Strip malformed grounding tags from markdown output

app.py

@@ -247,9 +247,27 @@ def clean_output(text, include_images=False):
                 text = text.replace(match[0], '', 1)
         else:
             text = re.sub(rf'(?m)^[^\n]*{re.escape(match[0])}[^\n]*\n?', '', text)
-    
+
+    text = _strip_malformed_grounding(text)
     return text.strip()
 
+def _strip_malformed_grounding(text: str) -> str:
+    """Remove incomplete grounding tags that can leak into OCR markdown/text."""
+    if not text:
+        return ""
+
+    line_patterns = [
+        r'(?m)^[^\n]*<\|ref\|>.*?<\|/ref\|><\|det\|>.*?(?:<\|/det\|>)?[^\n]*\n?',
+        r'(?m)^[^\n]*<\|det\|>.*?(?:<\|/det\|>)?[^\n]*\n?',
+        r'(?m)^[^\n]*<\|/?ref\|>[^\n]*\n?',
+    ]
+    for p in line_patterns:
+        text = re.sub(p, '', text)
+
+    text = re.sub(r'<\|/?ref\|>', '', text)
+    text = re.sub(r'<\|/?det\|>', '', text)
+    return text
+
 PREVIEW_CSS = """
 <style>
 .math-preview {

tests/test_clean_output.py

@@ -0,0 +1,54 @@
+import ast
+import pathlib
+import re
+import unittest
+
+
+def _load_clean_output():
+    app_path = pathlib.Path(__file__).resolve().parents[1] / "app.py"
+    source = app_path.read_text(encoding="utf-8")
+    module = ast.parse(source, filename=str(app_path))
+
+    wanted = {
+        "_strip_malformed_grounding",
+        "clean_output",
+    }
+    fn_nodes = [n for n in module.body if isinstance(n, ast.FunctionDef) and n.name in wanted]
+    fn_nodes.sort(key=lambda n: n.lineno)
+
+    test_mod = ast.Module(body=fn_nodes, type_ignores=[])
+    code = compile(test_mod, filename=str(app_path), mode="exec")
+
+    scope = {
+        "re": re,
+    }
+    exec(code, scope)
+    return scope["clean_output"]
+
+
+class CleanOutputTests(unittest.TestCase):
+    def test_removes_truncated_grounding_artifact_line(self):
+        clean_output = _load_clean_output()
+        raw = (
+            "\\[ \\frac{18x-34}{(2x-3)^2} \\]\n"
+            "<|ref|>equation<|/ref|><|det|>[[50, 0, 450, 100]]\n"
+        )
+
+        cleaned = clean_output(raw, include_images=True)
+        self.assertIn("\\[ \\frac{18x-34}{(2x-3)^2} \\]", cleaned)
+        self.assertNotIn("<|ref|>", cleaned)
+        self.assertNotIn("<|det|>", cleaned)
+        self.assertNotIn("[[50, 0, 450, 100]]", cleaned)
+
+    def test_replaces_full_image_reference(self):
+        clean_output = _load_clean_output()
+        raw = "prefix\n<|ref|>image<|/ref|><|det|>[[0,0,100,100]]<|/det|>\nsuffix"
+
+        cleaned = clean_output(raw, include_images=True)
+        self.assertIn("**[Figure 1]**", cleaned)
+        self.assertNotIn("<|ref|>", cleaned)
+        self.assertNotIn("<|det|>", cleaned)
+
+
+if __name__ == "__main__":
+    unittest.main()