Merge branch 'main' into main

Author: yrangana

CHANGELOG.md

@@ -1,3 +1,30 @@
+## 1.1.1
+
+* Add NotImplementedError when trying to single index a TextRegions, reflecting the fact that it won't behave correctly at the moment.
+
+## 1.1.0
+
+* Enhancement: Add `TextSource` to track where the text of an element came from
+* Enhancement: Refactor `__post_init__` of `TextRegions` and `LayoutElement` slightly to automate initialization
+
+## 1.0.10
+
+* Remove merging logic that's no longer used
+
+## 1.0.9
+
+* Make OD model loading thread safe
+
+## 1.0.8
+
+* Enhancement: Optimized `zoom_image` (codeflash)
+* Enhancement: Optimized `cells_to_html` for an 8% speedup in some cases (codeflash)
+* Enhancement: Optimized `outputs_to_objects` for an 88% speedup in some cases (codeflash)
+
+## 1.0.7
+
+* Fix a hardcoded file extension causing confusion in the logs
+
 ## 1.0.6
 
 * Add slicing through indexing for vectorized elements

test_unstructured_inference/inference/test_layout.py

@@ -13,6 +13,7 @@
     EmbeddedTextRegion,
     ImageTextRegion,
 )
+from unstructured_inference.constants import IsExtracted
 from unstructured_inference.models.unstructuredmodel import (
     UnstructuredElementExtractionModel,
     UnstructuredObjectDetectionModel,
@@ -34,7 +35,7 @@ def mock_initial_layout():
         6,
         8,
         text="A very repetitive narrative. " * 10,
-        source="Mock",
+        is_extracted=IsExtracted.TRUE,
     )
 
     title_block = EmbeddedTextRegion.from_coords(
@@ -43,7 +44,7 @@ def mock_initial_layout():
         3,
         4,
         text="A Catchy Title",
-        source="Mock",
+        is_extracted=IsExtracted.TRUE,
     )
 
     return [text_block, title_block]

test_unstructured_inference/inference/test_layout_element.py

@@ -1,10 +1,12 @@
 from unstructured_inference.inference.layoutelement import LayoutElement, TextRegion
+from unstructured_inference.constants import IsExtracted, Source
 
 
-def test_layout_element_do_dict(mock_layout_element):
+def test_layout_element_to_dict(mock_layout_element):
     expected = {
         "coordinates": ((100, 100), (100, 300), (300, 300), (300, 100)),
         "text": "Sample text",
+        "is_extracted": None,
         "type": "Text",
         "prob": None,
         "source": None,
@@ -18,3 +20,31 @@ def test_layout_element_from_region(mock_rectangle):
     region = TextRegion(bbox=mock_rectangle)
 
     assert LayoutElement.from_region(region) == expected
+
+
+def test_layoutelement_inheritance_works_correctly():
+    """Test that LayoutElement properly inherits from TextRegion without conflicts"""
+    from unstructured_inference.inference.elements import TextRegion
+
+    # Create a TextRegion with both source and text_source
+    region = TextRegion.from_coords(
+        0, 0, 10, 10, text="test", source=Source.YOLOX, is_extracted=IsExtracted.TRUE
+    )
+
+    # Convert to LayoutElement
+    element = LayoutElement.from_region(region)
+
+    # Check that both properties are preserved
+    assert element.source == Source.YOLOX, "LayoutElement should inherit source from TextRegion"
+    assert (
+        element.is_extracted == IsExtracted.TRUE
+    ), "LayoutElement should inherit is_extracted from TextRegion"
+
+    # Check that to_dict() works correctly
+    d = element.to_dict()
+    assert d["source"] == Source.YOLOX
+    assert d["is_extracted"] == IsExtracted.TRUE
+
+    # Check that we can set source directly on LayoutElement
+    element.source = Source.DETECTRON2_ONNX
+    assert element.source == Source.DETECTRON2_ONNX

test_unstructured_inference/models/test_model.py

@@ -1,4 +1,5 @@
 import json
+import threading
 from typing import Any
 from unittest import mock
 
@@ -40,6 +41,49 @@ def test_get_model(monkeypatch):
         assert isinstance(models.get_model("yolox"), MockModel)
 
 
+def test_get_model_threaded(monkeypatch):
+    """Test that get_model works correctly when called from multiple threads simultaneously."""
+    monkeypatch.setattr(models, "models", {})
+
+    # Results and exceptions from threads will be stored here
+    results = []
+    exceptions = []
+
+    def get_model_worker(thread_id):
+        """Worker function for each thread."""
+        try:
+            model = models.get_model("yolox")
+            results.append((thread_id, model))
+        except Exception as e:
+            exceptions.append((thread_id, e))
+
+    # Create and start multiple threads
+    num_threads = 10
+    threads = []
+
+    with mock.patch.dict(models.model_class_map, {"yolox": MockModel}):
+        for i in range(num_threads):
+            thread = threading.Thread(target=get_model_worker, args=(i,))
+            threads.append(thread)
+            thread.start()
+
+        # Wait for all threads to complete
+        for thread in threads:
+            thread.join()
+
+    # Verify no exceptions occurred
+    assert len(exceptions) == 0, f"Exceptions occurred in threads: {exceptions}"
+
+    # Verify all threads got results
+    assert len(results) == num_threads, f"Expected {num_threads} results, got {len(results)}"
+
+    # Verify all results are MockModel instances
+    for thread_id, model in results:
+        assert isinstance(
+            model, MockModel
+        ), f"Thread {thread_id} got unexpected model type: {type(model)}"
+
+
 def test_register_new_model():
     assert "foo" not in models.model_class_map
     assert "foo" not in models.model_config_map