Merge pull request #55 from thombashi/fix_validate_filepath

Fix validation functions of filepaths

.github/workflows/release.yml

@@ -71,7 +71,7 @@ jobs:
           path: ./dist
 
       - name: Sign the dists with Sigstore
-        uses: sigstore/gh-action-sigstore-python@v2.1.1
+        uses: sigstore/gh-action-sigstore-python@v3.0.0
         with:
           inputs: >-
             ./dist/*.tar.gz

examples/pathvalidate_examples.ipynb

@@ -30,14 +30,14 @@
     "from pathvalidate import ValidationError, validate_filename\n",
     "\n",
     "try:\n",
-    "    validate_filename(\"fi:l*e/p\\\"a?t>h|.t<xt\")\n",
+    "    validate_filename('fi:l*e/p\"a?t>h|.t<xt')\n",
     "except ValidationError as e:\n",
     "    print(f\"{e}\\n\", file=sys.stderr)\n",
     "\n",
     "try:\n",
     "    validate_filename(\"COM1\")\n",
     "except ValidationError as e:\n",
-    "    print(f\"{e}\\n\", file=sys.stderr)\n"
+    "    print(f\"{e}\\n\", file=sys.stderr)"
    ]
   },
   {
@@ -81,7 +81,7 @@
     "from pathvalidate import ValidationError, validate_filepath\n",
     "\n",
     "try:\n",
-    "    validate_filepath(\"fi:l*e/p\\\"a?t>h|.t<xt\")\n",
+    "    validate_filepath('fi:l*e/p\"a?t>h|.t<xt')\n",
     "except ValidationError as e:\n",
     "    print(e, file=sys.stderr)"
    ]
@@ -105,7 +105,7 @@
    "source": [
     "from pathvalidate import sanitize_filename\n",
     "\n",
-    "fname = \"fi:l*e/p\\\"a?t>h|.t<xt\"\n",
+    "fname = 'fi:l*e/p\"a?t>h|.t<xt'\n",
     "print(f\"{fname} -> {sanitize_filename(fname)}\\n\")\n",
     "\n",
     "fname = \"\\0_a*b:c<d>e%f/(g)h+i_0.txt\"\n",
@@ -131,7 +131,7 @@
    "source": [
     "from pathvalidate import sanitize_filepath\n",
     "\n",
-    "fpath = \"fi:l*e/p\\\"a?t>h|.t<xt\"\n",
+    "fpath = 'fi:l*e/p\"a?t>h|.t<xt'\n",
     "print(f\"{fpath} -> {sanitize_filepath(fpath)}\\n\")\n",
     "\n",
     "fpath = \"\\0_a*b:c<d>e%f/(g)h+i_0.txt\"\n",
@@ -177,7 +177,7 @@
    "source": [
     "from pathvalidate import is_valid_filename, sanitize_filename\n",
     "\n",
-    "fname = \"fi:l*e/p\\\"a?t>h|.t<xt\"\n",
+    "fname = 'fi:l*e/p\"a?t>h|.t<xt'\n",
     "print(f\"is_valid_filename('{fname}') return {is_valid_filename(fname)}\\n\")\n",
     "\n",
     "sanitized_fname = sanitize_filename(fname)\n",
@@ -203,7 +203,7 @@
    "source": [
     "from pathvalidate import is_valid_filepath, sanitize_filepath\n",
     "\n",
-    "fpath = \"fi:l*e/p\\\"a?t>h|.t<xt\"\n",
+    "fpath = 'fi:l*e/p\"a?t>h|.t<xt'\n",
     "print(f\"is_valid_filepath('{fpath}') return {is_valid_filepath(fpath)}\\n\")\n",
     "\n",
     "sanitized_fpath = sanitize_filepath(fpath)\n",
@@ -229,13 +229,17 @@
    "source": [
     "from pathvalidate import sanitize_filename, ValidationError\n",
     "\n",
+    "\n",
     "def add_trailing_underscore(e: ValidationError) -> str:\n",
     "    if e.reusable_name:\n",
     "        return e.reserved_name\n",
     "\n",
     "    return f\"{e.reserved_name}_\"\n",
     "\n",
-    "sanitize_filename(\".\", reserved_name_handler=add_trailing_underscore, additional_reserved_names=[\".\"])\n"
+    "\n",
+    "sanitize_filename(\n",
+    "    \".\", reserved_name_handler=add_trailing_underscore, additional_reserved_names=[\".\"]\n",
+    ")"
    ]
   }
  ],

pathvalidate/_common.py

@@ -2,9 +2,11 @@
 .. codeauthor:: Tsuyoshi Hombashi <[email protected]>
 """
 
+import ntpath
 import platform
 import re
 import string
+import sys
 from pathlib import PurePath
 from typing import Any, List, Optional
 
