feat(checks): Add MultipleCapitalCheck for issue #8753 (#17265)

Add MultipleCapitalCheck to detect mistakenly used capital letters.

Fixes #8753

Co-authored-by: Michal Čihař <michal@cihar.com>

Author: moonchoe

docs/changes.rst

@@ -5,6 +5,8 @@ Weblate 5.16
 
 .. rubric:: New features
 
+* :ref:`check-multiple-capital` quality check.
+
 .. rubric:: Improvements
 
 * Delete announcements permission can be assigned to teams, see :ref:`privileges`.

docs/user/checks.rst

@@ -1476,6 +1476,25 @@ Specifics on how each plural form is used can be found in the string definition.
 Failing to fill in plural forms will in some cases lead to displaying nothing when
 the plural form is in use.
 
+.. _check-multiple-capital:
+
+Multiple capitals
+~~~~~~~~~~~~~~~~~
+
+.. versionadded:: 5.16
+
+:Summary: Translation contains words with multiple misplaced capital letters.
+:Scope: translated strings
+:Check class: ``weblate.checks.chars.MultipleCapitalCheck``
+:Check identifier: ``multiple_capital``
+:Trigger: This check is always enabled but can be ignored using a flag.
+:Flag to ignore: ``ignore-multiple-capital``
+
+Checks for misplaced capitalization by detecting words that contain consecutive
+uppercase letters in otherwise lowercase or normally capitalized text (for
+example, ``HEllo`` or ``CAmelCase``). Strings that contain capitalization in the
+source string are allowed to contain capitalization in the translation.
+
 .. _check-kabyle-characters:
 
 Non‑standard characters in Kabyle

weblate/checks/chars.py

@@ -8,6 +8,7 @@
 import unicodedata
 from typing import TYPE_CHECKING, ClassVar
 
+import regex
 from django.utils.translation import gettext_lazy
 
 from weblate.checks.base import CountingCheck, TargetCheck, TargetCheckParametrized
@@ -579,3 +580,23 @@ def get_fixup(self, unit: Unit) -> Iterable[FixupType] | None:
                 "gu",
             ),
         ]
+
+
+class MultipleCapitalCheck(TargetCheck):
+    """Multiple capitals check."""
+
+    check_id = "multiple_capital"
+    name = gettext_lazy("Multiple capitals")
+    description = gettext_lazy(
+        "Translation contains words with multiple misplaced capital letters."
+    )
+
+    # matches sequences of 2+ uppercase letters in *any language*
+    UPPERCASE_SEQ = regex.compile(r"\p{Lu}{2,}")
+
+    def check_single(self, source: str, target: str, unit: Unit) -> bool:
+        # Flag if any uppercase sequence is present in target and not present in the source
+        return (
+            self.UPPERCASE_SEQ.search(target) is not None
+            and not self.UPPERCASE_SEQ.search(source) is not None
+        )

weblate/checks/models.py

@@ -65,6 +65,7 @@ class WeblateChecksConf(AppConf):
         "weblate.checks.chars.EndEllipsisCheck",
         "weblate.checks.chars.EndSemicolonCheck",
         "weblate.checks.chars.MaxLengthCheck",
+        "weblate.checks.chars.MultipleCapitalCheck",
         "weblate.checks.chars.KashidaCheck",
         "weblate.checks.chars.PunctuationSpacingCheck",
         "weblate.checks.chars.KabyleCharactersCheck",

weblate/checks/tests/test_chars_checks.py

@@ -23,6 +23,7 @@
     KabyleCharactersCheck,
     KashidaCheck,
     MaxLengthCheck,
+    MultipleCapitalCheck,
     NewLineCountCheck,
     PunctuationSpacingCheck,
     ZeroWidthSpaceCheck,
@@ -512,3 +513,34 @@ def test_skip(self) -> None:
             ),
             "el",
         )
+
+
+class MultipleCapitalCheckTest(CheckTestCase):
+    check = MultipleCapitalCheck()
+
+    def setUp(self) -> None:
+        super().setUp()
+        self.test_good_matching = ("Hello", "Hello", "")
+        self.test_failure_1 = ("Hello", "HEllo", "")
+        self.test_failure_2 = ("camel case", "CAmelCase", "")
+        self.test_failure_3 = ("sigma", "ΣIGMA", "")
+
+    def test_acronyms(self) -> None:
+        self.do_test(
+            False,
+            (
+                "Welcome NATO",
+                "Bonjour OTAN",
+                "",
+            ),
+            "fr",
+        )
+        self.do_test(
+            False,
+            (
+                "Welcome NATO",
+                "Vítej NATO",
+                "",
+            ),
+            "cs",
+        )

weblate/settings_docker.py

@@ -1108,6 +1108,7 @@
     "weblate.checks.chars.EndEllipsisCheck",
     "weblate.checks.chars.EndSemicolonCheck",
     "weblate.checks.chars.MaxLengthCheck",
+    "weblate.checks.chars.MultipleCapitalCheck",
     "weblate.checks.chars.KashidaCheck",
     "weblate.checks.chars.PunctuationSpacingCheck",
     "weblate.checks.chars.KabyleCharactersCheck",

weblate/settings_example.py

@@ -724,6 +724,7 @@
 #     "weblate.checks.chars.EndEllipsisCheck",
 #     "weblate.checks.chars.EndSemicolonCheck",
 #     "weblate.checks.chars.MaxLengthCheck",
+#     "weblate.checks.chars.MultipleCapitalCheck",
 #     "weblate.checks.chars.KashidaCheck",
 #     "weblate.checks.chars.PunctuationSpacingCheck",
 #     "weblate.checks.chars.KabyleCharactersCheck",