Merge branch 'main' into update_ruff_invocation

CHANGELOG.md

@@ -1,12 +1,21 @@
-## 0.16.17-dev1
+## 0.16.18-dev1
+
+### Enhancements
+
+### Features
+
+### Fixes
+ **Correct deprecated `ruff` invocation in `make tidy`**.  This will future-proof it or avoid surprises if someone happens to upgrade Ruff.
+
+## 0.16.17
 
 ### Enhancements
 - **Refactoring the VoyageAI integration** to use voyageai package directly, allowing extra features.
 
 ### Features
 
 ### Fixes
-- **Correct deprecated `ruff` invocation in `make tidy`**.  This will future-proof it or avoid surprises if someone happens to upgrade Ruff.
+- **Fix a bug where `build_layout_elements_from_cor_regions` incorrectly joins texts in wrong order**.
 
 ## 0.16.16
 

test_unstructured/partition/pdf_image/test_inference_utils.py

@@ -91,3 +91,79 @@ def test_build_layout_elements_from_ocr_regions_with_multi_line_text(mock_embedd
             ),
         ]
     )
+
+
+def test_build_layout_elements_from_ocr_regions_with_repeated_texts(mock_embedded_text_regions):
+    mock_embedded_text_regions.extend(
+        [
+            LayoutElement.from_coords(
+                x1=453.00277777777774,
+                y1=417.319341111111,
+                x2=711.5338541666665,
+                y2=458.28571222222206,
+                text="LayoutParser",
+                type=ElementType.UNCATEGORIZED_TEXT,
+            ),
+            LayoutElement.from_coords(
+                x1=453.00277777777774,
+                y1=468.319341111111,
+                x2=711.5338541666665,
+                y2=478.28571222222206,
+                text="for",
+                type=ElementType.UNCATEGORIZED_TEXT,
+            ),
+            LayoutElement.from_coords(
+                x1=453.00277777777774,
+                y1=488.319341111111,
+                x2=711.5338541666665,
+                y2=500.28571222222206,
+                text="Deep",
+                type=ElementType.UNCATEGORIZED_TEXT,
+            ),
+            LayoutElement.from_coords(
+                x1=453.00277777777774,
+                y1=510.319341111111,
+                x2=711.5338541666665,
+                y2=550.28571222222206,
+                text="Learning",
+                type=ElementType.UNCATEGORIZED_TEXT,
+            ),
+        ]
+    )
+    text = (
+        "LayoutParser: \n\nA Unified Toolkit for Deep Learning Based Document Image\n\n"
+        "LayoutParser for Deep Learning"
+    )
+    elements = build_layout_elements_from_ocr_regions(
+        TextRegions.from_list(mock_embedded_text_regions),
+        text,
+        group_by_ocr_text=True,
+    )
+    assert elements == LayoutElements.from_list(
+        [
+            LayoutElement.from_coords(
+                x1=453.00277777777774,
+                y1=317.319341111111,
+                x2=711.5338541666665,
+                y2=358.28571222222206,
+                text="LayoutParser:",
+                type=ElementType.UNCATEGORIZED_TEXT,
+            ),
+            LayoutElement.from_coords(
+                x1=437.83888888888885,
+                y1=317.319341111111,
+                x2=1256.334784222222,
+                y2=406.9837855555556,
+                text="A Unified Toolkit for Deep Learning Based Document Image",
+                type=ElementType.UNCATEGORIZED_TEXT,
+            ),
+            LayoutElement.from_coords(
+                x1=453.00277777777774,
+                y1=417.319341111111,
+                x2=711.5338541666665,
+                y2=550.28571222222206,
+                text="LayoutParser for Deep Learning",
+                type=ElementType.UNCATEGORIZED_TEXT,
+            ),
+        ]
+    )

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.16.17-dev1"  # pragma: no cover
+__version__ = "0.16.17"  # pragma: no cover

unstructured/partition/pdf_image/inference_utils.py

@@ -62,12 +62,12 @@ def build_layout_elements_from_ocr_regions(
                     break
                 if text in words:
                     regions.append(indices[mask][i])
-                    mask[mask][i] = False
                     words.remove(text)
 
             if not regions:
                 continue
 
+            mask[regions] = False
             grouped_regions.append(ocr_regions.slice(regions))
     else:
         grouped_regions = partition_groups_from_regions(ocr_regions)