wikitextparser 0.56.2__tar.gz → 0.56.3__tar.gz

wikitextparser-0.56.3/.vscode/settings.json

@@ -0,0 +1,7 @@
+{
+    "python.testing.pytestArgs": [
+        "tests"
+    ],
+    "python.testing.unittestEnabled": false,
+    "python.testing.pytestEnabled": true
+}

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/CHANGELOG.rst

@@ -1,3 +1,8 @@
+v0.56.3 (2024-10-18)
+--------------------
+* Fixed a bug in detecting HTML tags nested in wiki markup. (#140)
+* Improved type hints.
+
 v0.56.2
 -------
 * Fixed a bug in ``external_links`` property where ``|`` was recognized as part of the link by mistake. (#139)

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: wikitextparser
-Version: 0.56.2
+Version: 0.56.3
 Summary: A simple parsing tool for MediaWiki's wikitext markup.
 Keywords: MediaWiki,wikitext,parser
 Author-email: 5j9 <5j9@users.noreply.github.com>

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/pyproject.toml

@@ -50,6 +50,7 @@ line-length = 79
 format.quote-style = 'single'
 lint.isort.combine-as-imports = true
 lint.extend-select = [
+    'FA',  # flake8-future-annotations
     'I',  # isort
     'UP',  # pyupgrade
 ]
@@ -60,3 +61,12 @@ lint.ignore = [
 
 [tool.pytest.ini_options]
 addopts = '--quiet --tb=short'
+
+[tool.pyright]
+typeCheckingMode = 'standard'
+reportDeprecated = "warning"
+reportPropertyTypeMismatch = "warning"
+reportUnnecessaryCast = "warning"
+reportUnnecessaryContains = "warning"
+reportUnnecessaryIsInstance = "warning"
+reportUnnecessaryTypeIgnoreComment = true

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/wikitextparser/__init__.py

@@ -1,5 +1,5 @@
 # Scheme: [N!]N(.N)*[{a|b|rc}N][.postN][.devN]
-__version__ = '0.56.2'
+__version__ = '0.56.3'
 
 from . import _wikitext
 from ._argument import Argument  # noqa: F401

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/wikitextparser/_argument.py

@@ -1,7 +1,10 @@
-from typing import Dict, List, MutableSequence, Optional, Union
+from __future__ import annotations
 
-from regex import DOTALL, MULTILINE
+from typing import MutableSequence
 
+from regex import DOTALL, MULTILINE, Match
+
+from ._spans import TypeToSpans
 from ._wikitext import SECTION_HEADING, SubWikiText, rc
 
 ARG_SHADOW_FULLMATCH = rc(
@@ -25,28 +28,28 @@ class Argument(SubWikiText):
 
     def __init__(
         self,
-        string: Union[str, MutableSequence[str]],
-        _type_to_spans: Optional[Dict[str, List[List[int]]]] = None,
-        _span: Optional[List[int]] = None,
-        _type: Optional[Union[str, int]] = None,
-        _parent: 'SubWikiTextWithArgs' = None,
+        string: str | MutableSequence[str],
+        _type_to_spans: TypeToSpans | None = None,
+        _span: list[int] | None = None,
+        _type: str | int | None = None,
+        _parent: SubWikiTextWithArgs | None = None,
     ):
         super().__init__(string, _type_to_spans, _span, _type)
         self._parent = _parent or self
         self._shadow_match_cache = None, None
 
     @property
-    def _shadow_match(self):
+    def _shadow_match(self) -> Match[bytes]:
         cached_shadow_match, cache_string = self._shadow_match_cache
         self_string = str(self)
         if cache_string == self_string:
-            return cached_shadow_match
+            return cached_shadow_match  # type: ignore
         ss, se, _, _ = self._span_data
         parent = self._parent
         ps = parent._span_data[0]
         shadow_match = ARG_SHADOW_FULLMATCH(parent._shadow[ss - ps : se - ps])
         self._shadow_match_cache = shadow_match, self_string
-        return shadow_match
+        return shadow_match  # type: ignore
 
