pocketutils.tools.path_tools.PathTools.sanitize_path_node()

Complexity

Conditions

22

Size

Total Lines	128
Code Lines	79

Duplication

Lines	0
Ratio	0 %

Importance

Changes

0

import sys
    @classmethod
    def sanitize_path_node(
        cls,

Wrong hanging indentation before block (add 4 spaces).

        bit: PathLike,

Wrong hanging indentation before block (add 4 spaces).

        is_file: Optional[bool] = None,

Wrong hanging indentation before block (add 4 spaces).

        is_root_or_drive: Optional[bool] = None,

Wrong hanging indentation before block (add 4 spaces).

        include_fat: bool = False,

Wrong hanging indentation before block (add 4 spaces).

    ) -> str:
        r"""
        Sanitizes a path node such that it will be fine for major OSes and filesystems.
        For example:
        - 'plums;and/or;apples' becomes 'plums_and_or_apples' (escaped ; and /)
        - 'null.txt' becomes '_null_.txt' ('null' is forbidden in Windows)
        - 'abc  ' becomes 'abc' (no trailing spaces)
        The behavior is platform-independent -- os, sys, and pathlib are not used.
        For ex, calling sanitize_path_node(r'C:\') returns r'C:\' on both Windows and Linux
        If you want to sanitize a whole path, see sanitize_path instead.

        Args:
            bit: The node
            is_file: False for directories, True otherwise, None if unknown
            is_root_or_drive: True if known to be the root ('/') or a drive ('C:\'), None if unknown
            include_fat: Also make compatible with FAT filesystems

        Returns:
            A string
        """
        # since is_file and is_root_or_drive are both Optional[bool], let's be explicit and use 'is' for clarity

This line is too long as per the coding-style (112/100).

        if is_file is True and is_root_or_drive is True:
            raise ContradictoryRequestError("is_file and is_root_or_drive are both true")
        if is_file is True and is_root_or_drive is None:
            is_root_or_drive = False
        if is_root_or_drive is True and is_file is None:
            is_file = False
        source_bit = copy(str(bit))
        bit = str(bit).strip()
        # first, catch root or drive as long as is_root_or_drive is not false
        # if is_root_or_drive is True (which is a weird call), then fail if it's not
        # otherwise, it's not a root or drive letter, so keep going
        if is_root_or_drive is not False:
            # \ is allowed in Windows
            if bit in ["/", "\\"]:
                return bit
            m = regex.compile(r"^([A-Z]:)(?:\\)?$", flags=regex.V1).fullmatch(bit)

Variable name "m" doesn't conform to snake_case naming style ('([^\\W\\dA-Z][^\\WA-Z]2,|_[^\\WA-Z]*|__[^\\WA-Z\\d_][^\\WA-Z]+__)$' pattern)

            # this is interesting
            # for bit=='C:' and is_root_or_drive=None,
            # it could be either a drive letter
            # or a file path that should be corrected to 'C_'
            # I guess here we're going with a drive letter
            if m is not None:
                # we need C:\ and not C: because:
                # Path('C:\\', '5').is_absolute() is True
                # but Path('C:', '5').is_absolute() is False
                # unfortunately, doing Path('C:\\', '5') on Linux gives 'C:\\/5'
                # I can't handle that here, but sanitize_path() will account for it
                return m.group(1) + "\\"
            if is_root_or_drive is True:
                raise IllegalPathError(f"Node '{bit}' is not the root or a drive letter")
        # note that we can't call WindowsPath.is_reserved because it can't be instantiated on non-Linux

This line is too long as per the coding-style (103/100).

        # also, these appear to be different from the ones defined there
        bad_chars = {
            "<",
            ">",
            ":",
            '"',
            "|",
            "?",
            "*",
            "\\",
            "/",
            *{chr(c) for c in range(128, 128 + 33)},
            *{chr(c) for c in range(0, 32)},
            "\t",
        }
        # don't handle Long UNC paths
        # also cannot be blank or whitespace
        # the $ suffixed ones are for FAT
        # no CLOCK$, even with an ext
        # also no SCREEN$
        bad_strs = {
            "CON",
            "PRN",
            "AUX",
            "NUL",
            "COM1",
            "COM2",
            "COM3",
            "COM4",
            "COM5",
            "COM6",
            "COM7",
            "COM8",
            "COM9",
            "LPT1",
            "LPT2",
            "LPT3",
            "LPT4",
            "LPT5",
            "LPT6",
            "LPT7",
            "LPT8",
            "LPT9",
        }
        if include_fat:
            bad_strs += {"$IDLE$", "CONFIG$", "KEYBD$", "SCREEN$", "CLOCK$", "LST"}
        # just dots is invalid
        if set(bit.replace(" ", "")) == "." and bit not in ["..", "."]:
            raise IllegalPathError(f"Node '{source_bit}' is invalid")
        for q in bad_chars:

Variable name "q" doesn't conform to snake_case naming style ('([^\\W\\dA-Z][^\\WA-Z]2,|_[^\\WA-Z]*|__[^\\WA-Z\\d_][^\\WA-Z]+__)$' pattern)

            bit = bit.replace(q, "_")
        if bit.upper() in bad_strs:
            # arbitrary decision
            bit = "_" + bit + "_"
        else:
            stub, ext = os.path.splitext(bit)
            if stub.upper() in bad_strs:
                bit = "_" + stub + "_" + ext
        if bit.strip() == "":
            raise IllegalPathError(f"Node '{source_bit}' is empty or contains only whitespace")
        # do this after
        if len(bit) > 254:
            raise IllegalPathError(f"Node '{source_bit}' has more than 254 characters")
        bit = bit.strip()
        if is_file is not True and (bit == "." or bit == ".."):

Consider merging these comparisons with "in" to "bit in ('.', '..')"

            return bit
        # never allow '.' (or ' ') to end a filename
        bit = bit.rstrip(".")
        return bit