@@ -39,6 +41,19 @@ def to_str(name: PathType) -> str:
     return name
 
 
+def is_nt_abspath(value: str) -> bool:
+    ver_info = sys.version_info[:2]
+    if ver_info <= (3, 10):
+        if value.startswith("\\\\"):
+            return True
+    elif ver_info >= (3, 13):
+        return ntpath.isabs(value)
+
+    drive, _tail = ntpath.splitdrive(value)
+
+    return ntpath.isabs(value) and len(drive) > 0
+
+
 def is_null_string(value: Any) -> bool:
     if value is None:
         return True

pathvalidate/_filename.py

@@ -3,15 +3,14 @@
 """
 
 import itertools
-import ntpath
 import posixpath
 import re
 import warnings
 from pathlib import Path, PurePath
 from typing import Optional, Pattern, Sequence, Tuple
 
 from ._base import AbstractSanitizer, AbstractValidator, BaseFile, BaseValidator
-from ._common import findall_to_str, to_str, truncate_str, validate_pathtype
+from ._common import findall_to_str, is_nt_abspath, to_str, truncate_str, validate_pathtype
 from ._const import DEFAULT_MIN_LEN, INVALID_CHAR_ERR_MSG_TMPL, Platform
 from ._types import PathType, PlatformType
 from .error import ErrorAttrKey, ErrorReason, InvalidCharError, ValidationError
@@ -55,7 +54,6 @@ def __init__(
             null_value_handler=null_value_handler,
             reserved_name_handler=reserved_name_handler,
             additional_reserved_names=additional_reserved_names,
-            platform_max_len=_DEFAULT_MAX_FILENAME_LEN,
             platform=platform,
             validate_after_sanitize=validate_after_sanitize,
             validator=fname_validator,
@@ -161,7 +159,6 @@ def __init__(
             fs_encoding=fs_encoding,
             check_reserved=check_reserved,
             additional_reserved_names=additional_reserved_names,
-            platform_max_len=_DEFAULT_MAX_FILENAME_LEN,
             platform=platform,
         )
 
@@ -208,7 +205,7 @@ def validate_abspath(self, value: str) -> None:
         )
 
         if self._is_windows(include_universal=True):
-            if ntpath.isabs(value):
+            if is_nt_abspath(value):
                 raise err
 
         if posixpath.isabs(value):

pathvalidate/_filepath.py

@@ -11,7 +11,7 @@
 from typing import List, Optional, Pattern, Sequence, Tuple
 
 from ._base import AbstractSanitizer, AbstractValidator, BaseFile, BaseValidator
-from ._common import findall_to_str, to_str, validate_pathtype
+from ._common import findall_to_str, is_nt_abspath, to_str, validate_pathtype
 from ._const import _NTFS_RESERVED_FILE_NAMES, DEFAULT_MIN_LEN, INVALID_CHAR_ERR_MSG_TMPL, Platform
 from ._filename import FileNameSanitizer, FileNameValidator
 from ._types import PathType, PlatformType
@@ -178,7 +178,8 @@ def __init__(
 
         self.__fname_validator = FileNameValidator(
             min_len=min_len,
-            max_len=max_len,
+            max_len=self.max_len,
+            fs_encoding=fs_encoding,
             check_reserved=check_reserved,
             additional_reserved_names=additional_reserved_names,
             platform=platform,
@@ -229,7 +230,7 @@ def validate(self, value: PathType) -> None:
             if not entry or entry in (".", ".."):
                 continue
 
-            self.__fname_validator._validate_reserved_keywords(entry)
+            self.__fname_validator.validate(entry)
 
         if self._is_windows(include_universal=True):
             self.__validate_win_filepath(unicode_filepath)
@@ -238,7 +239,18 @@ def validate(self, value: PathType) -> None:
 
     def validate_abspath(self, value: PathType) -> None:
         is_posix_abs = posixpath.isabs(value)
-        is_nt_abs = ntpath.isabs(value)
+        is_nt_abs = is_nt_abspath(to_str(value))
+
+        if any([self._is_windows() and is_nt_abs, self._is_posix() and is_posix_abs]):
+            return
+
+        if self._is_universal() and any([is_nt_abs, is_posix_abs]):
+            ValidationError(
+                "platform-independent absolute file path is not supported",
+                platform=self.platform,
+                reason=ErrorReason.MALFORMED_ABS_PATH,
+            )
+
         err_object = ValidationError(
             description=(
                 "an invalid absolute file path ({}) for the platform ({}).".format(
@@ -251,25 +263,13 @@ def validate_abspath(self, value: PathType) -> None:
             reason=ErrorReason.MALFORMED_ABS_PATH,
         )
 
-        if any([self._is_windows() and is_nt_abs, self._is_linux() and is_posix_abs]):
-            return
-
-        if self._is_universal() and any([is_posix_abs, is_nt_abs]):
-            ValidationError(
-                description=(
-                    ("POSIX style" if is_posix_abs else "NT style")
-                    + " absolute file path found. expected a platform-independent file path."
-                ),
-                platform=self.platform,
-                reason=ErrorReason.MALFORMED_ABS_PATH,
-            )
-
         if self._is_windows(include_universal=True) and is_posix_abs:
             raise err_object
 
