Add markdown syntax check

Fixes #2495 Signed-off-by: Michal Čihař <[email protected]>
WeblateOrg · Feb 7, 2019 · b878711 · b878711
1 parent 33732ec
commit b878711
Show file tree

Hide file tree

Showing 5 changed files with 92 additions and 0 deletions.
diff --git a/docs/user/checks.rst b/docs/user/checks.rst
@@ -462,7 +462,20 @@ Markdown links do not match source.
 
    `Markdown links`_
 
+
+.. _check-md-syntax:
+
+Markdown syntax
+~~~~~~~~~~~~~~~
+
+Markdown syntax does not match source
+
+.. seealso::
+
+   `Markdown span elements`_
+
 .. _Markdown links: https://daringfireball.net/projects/markdown/syntax#link
+.. _Markdown span elements: https://daringfireball.net/projects/markdown/syntax#span
 
 Source checks
 -------------

diff --git a/weblate/checks/markup.py b/weblate/checks/markup.py
@@ -45,6 +45,19 @@
     r'(?:\[[^^\]]*\]|[^\[\]]|\](?=[^\[]*\]))*'
     r')\]\s*\[([^^\]]*)\]'
 )
+MD_SYNTAX = re.compile(
+    r'(_{2})(?:[\s\S]+?)_{2}(?!_)'  # __word__
+    r'|'
+    r'(\*{2})(?:[\s\S]+?)\*{2}(?!\*)'  # **word**
+    r'|'
+    r'\b(_)(?:(?:__|[^_])+?)_\b'  # _word_
+    r'|'
+    r'(\*)(?:(?:\*\*|[^\*])+?)\*(?!\*)'  # *word*
+    r'|'
+    r'(`+)\s*(?:[\s\S]*?[^`])\s*\5(?!`)'  # `code`
+    r'|'
+    r'(~~)(?=\S)(?:[\s\S]*?\S)~~' # ~~word~~
+)
 
 XML_MATCH = re.compile(r'<[^>]+>')
 XML_ENTITY_MATCH = re.compile(r'&#?\w+;')
@@ -230,3 +243,41 @@ def check_single(self, source, target, unit):
         # We don't check actual link targets as those might
         # be localized as well (consider links to Wikipedia)
         return len(src_match) != len(tgt_match)
+
+
+class MarkdownSyntaxCheck(MarkdownBaseCheck):
+    check_id = 'md-syntax'
+    name = _('Markdown syntax')
+    description = _('Markdown syntax does not match source')
+
+    @staticmethod
+    def extract_match(match):
+        for i in range(6):
+            if match[i]:
+                return match[i]
+        return None
+
+    def check_single(self, source, target, unit):
+        src_match = MD_SYNTAX.findall(source)
+        tgt_match = MD_SYNTAX.findall(target)
+
+        src_tags = {self.extract_match(x) for x in src_match}
+        tgt_tags = {self.extract_match(x) for x in tgt_match}
+
+        return src_tags != tgt_tags
+
+    def check_highlight(self, source, unit):
+        if self.should_skip(unit):
+            return []
+        ret = []
+        for match in MD_SYNTAX.finditer(source):
+            value = ''
+            for i in range(6):
+                value = match.group(i + 1)
+                if value:
+                    break
+            start = match.start()
+            end = match.end()
+            ret.append((start, start + len(value), value))
+            ret.append((end - len(value), end, value))
+        return ret
diff --git a/weblate/checks/models.py b/weblate/checks/models.py
@@ -69,6 +69,7 @@ class WeblateChecksConf(AppConf):
         'weblate.checks.markup.XMLTagsCheck',
         'weblate.checks.markup.MarkdownRefLinkCheck',
         'weblate.checks.markup.MarkdownLinkCheck',
+        'weblate.checks.markup.MarkdownSyntaxCheck',
         'weblate.checks.source.OptionalPluralCheck',
         'weblate.checks.source.EllipsisCheck',
         'weblate.checks.source.MultipleFailingCheck',

diff --git a/weblate/checks/tests/test_markup_checks.py b/weblate/checks/tests/test_markup_checks.py
@@ -29,6 +29,7 @@
     XMLValidityCheck,
     MarkdownRefLinkCheck,
     MarkdownLinkCheck,
+    MarkdownSyntaxCheck,
 )
 from weblate.checks.tests.test_checks import CheckTestCase
 
@@ -180,3 +181,28 @@ def setUp(self):
             '[Použij Weblate]',
             'md-text'
         )
+
+
+class MarkdownSyntaxCheckTest(CheckTestCase):
+    check = MarkdownSyntaxCheck()
+
+    def setUp(self):
+        super(MarkdownSyntaxCheckTest, self).setUp()
+        self.test_good_matching = ('**string**', '**string**', 'md-text')
+        self.test_good_none = ('string', 'string', 'md-text')
+        self.test_good_flag = ('**string**', 'string', '')
+        self.test_failure_1 = ('**string**', '*string*', 'md-text')
+        self.test_failure_2 = ('~~string~~', '*string*', 'md-text')
+        self.test_failure_3 = ('_string_', '*string*', 'md-text')
+        self.test_highlight = (
+            'md-text',
+            '**string** ~~strike~~ `code`',
+            [
+                (0, 2, '**'),
+                (8, 10, '**'),
+                (11, 13, '~~'),
+                (19, 21, '~~'),
+                (22, 23, '`'),
+                (27, 28, '`'),
+            ]
+        )
diff --git a/weblate/settings_example.py b/weblate/settings_example.py
@@ -694,6 +694,7 @@
 #     'weblate.checks.markup.XMLTagsCheck',
 #     'weblate.checks.markup.MarkdownRefLinkCheck',
 #     'weblate.checks.markup.MarkdownLinkCheck',
+#     'weblate.checks.markup.MarkdownSyntaxCheck',
 #     'weblate.checks.source.OptionalPluralCheck',
 #     'weblate.checks.source.EllipsisCheck',
 #     'weblate.checks.source.MultipleFailingCheck',