enhancement: raise notimplementederror on single index rather than random error (#450)

<!-- CURSOR_SUMMARY -->
> [!NOTE]
> Explicitly raises NotImplementedError when single-indexing
`TextRegions`; adds a skipped test, updates changelog, and bumps version
to 1.1.1.
> 
> - **Elements**:
> - `unstructured_inference/inference/elements.py`: Single-element
indexing of `TextRegions` now raises `NotImplementedError` in `slice()`
(used by `__getitem__`).
> - **Tests**:
> - `test_unstructured_inference/test_elements.py`: Adds a skipped test
covering single-element access expectations.
> - **Release**:
>   - `__version__` bumped to `1.1.1`.
> - `CHANGELOG.md`: Notes the new `NotImplementedError` behavior for
single indexing.
> 
> <sup>Written by [Cursor
Bugbot](https://cursor.com/dashboard?tab=bugbot) for commit
64b9438. This will update automatically
on new commits. Configure
[here](https://cursor.com/dashboard?tab=bugbot).</sup>
<!-- /CURSOR_SUMMARY -->

Author: qued

CHANGELOG.md

@@ -1,3 +1,7 @@
+## 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

test_unstructured_inference/test_elements.py

@@ -593,3 +593,31 @@ def test_textregions_from_coords_accepts_source():
 
     assert region.source == Source.YOLOX
     assert region.is_extracted
+
+
+@pytest.mark.skip(reason="Not implemented")
+def test_textregions_allows_for_single_element_access_and_returns_textregion_with_correct_values():
+    """Test that TextRegions allows for single element access and returns a TextRegion with the
+    correct values"""
+
+    regions = [
+        TextRegion.from_coords(
+            0, 0, 10, 10, text="first", source=Source.YOLOX, is_extracted=IsExtracted.TRUE
+        ),
+        TextRegion.from_coords(
+            0,
+            0,
+            20,
+            20,
+            text="second",
+            source=Source.DETECTRON2_ONNX,
+            is_extracted=IsExtracted.PARTIAL,
+        ),
+    ]
+    text_regions = TextRegions.from_list(regions)
+    for i, region in enumerate(regions):
+        sliced = text_regions[i]
+        assert isinstance(sliced, TextRegion)
+        assert sliced.text == region.text
+        assert sliced.source == region.source
+        assert sliced.is_extracted is region.is_extracted

unstructured_inference/__version__.py

@@ -1 +1 @@
-__version__ = "1.1.0"  # pragma: no cover
+__version__ = "1.1.1"  # pragma: no cover

unstructured_inference/inference/elements.py

@@ -253,6 +253,9 @@ def slice(self, indices) -> TextRegions:
         # NOTE(alan): I would expect if I try to access a single element, it should return a
         # TextRegion, not a TextRegions. Currently, you get an error when trying to access a single
         # element.
+        if self.element_coords[indices].ndim == 1:
+            # We've indexed a single element. For now this isn't implemented.
+            raise NotImplementedError("Slicing a single element is not implemented")
         return TextRegions(
             element_coords=self.element_coords[indices],
             texts=self.texts[indices],