fix(md): keep every line of pipe-led text that has no delimiter row

docling/backend/md_backend.py

@@ -343,6 +343,20 @@ def _append_table_text(self, text: str) -> None:
 
     def _close_table(self, doc: DoclingDocument):
         self.in_pipeless_table = False
+        if self.in_table and not (
+            len(self.md_table_buffer) > 1
+            and MarkdownDocumentBackend._is_delimiter_row(self.md_table_buffer[1])
+        ):
+            # GFM: a table needs a delimiter row right after its header, so
+            # lines that merely start with a pipe are a paragraph. Building a
+            # table here would drop the second line, taken for that row.
+            text = " ".join(
+                unescape(row.strip()) for row in self.md_table_buffer if row.strip()
+            )
+            self.in_table = False
+            self.md_table_buffer = []
+            if text:
+                doc.add_text(label=DocItemLabel.TEXT, text=text)
         if self.in_table:
             _log.debug("=== TABLE START ===")
             for md_table_row in self.md_table_buffer:

tests/test_backend_markdown.py

@@ -412,6 +412,24 @@ def test_convert_pipes_in_prose_stay_text():
         assert conv_result.document.tables == []
 
 
+def test_convert_leading_pipe_lines_without_delimiter_row_keep_all_text():
+    """
+    GFM: a table needs a delimiter row right after its header. Without one,
+    lines that start with a pipe are a paragraph, and the second of them used
+    to be dropped as if it were that delimiter row.
+    """
+    cases = [
+        ("| a | b |\n| 1 | 2 |\n| 3 | 4 |\n", "| a | b | | 1 | 2 | | 3 | 4 |"),
+        ("| quoted note\nsecond line\n", "| quoted note second line"),
+    ]
+
+    for markdown, expected in cases:
+        conv_result = get_converter().convert_string(markdown, format=InputFormat.MD)
+        assert conv_result.status == ConversionStatus.SUCCESS
+        assert conv_result.document.tables == []
+        assert [item.text for item in conv_result.document.texts] == [expected]
+
+
 def test_convert_pipeless_table_does_not_leak_into_later_text():
     """
     Regression guard: