Merge branch 'main' of github.com:DS4SD/docling into nli/layout_heron

CHANGELOG.md

@@ -1,3 +1,14 @@
+## [v2.42.1](https://github.com/docling-project/docling/releases/tag/v2.42.1) - 2025-07-22
+
+### Fix
+
+* Keep formula clusters also when empty ([#1970](https://github.com/docling-project/docling/issues/1970)) ([`67441ca`](https://github.com/docling-project/docling/commit/67441ca4188d532c79df788c461e7f6f7d2f8170))
+
+### Documentation
+
+* Enrich existing DoclingDocument ([#1969](https://github.com/docling-project/docling/issues/1969)) ([`90a7cc4`](https://github.com/docling-project/docling/commit/90a7cc4bdda7272cd87d6f4ab3c0b7966f6e9c73))
+* Add documentation for confidence scores ([#1912](https://github.com/docling-project/docling/issues/1912)) ([`5d98bce`](https://github.com/docling-project/docling/commit/5d98bcea1bd03aff426f903211c931620ff8fcc1))
+
 ## [v2.42.0](https://github.com/docling-project/docling/releases/tag/v2.42.0) - 2025-07-18
 
 ### Feature

docling/utils/layout_postprocessor.py

@@ -267,9 +267,14 @@ class LayoutPostprocessor:
         # Initial cell assignment
         clusters = self._assign_cells_to_clusters(clusters)
 
-        # Remove clusters with no cells (if keep_empty_clusters is False)
+        # Remove clusters with no cells (if keep_empty_clusters is False),
+        # but always keep clusters with label DocItemLabel.FORMULA
         if not self.options.keep_empty_clusters:
-            clusters = [cluster for cluster in clusters if cluster.cells]
+            clusters = [
+                cluster
+                for cluster in clusters
+                if cluster.cells or cluster.label == DocItemLabel.FORMULA
+            ]
 
         # Handle orphaned cells
         unassigned = self._find_unassigned_cells(clusters)

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "docling"
-version = "2.42.0"  # DO NOT EDIT, updated automatically
+version = "2.42.1"  # 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 = [

uv.lock

@@ -806,7 +806,7 @@ wheels = [
 
 [[package]]
 name = "docling"
-version = "2.42.0"
+version = "2.42.1"
 source = { editable = "." }
 dependencies = [
     { name = "accelerate" },