     @property
     def name(self) -> str:

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/wikitextparser/_cell.py

@@ -1,8 +1,10 @@
-from typing import Dict, List, Match, MutableSequence, Union
+from __future__ import annotations
 
-from regex import DOTALL, VERBOSE
+from typing import MutableSequence
 
-from ._spans import ATTRS_MATCH
+from regex import DOTALL, VERBOSE, Match
+
+from ._spans import ATTRS_MATCH, TypeToSpans
 from ._tag import SubWikiTextWithAttrs
 from ._wikitext import rc
 
@@ -144,15 +146,14 @@ class Cell(SubWikiTextWithAttrs):
 
     def __init__(
         self,
-        string: Union[str, MutableSequence[str]],
+        string: str | MutableSequence[str],
         header: bool = False,
-        _type_to_spans: Dict[str, List[List[int]]] = None,
-        _span: int = None,
-        _type: int = None,
-        _match: Match = None,
-        _attrs_match: Match = None,
+        _type_to_spans: TypeToSpans | None = None,
+        _span: list | None = None,
+        _type: int | None = None,
+        _match: Match | None = None,
+        _attrs_match: Match | None = None,
     ) -> None:
-        """Initialize the object."""
         super().__init__(string, _type_to_spans, _span, _type)
         self._header = header
         if _match:
@@ -175,7 +176,7 @@ class Cell(SubWikiTextWithAttrs):
             self._attrs_match_cache = self._match_cache = None, None
 
     @property
-    def _match(self):
+    def _match(self) -> Match[bytes]:
         """Return the match object for the current tag. Cache the result.
 
         Be extra careful when using this property. The position of match
@@ -185,7 +186,7 @@ class Cell(SubWikiTextWithAttrs):
         cache_match, cache_string = self._match_cache
         string = self.string
         if cache_string == string:
-            return cache_match
+            return cache_match  # type: ignore
         shadow = self._shadow
         if shadow[0] == 10:  # ord('\n')
             m = NEWLINE_CELL_MATCH(shadow)
@@ -196,7 +197,7 @@ class Cell(SubWikiTextWithAttrs):
             m = INLINE_NONHAEDER_CELL_MATCH(shadow)
         self._match_cache = m, string
         self._attrs_match_cache = None, None
-        return m
+        return m  # type: ignore
 
     @property
     def value(self) -> str:

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/wikitextparser/_comment_bold_italic.py

@@ -1,7 +1,10 @@
-from typing import Dict, List, MutableSequence, Optional, Tuple, Union
+from __future__ import annotations
 
-from regex import DOTALL, MULTILINE
+from typing import MutableSequence
 
+from regex import DOTALL, MULTILINE, Match
+
+from ._spans import TypeToSpans
 from ._wikitext import SubWikiText, rc
 
 COMMENT_PATTERN = r'<!--[\s\S]*?(?>-->|\Z)'
@@ -29,13 +32,16 @@ class Comment(SubWikiText):
         return s[4:]
 
     @property
-    def comments(self) -> List['Comment']:
+    def comments(self) -> list[Comment]:
         return []
 
 
 class BoldItalic(SubWikiText):
     __slots__ = ()
 
+    @property
+    def _match(self) -> Match[str]: ...
+
     @property
     def text(self) -> str:
         """Return text value of self (without triple quotes)."""
@@ -49,7 +55,7 @@ class BoldItalic(SubWikiText):
         self[b:e] = s
 
     @property
-    def _content_span(self) -> Tuple[int, int]:
+    def _content_span(self) -> tuple[int, int]:
         # noinspection PyUnresolvedReferences
         return self._match.span(1)
 
@@ -58,8 +64,8 @@ class Bold(BoldItalic):
     __slots__ = ()
 
