Fix markdown parsing for mistral

allenai · Mar 6, 2025 · e144200 · e144200
1 parent bdc0d75
commit e144200
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/olmocr/bench/tests.py b/olmocr/bench/tests.py
@@ -194,7 +194,7 @@ def parse_markdown_tables(self, md_content: str) -> List[np.ndarray]:
             A list of numpy arrays, each representing a parsed table
         """
         # Extract all tables from markdown
-        table_pattern = r'(\|(?:[^|]*\|)+)\s*\n\|(?:[:-]+\|)+\s*\n((?:\|(?:[^|]*\|)+\s*\n)+)'
+        table_pattern = r'(\|(?:[^|]*\|)+)\s*\n\|(?:[ :-]+\|)+\s*\n((?:\|(?:[^|]*\|)+\s*\n)+)'
         table_matches = re.finditer(table_pattern, md_content)
 
         parsed_tables = []