docling-core 2.26.2__py3-none-any.whl → 2.26.4__py3-none-any.whl

docling_core/experimental/serializer/common.py

@@ -69,7 +69,9 @@ def _iterate_items(
     node: Optional[NodeItem] = None,
     traverse_pictures: bool = False,
     add_page_breaks: bool = False,
+    visited: Optional[set[str]] = None,
 ):
+    my_visited: set[str] = visited if visited is not None else set()
     prev_page_nr: Optional[int] = None
     page_break_i = 0
     for item, _ in doc.iterate_items(
@@ -78,10 +80,33 @@ def _iterate_items(
         included_content_layers=layers,
         traverse_pictures=traverse_pictures,
     ):
-        if isinstance(item, DocItem):
-            if item.prov:
+        if add_page_breaks:
+            if (
+                isinstance(item, (UnorderedList, OrderedList, InlineGroup))
+                and item.self_ref not in my_visited
+            ):
+                # if group starts with new page, yield page break before group node
+                my_visited.add(item.self_ref)
+                for it in _iterate_items(
+                    doc=doc,
+                    layers=layers,
+                    node=item,
+                    traverse_pictures=traverse_pictures,
+                    add_page_breaks=add_page_breaks,
+                    visited=my_visited,
+                ):
+                    if isinstance(it, DocItem) and it.prov:
+                        page_no = it.prov[0].page_no
+                        if prev_page_nr is not None and page_no > prev_page_nr:
+                            yield _PageBreakNode(
+                                self_ref=f"#/pb/{page_break_i}",
+                                prev_page=prev_page_nr,
+                                next_page=page_no,
+                            )
+                        break
+            elif isinstance(item, DocItem) and item.prov:
                 page_no = item.prov[0].page_no
-                if add_page_breaks and (prev_page_nr is None or page_no > prev_page_nr):
+                if prev_page_nr is None or page_no > prev_page_nr:
                     if prev_page_nr is not None:  # close previous range
                         yield _PageBreakNode(
                             self_ref=f"#/pb/{page_break_i}",

{docling_core-2.26.2.dist-info → docling_core-2.26.4.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: docling-core
-Version: 2.26.2
+Version: 2.26.4
 Summary: A python library to define and validate data types in Docling.
 Home-page: https://github.com/docling-project
 License: MIT
@@ -36,7 +36,7 @@ Requires-Dist: pyyaml (>=5.1,<7.0.0)
 Requires-Dist: semchunk (>=2.2.0,<3.0.0) ; extra == "chunking"
 Requires-Dist: tabulate (>=0.9.0,<0.10.0)
 Requires-Dist: transformers (>=4.34.0,<5.0.0) ; extra == "chunking"
-Requires-Dist: typer (>=0.12.5,<0.13.0)
+Requires-Dist: typer (>=0.12.5,<0.16.0)
 Requires-Dist: typing-extensions (>=4.12.2,<5.0.0)
 Project-URL: Repository, https://github.com/docling-project/docling-core
 Description-Content-Type: text/markdown

{docling_core-2.26.2.dist-info → docling_core-2.26.4.dist-info}/RECORD

@@ -4,7 +4,7 @@ docling_core/cli/view.py,sha256=gwxSBYhGqwznMR8pdXaEuAh2bjFD5X_g11xFYSgFgtM,1764
 docling_core/experimental/__init__.py,sha256=XnAVSUHbA6OFhNSpoYqSD3u83-xVaUaki1DIKFw69Ew,99
 docling_core/experimental/serializer/__init__.py,sha256=CECQlMoCDUxkg4RAUdC3itA3I3qFhKhe2HcYghN6_xw,105
 docling_core/experimental/serializer/base.py,sha256=1sD1v5rWC4MT_Y6BWpMDjUAwuEqC0TR9YjQJZlhPt50,5901
-docling_core/experimental/serializer/common.py,sha256=z80B2BzUdDfp_HgZ1KA64vK-oV07jcgxLs1XyCgp7sI,16152
+docling_core/experimental/serializer/common.py,sha256=iQUJPRZUhpGMi_s4makkZcINy5sdtxu2ehS9N8lnoMM,17332
 docling_core/experimental/serializer/doctags.py,sha256=e97FJHh77x--g2t1O2YprBzF8lkihn_xOr59EjnR7ag,17794
 docling_core/experimental/serializer/html.py,sha256=C-xf_PH28R6DoISFDoZwBK67rsfFmTemgZ_VW2i0haY,31896
 docling_core/experimental/serializer/html_styles.py,sha256=-jBwS4EU7yfKoz0GSoxhwx90OmIKieO6TwPw57IuxcA,4692
@@ -65,8 +65,8 @@ docling_core/utils/generate_jsonschema.py,sha256=uNX1O5XnjyB5nA66XqZXTt3YbGuR2ty
 docling_core/utils/legacy.py,sha256=SqNQAxl97aHfoJEsC9vZcMJg5FNkmqKPFi-wdSrnfI0,24442
 docling_core/utils/validate.py,sha256=aQ11UbFyl8iD_N7yTTZmm_VVeXz8KcCyn3GLXgkfYRM,2049
 docling_core/utils/validators.py,sha256=azcrndLzhNkTWnbFSu9shJ5D3j_znnLrIFA5R8hzmGU,2798
-docling_core-2.26.2.dist-info/LICENSE,sha256=2M9-6EoQ1sxFztTOkXGAtwUDJvnWaAHdB9BYWVwGkIw,1087
-docling_core-2.26.2.dist-info/METADATA,sha256=motwJC3ixvU3oYstkDMsu5YmnHryDzuFfn3udzpDGrY,5843
-docling_core-2.26.2.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-docling_core-2.26.2.dist-info/entry_points.txt,sha256=oClcdb2L2RKx4jdqUykY16Kum_f0_whwWhGzIodyidc,216
-docling_core-2.26.2.dist-info/RECORD,,
+docling_core-2.26.4.dist-info/LICENSE,sha256=2M9-6EoQ1sxFztTOkXGAtwUDJvnWaAHdB9BYWVwGkIw,1087
+docling_core-2.26.4.dist-info/METADATA,sha256=S9qfeO5R35zkBwzIYnd5q3-O3LBl3-WOsQRlL50klqw,5843
+docling_core-2.26.4.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+docling_core-2.26.4.dist-info/entry_points.txt,sha256=oClcdb2L2RKx4jdqUykY16Kum_f0_whwWhGzIodyidc,216
+docling_core-2.26.4.dist-info/RECORD,,