docling 2.55.0__tar.gz → 2.55.1__tar.gz

{docling-2.55.0 → docling-2.55.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling
-Version: 2.55.0
+Version: 2.55.1
 Summary: SDK and CLI for parsing PDF, DOCX, HTML, and more, to a unified document representation for powering downstream workflows such as gen AI applications.
 Author-email: Christoph Auer <cau@zurich.ibm.com>, Michele Dolfi <dol@zurich.ibm.com>, Maxim Lysak <mly@zurich.ibm.com>, Nikos Livathinos <nli@zurich.ibm.com>, Ahmed Nassar <ahn@zurich.ibm.com>, Panos Vagenas <pva@zurich.ibm.com>, Peter Staar <taa@zurich.ibm.com>
 License-Expression: MIT

{docling-2.55.0 → docling-2.55.1}/docling/backend/md_backend.py

@@ -249,7 +249,10 @@ class MarkdownDocumentBackend(DeclarativeDocumentBackend):
 
         # Iterates over all elements in the AST
         # Check for different element types and process relevant details
-        if isinstance(element, marko.block.Heading) and len(element.children) > 0:
+        if (
+            isinstance(element, marko.block.Heading)
+            or isinstance(element, marko.block.SetextHeading)
+        ) and len(element.children) > 0:
             self._close_table(doc)
             _log.debug(
                 f" - Heading level {element.level}, content: {element.children[0].children}"  # type: ignore

{docling-2.55.0 → docling-2.55.1}/docling/cli/main.py

@@ -355,6 +355,13 @@ def convert(  # noqa: C901
             help="Replace any existing text with OCR generated text over the full content.",
         ),
     ] = False,
+    tables: Annotated[
+        bool,
+        typer.Option(
+            ...,
+            help="If enabled, the table structure model will be used to extract table information.",
+        ),
+    ] = True,
     ocr_engine: Annotated[
         str,
         typer.Option(
@@ -591,7 +598,7 @@ def convert(  # noqa: C901
                 accelerator_options=accelerator_options,
                 do_ocr=ocr,
                 ocr_options=ocr_options,
-                do_table_structure=True,
+                do_table_structure=tables,
                 do_code_enrichment=enrich_code,
                 do_formula_enrichment=enrich_formula,
                 do_picture_description=enrich_picture_description,

{docling-2.55.0 → docling-2.55.1}/docling/models/readingorder_model.py

@@ -9,6 +9,7 @@ from docling_core.types.doc import (
     NodeItem,
     ProvenanceItem,
     RefItem,
+    RichTableCell,
     TableData,
 )
 from docling_core.types.doc.document import ContentLayer
@@ -103,6 +104,22 @@ class ReadingOrderModel:
             else:
                 doc.add_text(parent=doc_item, label=c_label, text=c_text, prov=c_prov)
 
+    def _create_rich_cell_group(
+        self, element: BasePageElement, doc: DoclingDocument, table_item: NodeItem
+    ) -> RefItem:
+        """Create a group containing all child elements for a rich table cell."""
+        group_name = f"rich_cell_group_{len(doc.tables)}_0_0"
+        group_element = doc.add_group(
+            label=GroupLabel.UNSPECIFIED,
+            name=group_name,
+            parent=table_item,
+        )
+
+        # Add all child elements to the group
+        self._add_child_elements(element, group_element, doc)
+
+        return group_element.get_ref()
+
     def _readingorder_elements_to_docling_doc(
         self,
         conv_res: ConversionResult,
@@ -197,11 +214,21 @@ class ReadingOrderModel:
                             )
 
             elif isinstance(element, Table):
-                tbl_data = TableData(
-                    num_rows=element.num_rows,
-                    num_cols=element.num_cols,
-                    table_cells=element.table_cells,
-                )
+                # Check if table has no structure prediction
+                if element.num_rows == 0 and element.num_cols == 0:
+                    # Only create 1x1 table if there are children to put in it
+                    if element.cluster.children:
+                        # Create minimal 1x1 table with rich cell containing all children
+                        tbl_data = TableData(num_rows=1, num_cols=1, table_cells=[])
+                    else:
+                        # Create empty table with no structure
+                        tbl_data = TableData(num_rows=0, num_cols=0, table_cells=[])
+                else:
+                    tbl_data = TableData(
+                        num_rows=element.num_rows,
+                        num_cols=element.num_cols,
+                        table_cells=element.table_cells,
+                    )
 
                 prov = ProvenanceItem(
                     page_no=element.page_no + 1,
@@ -231,6 +258,30 @@ class ReadingOrderModel:
 
                         tbl.footnotes.append(new_footnote_item.get_ref())
 
+                # Handle case where table has no structure prediction but has children
+                if (
+                    element.num_rows == 0
+                    and element.num_cols == 0
+                    and element.cluster.children
+                ):
+                    # Create rich cell containing all child elements
+                    rich_cell_ref = self._create_rich_cell_group(element, out_doc, tbl)
+
+                    # Create rich table cell spanning the entire 1x1 table
+                    rich_cell = RichTableCell(
+                        text="",  # Empty text since content is in the group
+                        row_span=1,
+                        col_span=1,
+                        start_row_offset_idx=0,
+                        end_row_offset_idx=1,
+                        start_col_offset_idx=0,
+                        end_col_offset_idx=1,
+                        column_header=False,
+                        row_header=False,
+                        ref=rich_cell_ref,
+                    )
+                    out_doc.add_table_cell(table_item=tbl, cell=rich_cell)
+
                 # TODO: Consider adding children of Table.
 
             elif isinstance(element, FigureElement):

{docling-2.55.0 → docling-2.55.1}/docling.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling
-Version: 2.55.0
+Version: 2.55.1
 Summary: SDK and CLI for parsing PDF, DOCX, HTML, and more, to a unified document representation for powering downstream workflows such as gen AI applications.
 Author-email: Christoph Auer <cau@zurich.ibm.com>, Michele Dolfi <dol@zurich.ibm.com>, Maxim Lysak <mly@zurich.ibm.com>, Nikos Livathinos <nli@zurich.ibm.com>, Ahmed Nassar <ahn@zurich.ibm.com>, Panos Vagenas <pva@zurich.ibm.com>, Peter Staar <taa@zurich.ibm.com>
 License-Expression: MIT

{docling-2.55.0 → docling-2.55.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "docling"
-version = "2.55.0"  # DO NOT EDIT, updated automatically
+version = "2.55.1"  # DO NOT EDIT, updated automatically
 description = "SDK and CLI for parsing PDF, DOCX, HTML, and more, to a unified document representation for powering downstream workflows such as gen AI applications."
 license = "MIT"
 keywords = [
@@ -123,6 +123,7 @@ dev = [
     "pytest~=8.3",
     "pytest-cov>=6.1.1",
     "pytest-dependency~=0.6",
+    "pytest-durations~=1.6.1",
     "pytest-xdist~=3.3",
     "ipykernel~=6.29",
     "ipywidgets~=8.1",