     @property
-    def _match(self):
-        return BOLD_FULLMATCH(self.string)
+    def _match(self) -> Match[str]:
+        return BOLD_FULLMATCH(self.string)  # type: ignore
 
 
 class Italic(BoldItalic):
@@ -67,10 +73,10 @@ class Italic(BoldItalic):
 
     def __init__(
         self,
-        string: Union[str, MutableSequence[str]],
-        _type_to_spans: Optional[Dict[str, List[List[int]]]] = None,
-        _span: Optional[List[int]] = None,
-        _type: Optional[Union[str, int]] = None,
+        string: str | MutableSequence[str],
+        _type_to_spans: TypeToSpans | None = None,
+        _span: list[int] | None = None,
+        _type: str | int | None = None,
         end_token: bool = True,
     ):
         """Initialize the Italic object.
@@ -82,7 +88,7 @@ class Italic(BoldItalic):
         self.end_token = end_token
 
     @property
-    def _match(self):
+    def _match(self) -> Match[str]:
         if self.end_token:
-            return ITALIC_FULLMATCH(self.string)
-        return ITALIC_NOEND_FULLMATCH(self.string)
+            return ITALIC_FULLMATCH(self.string)  # type: ignore
+        return ITALIC_NOEND_FULLMATCH(self.string)  # type: ignore

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/wikitextparser/_externallink.py

@@ -1,4 +1,4 @@
-from typing import List, Optional
+from __future__ import annotations
 
 from ._wikitext import BRACKET_EXTERNAL_LINK_URL, IGNORECASE, SubWikiText, rc
 
@@ -28,7 +28,7 @@ class ExternalLink(SubWikiText):
             self[0 : len(self.url)] = newurl
 
     @property
-    def text(self) -> Optional[str]:
+    def text(self) -> str | None:
         """The text part (the part after the url).
 
         getter: Return None if this is a bare link or has no associated text.
@@ -74,5 +74,5 @@ class ExternalLink(SubWikiText):
         return self(0) == '['
 
     @property
-    def external_links(self) -> List['ExternalLink']:
+    def external_links(self) -> list[ExternalLink]:
         return []

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/wikitextparser/_parameter.py

@@ -1,4 +1,4 @@
-from typing import List, Optional, Tuple
+from __future__ import annotations
 
 from ._wikitext import WS, SubWikiText
 
@@ -35,7 +35,7 @@ class Parameter(SubWikiText):
         return '|' if self._shadow.find(124) != -1 else ''
 
     @property
-    def default(self) -> Optional[str]:
+    def default(self) -> str | None:
         """The default value of current parameter.
 
         getter: Return None if there is no default.
@@ -100,9 +100,9 @@ class Parameter(SubWikiText):
             ] = '{{{' + new_default_name + '|' + innermost_default + '}}}'
 
     @property
-    def parameters(self) -> List['Parameter']:
+    def parameters(self) -> list[Parameter]:
         return super().parameters[1:]
 
     @property
-    def _content_span(self) -> Tuple[int, int]:
+    def _content_span(self) -> tuple[int, int]:
         return 3, -3

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/wikitextparser/_parser_function.py

@@ -1,5 +1,7 @@
+from __future__ import annotations
+
 from bisect import insort
-from typing import Iterable, List, Tuple, Union
+from typing import Iterable
 
 from ._argument import Argument
 from ._wikilist import WikiList
@@ -19,7 +21,7 @@ class SubWikiTextWithArgs(SubWikiText):
     _first_arg_sep = 0
 
     @property
-    def _content_span(self) -> Tuple[int, int]:
+    def _content_span(self) -> tuple[int, int]:
         return 2, -2
 
     @property
@@ -32,7 +34,7 @@ class SubWikiTextWithArgs(SubWikiText):
         return self._nesting_level(('Template', 'ParserFunction'))
 
     @property
-    def arguments(self) -> List[Argument]:
+    def arguments(self) -> list[Argument]:
         """Parse template content. Create self.name and self.arguments."""
         shadow = self._shadow
         split_spans = self._name_args_matcher(shadow, 2, -2).spans('arg')
@@ -65,8 +67,8 @@ class SubWikiTextWithArgs(SubWikiText):
         return arguments
 
     def get_lists(
-        self, pattern: Union[str, Iterable[str]] = (r'\#', r'\*', '[:;]')
-    ) -> List[WikiList]:
+        self, pattern: str | Iterable[str] = (r'\#', r'\*', '[:;]')
+    ) -> list[WikiList]:
         """Return the lists in all arguments.
 
