fix: use in-OTSL DocTags for rich table cells (#375)

Signed-off-by: Panos Vagenas <pva@zurich.ibm.com>

Author: vagenas

docling_core/types/doc/document.py

@@ -351,7 +351,9 @@ def _get_text(self, doc: Optional["DoclingDocument"] = None, **kwargs: Any) -> s
         from docling_core.transforms.serializer.markdown import MarkdownDocSerializer
 
         if doc is not None:
-            doc_serializer = MarkdownDocSerializer(doc=doc)
+            doc_serializer = kwargs.pop(
+                "doc_serializer", MarkdownDocSerializer(doc=doc)
+            )
             ser_res = doc_serializer.serialize(item=self.ref.resolve(doc=doc), **kwargs)
             return ser_res.text
         else:
@@ -1692,6 +1694,9 @@ def export_to_otsl(
         # Headers (column, row, section row):
         # "ched", "rhed", "srow"
 
+        from docling_core.transforms.serializer.doctags import DocTagsDocSerializer
+
+        doc_serializer = DocTagsDocSerializer(doc=doc)
         body = []
         nrows = self.data.num_rows
         ncols = self.data.num_cols
@@ -1705,7 +1710,9 @@ def export_to_otsl(
         for i in range(nrows):
             for j in range(ncols):
                 cell: TableCell = self.data.grid[i][j]
-                content = cell._get_text(doc=doc, **kwargs).strip()
+                content = cell._get_text(
+                    doc=doc, doc_serializer=doc_serializer, **kwargs
+                ).strip()
                 rowspan, rowstart = (
                     cell.row_span,
                     cell.start_row_offset_idx,

examples/rich_table_cells.ipynb

@@ -117,8 +117,9 @@
      "output_type": "stream",
      "text": [
       "<doctag><title>Rich tables</title>\n",
-      "<otsl><fcel>cell 0,0<fcel>cell 0,1<nl><fcel>cell 1,0<fcel>- list item 1\n",
-      "- list item 2<nl><fcel>cell 2,0<fcel>cell 2,1<nl></otsl>\n",
+      "<otsl><fcel>cell 0,0<fcel>cell 0,1<nl><fcel>cell 1,0<fcel><unordered_list><list_item>list item 1</list_item>\n",
+      "<list_item>list item 2</list_item>\n",
+      "</unordered_list><nl><fcel>cell 2,0<fcel>cell 2,1<nl></otsl>\n",
       "</doctag>\n"
      ]
     }
@@ -218,8 +219,9 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "<otsl><fcel>cell 0,0<fcel>cell 0,1<nl><fcel>cell 1,0<fcel>- list item 1\n",
-      "- list item 2<nl><fcel>cell 2,0<fcel>cell 2,1<nl></otsl>\n"
+      "<otsl><fcel>cell 0,0<fcel>cell 0,1<nl><fcel>cell 1,0<fcel><unordered_list><list_item>list item 1</list_item>\n",
+      "<list_item>list item 2</list_item>\n",
+      "</unordered_list><nl><fcel>cell 2,0<fcel>cell 2,1<nl></otsl>\n"
      ]
     }
    ],
@@ -237,8 +239,9 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "<fcel>cell 0,0<fcel>cell 0,1<nl><fcel>cell 1,0<fcel>- list item 1\n",
-      "- list item 2<nl><fcel>cell 2,0<fcel>cell 2,1<nl>\n"
+      "<fcel>cell 0,0<fcel>cell 0,1<nl><fcel>cell 1,0<fcel><unordered_list><list_item>list item 1</list_item>\n",
+      "<list_item>list item 2</list_item>\n",
+      "</unordered_list><nl><fcel>cell 2,0<fcel>cell 2,1<nl>\n"
      ]
     }
    ],

test/data/doc/rich_table.out.dt

@@ -1,6 +1,5 @@
 <doctag><title>Rich tables</title>
-<otsl><fcel>cell 0,0<fcel>cell 0,1<nl><fcel>cell 1,0<fcel>*text in italic*<nl><fcel>- list item 1
-- list item 2<fcel>cell 2,1<nl><fcel>cell 3,0<fcel>| inner cell 0,0   | inner cell 0,1   | inner cell 0,2   |
-|------------------|------------------|------------------|
-| inner cell 1,0   | inner cell 1,1   | inner cell 1,2   |<nl></otsl>
+<otsl><fcel>cell 0,0<fcel>cell 0,1<nl><fcel>cell 1,0<fcel><text>text in italic</text><nl><fcel><unordered_list><list_item>list item 1</list_item>
+<list_item>list item 2</list_item>
+</unordered_list><fcel>cell 2,1<nl><fcel>cell 3,0<fcel><otsl><fcel>inner cell 0,0<fcel>inner cell 0,1<fcel>inner cell 0,2<nl><fcel>inner cell 1,0<fcel>inner cell 1,1<fcel>inner cell 1,2<nl></otsl><nl></otsl>
 </doctag>