-        drive, _tail = ntpath.splitdrive(value)
-        if not self._is_windows() and drive and is_nt_abs:
-            raise err_object
+        if not self._is_windows():
+            drive, _tail = ntpath.splitdrive(value)
+            if drive and is_nt_abs:
+                raise err_object
 
     def __validate_unix_filepath(self, unicode_filepath: str) -> None:
         match = _RE_INVALID_PATH.findall(unicode_filepath)

test/test_filename.py

@@ -397,28 +397,32 @@ def test_normal_additional_reserved_names(self, value, arn, expected):
         assert is_valid_filename(value, additional_reserved_names=arn) == expected
 
     @pytest.mark.parametrize(
-        ["platform", "value", "expected"],
+        ["platform", "value", "expected", "reason"],
         [
-            [win_abspath, platform, None]
+            [win_abspath, platform, None, None]
             for win_abspath, platform in product(
                 ["linux", "macos", "posix"],
                 ["\\", "\\\\", "\\ ", "C:\\", "c:\\", "\\xyz", "\\xyz "],
             )
         ]
         + [
-            [win_abspath, platform, ValidationError]
+            [win_abspath, platform, ValidationError, ErrorReason.FOUND_ABS_PATH]
+            for win_abspath, platform in product(["windows", "universal"], ["\\\\", "C:\\", "c:\\"])
+        ]
+        + [
+            [win_abspath, platform, ValidationError, ErrorReason.INVALID_CHARACTER]
             for win_abspath, platform in product(
-                ["windows", "universal"], ["\\", "\\\\", "\\ ", "C:\\", "c:\\", "\\xyz", "\\xyz "]
+                ["windows", "universal"], ["\\", "\\ ", "\\xyz", "\\xyz "]
             )
         ],
     )
-    def test_win_abs_path(self, platform, value, expected):
+    def test_win_abs_path(self, platform, value, expected, reason):
         if expected is None:
             validate_filename(value, platform=platform)
         else:
             with pytest.raises(expected) as e:
                 validate_filename(value, platform=platform)
-            assert e.value.reason == ErrorReason.FOUND_ABS_PATH
+            assert e.value.reason == reason
 
     @pytest.mark.parametrize(
         ["value", "platform"],
@@ -465,21 +469,6 @@ def test_exception_null_value(self, value, expected):
             validate_filename(value)
         assert not is_valid_filename(value)
 
-    @pytest.mark.parametrize(
-        ["value", "expected"],
-        [
-            ["a" * 256, ValidationError],
-            [1, TypeError],
-            [True, TypeError],
-            [nan, TypeError],
-            [inf, TypeError],
-        ],
-    )
-    def test_exception(self, value, expected):
-        with pytest.raises(expected):
-            validate_filename(value)
-        assert not is_valid_filename(value)
-
 
 class Test_sanitize_filename:
     SANITIZE_CHARS = INVALID_WIN_FILENAME_CHARS + unprintable_ascii_chars

test/test_filepath.py

@@ -289,7 +289,7 @@ def test_minmax_len(self, value, min_len, max_len, expected):
         [
             ["linux", "/a/b/c.txt", None],
             ["linux", "C:\\a\\b\\c.txt", ValidationError],
-            ["windows", "/a/b/c.txt", None],
+            ["windows", "/a/b/c.txt", ValidationError],
             ["windows", "C:\\a\\b\\c.txt", None],
             ["universal", "/a/b/c.txt", ValidationError],
             ["universal", "C:\\a\\b\\c.txt", ValidationError],
@@ -360,24 +360,34 @@ def test_relative_path(self, test_platform, value, expected):
             with pytest.raises(expected):
                 validate_filepath(value, platform=test_platform)
 
+    @pytest.mark.parametrize(
+        ["platform", "value"],
+        [
+            ["linux", "period."],
+            ["linux", "space "],
+            ["linux", "space_and_period. "],
+        ],
+    )
+    def test_normal_space_or_period_at_tail(self, platform, value):
+        validate_filepath(value, platform=platform)
+        assert is_valid_filepath(value, platform=platform)
+
     @pytest.mark.parametrize(
         ["platform", "value"],
         [
             ["windows", "period."],
             ["windows", "space "],
             ["windows", "space_and_period ."],
             ["windows", "space_and_period. "],
-            ["linux", "period."],
-            ["linux", "space "],
-            ["linux", "space_and_period. "],
             ["universal", "period."],
             ["universal", "space "],
             ["universal", "space_and_period ."],
         ],
     )
-    def test_normal_space_or_period_at_tail(self, platform, value):
-        validate_filepath(value, platform=platform)
-        assert is_valid_filepath(value, platform=platform)
+    def test_exception_space_or_period_at_tail(self, platform, value):
+        with pytest.raises(ValidationError):
+            validate_filepath(value, platform=platform)
+        assert not is_valid_filepath(value, platform=platform)
 
     @pytest.mark.skipif(not is_faker_installed(), reason="requires faker")
     @pytest.mark.parametrize(