         For performance reasons it is usually preferred to get a specific
@@ -103,5 +105,5 @@ class ParserFunction(SubWikiTextWithArgs):
     _first_arg_sep = 58
 
     @property
-    def parser_functions(self) -> List['ParserFunction']:
+    def parser_functions(self) -> list[ParserFunction]:
         return super().parser_functions[1:]

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/wikitextparser/_section.py

@@ -1,4 +1,4 @@
-from typing import Optional
+from __future__ import annotations
 
 from ._wikitext import SubWikiText, rc
 
@@ -50,7 +50,7 @@ class Section(SubWikiText):
         del self[m.end(2) : m.end(2) + level_diff]
 
     @property
-    def title(self) -> Optional[str]:
+    def title(self) -> str | None:
         """The title of this section.
 
         getter: Return the title or None for lead sections or sections that

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/wikitextparser/_spans.py

@@ -1,7 +1,9 @@
 """Define the functions required for parsing wikitext into spans."""
 
+from __future__ import annotations
+
 from functools import partial
-from typing import Callable, Dict, Optional
+from typing import Callable, Dict, List, Union
 
 from regex import DOTALL, IGNORECASE, REVERSE, Match, compile as rc
 
@@ -210,7 +212,11 @@ HTML_END_TAG_FINDITER = rc(
 ).finditer
 
 
-def parse_to_spans(byte_array: bytearray) -> Dict[str, list]:
+# [stan_start: int, span_end: int, Match, byte_array]
+TypeToSpans = Dict[Union[str, int], List[List]]
+
+
+def parse_to_spans(byte_array: bytearray) -> TypeToSpans:
     """Calculate and set self._type_to_spans.
 
     Extracted spans will be removed from byte_array.
@@ -321,7 +327,7 @@ def extract_tag_extensions(
 def _parse_sub_spans(
     byte_array: bytearray,
     start: int,
-    end: Optional[int],
+    end: int | None,
     pms_append: Callable,
     pfs_append: Callable,
     tls_append: Callable,
@@ -336,7 +342,7 @@ def _parse_sub_spans(
         byte_array[ms:me] = byte_array[ms:me].translate(BRACKETS)
     while True:
         while True:
-            match: Optional[Match] = None
+            match: Match | None = None
             for match in WIKILINK_PARAM_FINDITER(byte_array, start, end):
                 ms, me = match.span()
                 if match[1] is None:

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/wikitextparser/_table.py

@@ -1,6 +1,8 @@
+from __future__ import annotations
+
 from bisect import insort_right
 from collections.abc import Mapping
-from typing import Any, Dict, List, Optional, Tuple, TypeVar, Union
+from typing import Any, TypeVar
 
 from regex import DOTALL, VERBOSE
 
@@ -86,7 +88,7 @@ class Table(SubWikiTextWithAttrs):
         return shadow
 
     @property
-    def _match_table(self) -> List[List[Any]]:
+    def _match_table(self) -> list[list[Any]]:
         """Return match_table."""
         table_shadow = self._table_shadow
         # Remove table-start and table-end marks.
@@ -141,7 +143,7 @@ class Table(SubWikiTextWithAttrs):
         strip: bool = True,
         row: int = None,
         column: int = None,
-    ) -> Union[List[List[str]], List[str], str]:
+    ) -> list[list[str]] | list[str] | str:
         """Return a list containing lists of row values.
 
         :param span: If true, calculate rows according to rowspans and colspans
@@ -209,7 +211,7 @@ class Table(SubWikiTextWithAttrs):
         row: int = None,
         column: int = None,
         span: bool = True,
-    ) -> Union[List[List[Cell]], List[Cell], Cell]:
+    ) -> list[list[Cell]] | list[Cell] | Cell:
         """Return a list of lists containing Cell objects.
 
