Bump policyengine-core to 3.23.5 for pandas 3.0 compatibility (#1492)

* Bump policyengine-core to 3.23.5 for pandas 3.0 compatibility

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* Add test for pandas 3.0 enum encoding compatibility

Adds tests to verify that policyengine-core 3.23.5 correctly handles
pandas Series with StringDtype index when encoding enums. This ensures
the fix using .iloc[0] for positional access works as expected.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* [TDD DEMO] Pin core to 3.23.4 to demonstrate test failure

This commit intentionally pins policyengine-core to 3.23.4 (before the
pandas 3.0 fix) to demonstrate that the test_pandas3_enum_encoding test
fails without the fix. The next commit will revert to >=3.23.5.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* [TDD DEMO] Bump core to >=3.23.5 to fix pandas 3.0 compatibility

This commit restores policyengine-core>=3.23.5, which includes the fix
for pandas 3.0 enum encoding (using .iloc[0] instead of array[0]).

The previous commit demonstrated the test failure. This commit shows
the test passing with the fix in place.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* Bump microdf-python to >=1.2.1 for pandas 3.0 compatibility

microdf 1.2.1 includes the fix for pandas 3.0 Copy-on-Write
compatibility in MicroDataFrame column access.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* Fix black formatting in test file

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* Bump policyengine-core to >=3.23.6 for pandas 3.0 StringArray fix

Core 3.23.6 fixes TypeError: unhashable type: 'StringArray' when using
pandas StringArray for parameter lookups (issue #429 in policyengine-core).

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

Author: MaxGhenis

changelog_entry.yaml

@@ -0,0 +1,4 @@
+- bump: patch
+  changes:
+    changed:
+      - Bumped policyengine-core minimum version to 3.23.5 for pandas 3.0 compatibility

policyengine_uk/tests/core/__init__.py

@@ -0,0 +1,44 @@
+"""Test pandas 3.0 compatibility with enum encoding.
+
+This test verifies that policyengine-core 3.23.5+ correctly handles
+pandas Series with StringDtype index when encoding enums.
+"""
+
+import numpy as np
+import pandas as pd
+import pytest
+
+from policyengine_core.enums import Enum
+
+
+class SampleEnum(Enum):
+    VALUE_A = "value_a"
+    VALUE_B = "value_b"
+
+
+def test_enum_encode_with_pandas_series():
+    """Test that Enum.encode works with pandas Series containing enum items.
+
+    In pandas 3.0, Series with StringDtype use label-based indexing by default.
+    This test verifies the fix in policyengine-core 3.23.5 that uses .iloc[0]
+    for positional access instead of array[0].
+    """
+    enum_items = [SampleEnum.VALUE_A, SampleEnum.VALUE_B, SampleEnum.VALUE_A]
+    series = pd.Series(enum_items)
+
+    # This would fail with KeyError: 0 before the fix
+    encoded = SampleEnum.encode(series)
+
+    assert len(encoded) == 3
+    assert list(encoded) == [0, 1, 0]
+
+
+def test_enum_encode_with_string_index():
+    """Test enum encoding with explicitly string-indexed Series."""
+    enum_items = [SampleEnum.VALUE_A, SampleEnum.VALUE_B]
+    series = pd.Series(enum_items, index=["a", "b"])
+
+    encoded = SampleEnum.encode(series)
+
+    assert len(encoded) == 2
+    assert list(encoded) == [0, 1]

policyengine_uk/tests/core/test_pandas3_enum_encoding.py

@@ -24,8 +24,8 @@ classifiers = [
 ]
 requires-python = ">=3.13,<3.14"
 dependencies = [
-    "policyengine-core>=3.23.0",
-    "microdf-python>=1.0.2",
+    "policyengine-core>=3.23.6",
+    "microdf-python>=1.2.1",
     "pydantic>=2.11.7",
     "tables>=3.10.2",
 ]