Merge pull request #266 from PolicyEngine/fix/region-values-comparison

fix: use region.values for Scotland comparisons in loss function

Author: vahid-ahmadi

.github/workflows/pull_request.yaml

@@ -24,7 +24,7 @@ jobs:
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip
-          pip install black==25.1.0
+          pip install black
       - name: Check formatting
         run: black . -l 79 --check
   test:

changelog_entry.yaml

@@ -0,0 +1,4 @@
+- bump: patch
+  changes:
+    fixed:
+      - Use region.values for Scotland comparisons in loss function to ensure consistent behavior with StringArray types

policyengine_uk_data/tests/test_scotland_uc_babies.py

@@ -18,9 +18,7 @@ def test_scotland_uc_households_child_under_1(baseline):
     Target: ~14,000 households (13,992 from Stat-Xplore November 2023)
     Source: DWP Stat-Xplore UC Households dataset
     """
-    region = baseline.calculate(
-        "region", map_to="household", period=2025
-    ).values
+    region = baseline.calculate("region", map_to="household", period=2025)
     uc = baseline.calculate("universal_credit", period=2025).values
     household_weight = baseline.calculate(
         "household_weight", map_to="household", period=2025
@@ -38,7 +36,7 @@ def test_scotland_uc_households_child_under_1(baseline):
     )
 
     scotland_uc_child_under_1 = (
-        (region == "SCOTLAND") & (uc > 0) & has_child_under_1
+        (region.values == "SCOTLAND") & (uc > 0) & has_child_under_1
     )
     total = (household_weight * scotland_uc_child_under_1).sum()
 

policyengine_uk_data/utils/loss.py

@@ -267,7 +267,7 @@ def pe_count(*variables):
     # Children under 16 in Scotland
     # Source: NRS mid-year population estimates
     # https://www.nrscotland.gov.uk/statistics-and-data/statistics/statistics-by-theme/population/population-estimates/mid-year-population-estimates
-    scotland_children_under_16 = (region == "SCOTLAND") & (age < 16)
+    scotland_children_under_16 = (region.values == "SCOTLAND") & (age < 16)
     df["ons/scotland_children_under_16"] = household_from_person(
         scotland_children_under_16
     )
@@ -276,7 +276,7 @@ def pe_count(*variables):
     # Source: NRS Vital Events - births registered in Scotland
     # https://www.nrscotland.gov.uk/publications/vital-events-reference-tables-2024/
     # ~46,000 births per year (45,763 in 2024)
-    scotland_babies_under_1 = (region == "SCOTLAND") & (age < 1)
+    scotland_babies_under_1 = (region.values == "SCOTLAND") & (age < 1)
     df["ons/scotland_babies_under_1"] = household_from_person(
         scotland_babies_under_1
     )