         :param span: If is True, rearrange the result according to colspan and
@@ -283,7 +285,7 @@ class Table(SubWikiTextWithAttrs):
         return table_cells[row][column]
 
     @property
-    def caption(self) -> Optional[str]:
+    def caption(self) -> str | None:
         """Caption of the table. Support get and set."""
         m = CAPTION_MATCH(self._shadow)
         if m:
@@ -317,7 +319,7 @@ class Table(SubWikiTextWithAttrs):
         return attrs_match
 
     @property
-    def caption_attrs(self) -> Optional[str]:
+    def caption_attrs(self) -> str | None:
         """Caption attributes. Support get and set operations."""
         m = CAPTION_MATCH(self._shadow)
         if m:
@@ -339,7 +341,7 @@ class Table(SubWikiTextWithAttrs):
                 self[m.end('preattrs') : end] = attrs
 
     @property
-    def row_attrs(self) -> List[dict]:
+    def row_attrs(self) -> list[dict]:
         """Row attributes.
 
         Use the setter of this property to set attributes for all rows.
@@ -363,7 +365,7 @@ class Table(SubWikiTextWithAttrs):
         return attrs
 
     @row_attrs.setter
-    def row_attrs(self, attrs: List[Mapping]):
+    def row_attrs(self, attrs: list[Mapping]):
         for row_match, attrs_dict in reversed(
             [*zip(FIND_ROWS(self._table_shadow), attrs)]
         ):
@@ -381,8 +383,8 @@ class Table(SubWikiTextWithAttrs):
 
 
 def _apply_attr_spans(
-    table_attrs: List[List[Dict[str, str]]], table_data: List[List[T]]
-) -> List[List[T]]:
+    table_attrs: list[list[dict[str, str]]], table_data: list[list[T]]
+) -> list[list[T]]:
     """Apply row and column spans and return table_data."""
     # The following code is based on the table forming algorithm described
     # at http://www.w3.org/TR/html5/tabular-data.html#processing-model-1
@@ -399,7 +401,7 @@ def _apply_attr_spans(
     # if not table_data:
     #     return table_data
     # 11
-    downward_growing_cells: List[Tuple[Optional[T], int, int]] = []
+    downward_growing_cells: list[tuple[T | None, int, int]] = []
     # 13, 18
     # Algorithm for processing rows
     for attrs_row, row in zip(table_attrs, table_data):

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/wikitextparser/_tag.py

@@ -8,7 +8,9 @@ For more info see:
 * https://www.mediawiki.org/wiki/HTML_restriction
 """
 
-from typing import Any, Dict, List, Optional, Tuple
+from __future__ import annotations
+
+from typing import Any
 
 from regex import DOTALL, VERBOSE
 
@@ -55,7 +57,7 @@ class SubWikiTextWithAttrs(SubWikiText):
     __slots__ = '_attrs_match'
 
     @property
-    def attrs(self) -> Dict[str, str]:
+    def attrs(self) -> dict[str, str]:
         """Return self attributes as a dictionary."""
         spans = self._attrs_match.spans
         string = self.string
@@ -73,7 +75,7 @@ class SubWikiTextWithAttrs(SubWikiText):
             string[s:e] for s, e in self._attrs_match.spans('attr_name')
         )
 
-    def get_attr(self, attr_name: str) -> Optional[str]:
+    def get_attr(self, attr_name: str) -> str | None:
         """Return the value of the last attribute with the given name.
 
