Replace word boundaries with negative assertions

standardebooks · Jul 25, 2024 · 3089cac · 3089cac
1 parent 788caf1
commit 3089cac
Show file tree

Hide file tree

Showing 3 changed files with 9 additions and 1 deletion.
diff --git a/se/se_epub_lint.py b/se/se_epub_lint.py
@@ -3618,7 +3618,7 @@ def lint(self, skip_lint_ignore: bool, allowed_messages: Optional[List[str]] = N
 						nodes = dom.xpath("/html/body//dd[contains(@epub:type, 'glossdef')]")
 						source_text = " ".join([node.inner_text() for node in nodes])
 					for glossary_index, glossary_value in enumerate(glossary_usage):
-						if glossary_value[1] is False and regex.search(r"\b\L<val>\b", source_text, flags=regex.IGNORECASE, val=[glossary_value[0]]):
+						if glossary_value[1] is False and regex.search(r"(?<!\w)\L<val>(?!\w)", source_text, flags=regex.IGNORECASE, val=[glossary_value[0]]):
 							glossary_usage[glossary_index] = (glossary_value[0], True)
 
 				# Test against word boundaries to not match `halftitlepage`

diff --git a/tests/lint/metadata/m-070/in/src/epub/glossary-search-key-map.xml b/tests/lint/metadata/m-070/in/src/epub/glossary-search-key-map.xml
@@ -6,4 +6,10 @@
 	<search-key-group href="text/glossary.xhtml#bar">
 		<match value="R+L=J"/>
 	</search-key-group>
+	<search-key-group href="text/glossary.xhtml#xyz">
+		<match value="M.O."/>
+	</search-key-group>
+	<search-key-group href="text/glossary.xhtml#abc">
+		<match value="’versal"/>
+	</search-key-group>
 </search-key-map>
diff --git a/tests/lint/metadata/m-070/in/src/epub/text/chapter-1.xhtml b/tests/lint/metadata/m-070/in/src/epub/text/chapter-1.xhtml
@@ -10,6 +10,8 @@
 			<h2 epub:type="ordinal z3998:roman">I</h2>
 			<p>He ate some food.</p>
 			<p>A common theory was R+L=J.</p>
+			<p>A ’versal truth.</p>
+			<p>An unknown M.O.</p>
 		</section>
 	</body>
 </html>