docling 2.28.1__tar.gz → 2.28.3__tar.gz

{docling-2.28.1 → docling-2.28.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: docling
-Version: 2.28.1
+Version: 2.28.3
 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.
 Home-page: https://github.com/docling-project/docling
 License: MIT
@@ -28,7 +28,7 @@ Provides-Extra: vlm
 Requires-Dist: accelerate (>=1.2.1,<2.0.0) ; (sys_platform != "darwin" or platform_machine != "x86_64") and (extra == "vlm")
 Requires-Dist: beautifulsoup4 (>=4.12.3,<5.0.0)
 Requires-Dist: certifi (>=2024.7.4)
-Requires-Dist: docling-core[chunking] (>=2.23.1,<3.0.0)
+Requires-Dist: docling-core[chunking] (>=2.24.1,<3.0.0)
 Requires-Dist: docling-ibm-models (>=3.4.0,<4.0.0)
 Requires-Dist: docling-parse (>=4.0.0,<5.0.0)
 Requires-Dist: easyocr (>=1.7,<2.0)

{docling-2.28.1 → docling-2.28.3}/docling/backend/html_backend.py

@@ -206,9 +206,9 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
         hlevel = int(element.name.replace("h", ""))
         text = element.text.strip()
 
-        if hlevel == 1:
-            self.content_layer = ContentLayer.BODY
+        self.content_layer = ContentLayer.BODY
 
+        if hlevel == 1:
             for key in self.parents.keys():
                 self.parents[key] = None
 
@@ -243,7 +243,7 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
             self.parents[hlevel] = doc.add_heading(
                 parent=self.parents[hlevel - 1],
                 text=text,
-                level=hlevel,
+                level=hlevel - 1,
                 content_layer=self.content_layer,
             )
 

{docling-2.28.1 → docling-2.28.3}/docling/backend/md_backend.py

@@ -212,9 +212,16 @@ class MarkdownDocumentBackend(DeclarativeDocumentBackend):
             traverse(element)
             snippet_text = "".join(strings)
             if len(snippet_text) > 0:
-                parent_item = doc.add_text(
-                    label=doc_label, parent=parent_item, text=snippet_text
-                )
+                if doc_label == DocItemLabel.SECTION_HEADER:
+                    parent_item = doc.add_heading(
+                        text=snippet_text,
+                        level=element.level - 1,
+                        parent=parent_item,
+                    )
+                else:
+                    parent_item = doc.add_text(
+                        label=doc_label, parent=parent_item, text=snippet_text
+                    )
 
         elif isinstance(element, marko.block.List):
             has_non_empty_list_items = False
@@ -232,12 +239,15 @@ class MarkdownDocumentBackend(DeclarativeDocumentBackend):
                     label=label, name=f"list", parent=parent_item
                 )
 
-        elif isinstance(element, marko.block.ListItem) and len(element.children) > 0:
+        elif (
+            isinstance(element, marko.block.ListItem)
+            and len(element.children) > 0
+            and isinstance((first_child := element.children[0]), marko.block.Paragraph)
+        ):
             self._close_table(doc)
             self._process_inline_text(parent_item, doc)
             _log.debug(" - List item")
 
-            first_child = element.children[0]
             snippet_text = str(first_child.children[0].children)  # type: ignore
             is_numbered = False
             if (

{docling-2.28.1 → docling-2.28.3}/docling/models/table_structure_model.py

@@ -5,7 +5,11 @@ from typing import Iterable, Optional, Union
 
 import numpy
 from docling_core.types.doc import BoundingBox, DocItemLabel, TableCell
-from docling_core.types.doc.page import BoundingRectangle
+from docling_core.types.doc.page import (
+    BoundingRectangle,
+    SegmentedPdfPage,
+    TextCellUnit,
+)
 from docling_ibm_models.tableformer.data_management.tf_predictor import TFPredictor
 from PIL import ImageDraw
 
@@ -218,9 +222,18 @@ class TableStructureModel(BasePageModel):
 
                     if len(table_bboxes):
                         for table_cluster, tbl_box in in_tables:
-
+                            # Check if word-level cells are available from backend:
+                            sp = page._backend.get_segmented_page()
+                            if sp is not None:
+                                tcells = sp.get_cells_in_bbox(
+                                    cell_unit=TextCellUnit.WORD,
+                                    bbox=table_cluster.bbox,
+                                )
+                            else:
+                                # Otherwise - we use normal (line/phrase) cells
+                                tcells = table_cluster.cells
                             tokens = []
-                            for c in table_cluster.cells:
+                            for c in tcells:
                                 # Only allow non empty stings (spaces) into the cells of a table
                                 if len(c.text.strip()) > 0:
                                     new_cell = copy.deepcopy(c)
@@ -229,7 +242,6 @@ class TableStructureModel(BasePageModel):
                                             scale=self.scale
                                         )
                                     )
-
                                     tokens.append(
                                         {
                                             "id": new_cell.index,

{docling-2.28.1 → docling-2.28.3}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "docling"
-version = "2.28.1"  # DO NOT EDIT, updated automatically
+version = "2.28.3"  # 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."
 authors = [
   "Christoph Auer <cau@zurich.ibm.com>",
@@ -46,7 +46,7 @@ packages = [{ include = "docling" }]
 ######################
 python = "^3.9"
 pydantic = "^2.0.0"
-docling-core = {extras = ["chunking"], version = "^2.23.1"}
+docling-core = {extras = ["chunking"], version = "^2.24.1"}
 docling-ibm-models = "^3.4.0"
 docling-parse = "^4.0.0"
 filetype = "^1.2.0"