Fix misleading name for overlap method

Resolves: https://github.com/alphagov/notifications-admin/pull/3980#discussion_r692919874

Previously it was unclear what kinds of areas this method returned,
and whether there would be any duplicates (due to the hierarchy of
areas we work with). This clarifies that.

In addition, the areas returned may not overlap with the custom one
[1], so we should reword to avoid falsely implying that. We could do
the overlap check as part of the method as an alternative, but that
would create extra work when calculating the ratio of intersection.
We could always add "overlapping areas" as a complementary method to
this one in future.

[1]: https://github.com/alphagov/notifications-admin/pull/3980#discussion_r692919874

app/broadcast_areas/models.py

@@ -163,10 +163,11 @@ class CustomBroadcastArea(BaseBroadcastArea):
     simple_polygons = polygons
 
     @cached_property
-    def overlapping_areas(self):
+    def nearby_electoral_wards(self):
         if not self.polygons:
             return []
         return broadcast_area_libraries.get_areas_with_simple_polygons([
+            # We only index electoral wards in the RTree
             overlap.data for overlap in rtree_index.query(
                 Rect(*self.polygons.bounds)
             )
@@ -176,7 +177,7 @@ class CustomBroadcastArea(BaseBroadcastArea):
     def count_of_phones(self):
         return sum(
             area.simple_polygons.ratio_of_intersection_with(self.polygons) * area.count_of_phones
-            for area in self.overlapping_areas
+            for area in self.nearby_electoral_wards
         )
 
 

tests/app/broadcast_areas/test_broadcast_area.py

@@ -386,7 +386,7 @@ def test_count_of_phones_for_custom_area(
     )
 
     assert sorted(
-        overlap.name for overlap in area.overlapping_areas
+        overlap.name for overlap in area.nearby_electoral_wards
     ) == expected_possible_overlaps
 
     assert close_enough(area.count_of_phones, expected_count_of_phones)