         Return None if the attr_name does not exist in self.
@@ -163,7 +165,7 @@ class Tag(SubWikiTextWithAttrs):
         self[start:end] = name
 
     @property
-    def contents(self) -> Optional[str]:
+    def contents(self) -> str | None:
         """Tag contents. Support both get and set operations.
 
         setter:
@@ -210,10 +212,10 @@ class Tag(SubWikiTextWithAttrs):
     def _extension_tags(self):
         return super()._extension_tags[1:]
 
-    def get_tags(self, name=None) -> List['Tag']:
+    def get_tags(self, name=None) -> list[Tag]:
         return super().get_tags(name)[1:]
 
     @property
-    def _content_span(self) -> Tuple[int, int]:
+    def _content_span(self) -> tuple[int, int]:
         s = self.string
         return s.find('>') + 1, s.rfind('<')

{wikitextparser-0.56.2 → wikitextparser-0.56.3}/wikitextparser/_template.py

@@ -1,4 +1,6 @@
-from typing import Dict, Iterable, List, Optional, Tuple, TypeVar
+from __future__ import annotations
+
+from typing import Iterable, TypeVar
 
 from regex import REVERSE
 
@@ -29,14 +31,14 @@ class Template(SubWikiTextWithArgs):
     _first_arg_sep = 124
 
     @property
-    def _content_span(self) -> Tuple[int, int]:
+    def _content_span(self) -> tuple[int, int]:
         return 2, -2
 
     def normal_name(
         self,
         rm_namespaces=('Template',),
         *,
-        code: str = None,
+        code: str | None = None,
         capitalize=False,
     ) -> str:
         """Return normal form of self.name.
@@ -110,7 +112,7 @@ class Template(SubWikiTextWithArgs):
             else:
                 names.add(name)
 
-    def rm_dup_args_safe(self, tag: str = None) -> None:
+    def rm_dup_args_safe(self, tag: str | None = None) -> None:
         """Remove duplicate arguments in a safe manner.
 
         Remove the duplicate arguments only in the following situations:
@@ -129,7 +131,7 @@ class Template(SubWikiTextWithArgs):
 
         Also see `rm_first_of_dup_args` function.
         """
-        name_to_lastarg_vals: Dict[str, Tuple[Argument, List[str]]] = {}
+        name_to_lastarg_vals: dict[str, tuple[Argument, list[str]]] = {}
         # Removing positional args affects their name. By reversing the list
         # we avoid encountering those kind of args.
         for arg in reversed(self.arguments):
@@ -171,9 +173,9 @@ class Template(SubWikiTextWithArgs):
         self,
         name: str,
         value: str,
-        positional: bool = None,
-        before: str = None,
-        after: str = None,
+        positional: bool | None = None,
+        before: str | None = None,
+        after: str | None = None,
         preserve_spacing=False,
     ) -> None:
         """Set the value for `name` argument. Add it if it doesn't exist.
@@ -272,14 +274,14 @@ class Template(SubWikiTextWithArgs):
                 # positional AND is to be added at the end of the template.
                 self.insert(-2, addstring)
 
-    def get_arg(self, name: str) -> Optional[Argument]:
+    def get_arg(self, name: str) -> Argument | None:
         """Return the last argument with the given name.
 
         Return None if no argument with that name is found.
         """
         return get_arg(name, reversed(self.arguments))
 
-    def has_arg(self, name: str, value: str = None) -> bool:
+    def has_arg(self, name: str, value: str | None = None) -> bool:
         """Return true if the is an arg named `name`.
 
         Also check equality of values if `value` is provided.
@@ -308,11 +310,11 @@ class Template(SubWikiTextWithArgs):
                 del arg[:]
 
     @property
-    def templates(self) -> List['Template']:
+    def templates(self) -> list[Template]:
         return super().templates[1:]
 
 
-def mode(list_: List[T]) -> T:
+def mode(list_: list[T]) -> T:
     """Return the most common item in the list.
 
     Return the first one if there are more than one most common items.
@@ -330,7 +332,7 @@ def mode(list_: List[T]) -> T:
     return max(set(list_), key=list_.count)
 
 
-def get_arg(name: str, args: Iterable[Argument]) -> Optional[Argument]:
+def get_arg(name: str, args: Iterable[Argument]) -> Argument | None:
     """Return the first argument in the args that has the given name.
 
     Return None if no such argument is found.