quality: Unescape HTML entities in Document Intelligence

microsoft · Jun 15, 2024 · 28194a4 · 28194a4
1 parent fabe09c
commit 28194a4
Showing 1 changed file with 3 additions and 2 deletions.
diff --git a/persistence/document_intelligence.py b/persistence/document_intelligence.py
@@ -13,6 +13,7 @@
     ParagraphRole,
 )
 import math
+import html
 
 
 class DocumentIntelligenceAnalyze(IAnalyze):
@@ -62,8 +63,8 @@ async def analyze(
                     None,  # Third, nothing
                 ),
             )
-            content = res.content
-            title = title_paragraph.content if title_paragraph else None
+            content = html.unescape(res.content)
+            title = html.unescape(title_paragraph.content) if title_paragraph else None
             langs = {lang.locale for lang in res.languages or [] if lang.confidence >= self._config.extract_lang_threshold}
 
         # Return title, content and langs