be
                            if len(data) == sizeCentralDir:
                                centdir = struct.unpack(structCentralDir, data)  # CD is the right size
                                if centdir[_CD_SIGNATURE] == stringCentralDir:
                                    return True  # First central directory entry  has correct magic number
            return False
        except Exception:  # catch all errors in case future python versions change the zipfile internals
            return False