fix: normalize official agricultural crop groups

backend/app/services/vector_feature_service.py

@@ -101,8 +101,25 @@ SEMANTIC_COUNT_LABELS = {
     "agriculture": "Landbouwgebruikspercelen",
 }
 
+# Sprint 205 initially normalized two official comma-separated ALZ group labels
+# mechanically. Keep those persisted values queryable while new artifacts use
+# the explicit controlled keys.
+SELECTION_FILTER_VALUE_ALIASES: dict[tuple[str, str], tuple[str, ...]] = {
+    ("main_crop_group_key", "grains_seeds_legumes"): ("granen,_zaden_en_peulvruchten",),
+    ("main_crop_group_key", "horticulture"): ("groenten,_kruiden_en_sierplanten",),
+}
+
 
 class VectorFeatureService:
+    @staticmethod
+    def _expanded_selection_filter_values(filter_property: str, filter_values: list[Any]) -> list[str]:
+        expanded: list[str] = []
+        for value in filter_values:
+            normalized = str(value)
+            expanded.append(normalized)
+            expanded.extend(SELECTION_FILTER_VALUE_ALIASES.get((filter_property, normalized), ()))
+        return list(dict.fromkeys(expanded))
+
     @staticmethod
     def _dataset_theme(dataset: Dataset) -> str | None:
         source_metadata = dataset.source_metadata if isinstance(dataset.source_metadata, dict) else {}
@@ -453,7 +470,10 @@ class VectorFeatureService:
                     details={"dataset_id": str(dataset.id), "filter_property": filter_property},
                     status_code=500,
                 )
-            normalized_filter_values = [str(value) for value in filter_values]
+            normalized_filter_values = VectorFeatureService._expanded_selection_filter_values(
+                filter_property,
+                filter_values,
+            )
             metric_filter += (
                 VectorFeature.properties_json.op("->>")(filter_property).in_(normalized_filter_values),
             )