Support aggregating local authority clusters

This applies some heuristics to try and keep the overall list of
areas short when many are selected in the same wider area.

Currently we only have relationship information between upper and
lower tier local authorities, so we can't / won't aggregate up to
Greater London (it's own special thing) or whole countries.

app/broadcast_areas/utils.py

@@ -1,9 +1,12 @@
+from collections import defaultdict
+
 from app.broadcast_areas.models import CustomBroadcastArea
 
 
 def aggregate_areas(areas):
     areas = _convert_custom_areas_to_wards(areas)
     areas = _aggregate_wards_by_local_authority(areas)
+    areas = _aggregate_lower_tier_authorities(areas)
     return areas
 
 
@@ -24,3 +27,39 @@ def _aggregate_wards_by_local_authority(areas):
         area.parent if area.id.startswith('wd20-')
         else area for area in areas
     }
+
+
+def _aggregate_lower_tier_authorities(areas):
+    results = set()
+    clusters = _cluster_lower_tier_authorities(areas)
+
+    for cluster in clusters:
+        # always keep lone area as itself
+        if len(cluster) == 1:
+            results |= set(cluster)
+        # aggregate multi-area cluster
+        elif len(cluster) > 3:
+            results |= {cluster[0].parent}
+        # aggregate many small clusters
+        elif len(clusters) > 1:
+            area = cluster[0]
+            results |= {area.parent or area}
+        # keep one small cluster in full
+        else:
+            results |= set(cluster)
+
+    return results
+
+
+def _cluster_lower_tier_authorities(areas):
+    result = defaultdict(lambda: [])
+
+    for area in areas:
+        # group lower tier authorities by "county"
+        if area.id.startswith('lad20-') and area.parent:
+            result[area.parent] += [area]
+        # leave countries, unitary authorities as-is
+        else:
+            result[area] = [area]
+
+    return result.values()

tests/app/broadcast_areas/custom_polygons.py

@@ -37,3 +37,19 @@ CHELTENHAM = [
     [51.9328, -2.0221],
     [51.9324, -2.1265],
 ]
+
+CHELTENHAM_AND_GLOUCESTER = [
+    [51.8820, -2.2920],
+    [51.8234, -2.2570],
+    [51.8883, -2.0262],
+    [51.9425, -2.0771],
+    [51.8820, -2.2920],
+]
+
+SEVERN_ESTUARY = [
+    [51.5719, -2.9388],
+    [51.4180, -2.7259],
+    [51.8493, -2.1958],
+    [51.8883, -2.3016],
+    [51.5719, -2.9388],
+]

tests/app/broadcast_areas/test_utils.py

@@ -7,12 +7,24 @@ from tests.app.broadcast_areas.custom_polygons import (
     BRISTOL,
     BURFORD,
     CHELTENHAM,
+    CHELTENHAM_AND_GLOUCESTER,
     SANTA_A,
+    SEVERN_ESTUARY,
     SKYE,
 )
 
 
 @pytest.mark.parametrize(('area_ids', 'expected_area_names'), [
+    (
+        [
+            'wd20-E05009336',  # Whitechapel, Tower Hamlets (electoral ward)
+            'wd20-E05009372',  # Hackney Central, Hackney (electoral ward)
+            'wd20-E05009374',  # Hackney Wick, Hackney (electoral ward)
+        ], [
+            'Hackney',  # in Greater London* (DB doesn't know this)
+            'Tower Hamlets',  # in Greater London*
+        ],
+    ),
     (
         [
             'wd20-E05004294',  # Hester’s Way, Cheltenham (electoral ward)
@@ -31,6 +43,16 @@ from tests.app.broadcast_areas.custom_polygons import (
             'Hackney',  # in Greater London* (DB doesn't know this)
         ],
     ),
+    (
+        [
+            'wd20-E05004294',  # Hester’s Way, Cheltenham (electoral ward)
+            'wd20-E05010981',  # Painswick & Upton, Stroud (electoral ward)
+            'wd20-E05009372',  # Hackney Central, Hackney (electoral ward)
+        ], [
+            'Gloucestershire',  # upper tier authority
+            'Hackney',  # in Greater London* (DB doesn't know this)
+        ],
+    ),
     (
         [
             'lad20-E07000037',  # High Peak (lower tier authority)
@@ -49,6 +71,17 @@ from tests.app.broadcast_areas.custom_polygons import (
             'High Peak',  # in Derbyshire
         ]
     ),
+    (
+        [
+            'lad20-E07000037',  # High Peak (lower tier authority)
+            'lad20-E07000035',  # Derbyshire Dales (lower tier authority)
+            'ctyua19-E10000028',  # Staffordshire (upper tier authority)
+        ],
+        [
+            'Derbyshire',  # upper tier authority
+            'Staffordshire',
+        ]
+    ),
     (
         [
             'ctry19-E92000001',  # England
@@ -84,6 +117,19 @@ def test_aggregate_areas(
             'Bristol, City of',
         ]
     ),
+    (
+        [SEVERN_ESTUARY], [
+            # Area covers various lower-tier authorities
+            # with more than three in Gloucestershire so
+            # we aggregate them into that
+            'Bristol, City of',
+            'Gloucestershire',
+            'Monmouthshire',
+            'Newport',
+            'North Somerset',
+            'South Gloucestershire',
+        ]
+    ),
     (
         [CHELTENHAM], [
             # Area covers three lower-tier authorities
@@ -93,14 +139,30 @@ def test_aggregate_areas(
             'Tewkesbury',
         ]
     ),
+    (
+        [CHELTENHAM_AND_GLOUCESTER], [
+            # Area covers more than 3 lower-tier authorities
+            # in Gloucestershire so we aggregate them
+            'Gloucestershire',
+        ]
+    ),
     (
         [BURFORD], [
-            # Area covers two lower-tier authorities
-            # in Gloucestershire (upper tier authority)
+            # Area covers one lower-tier authority in
+            # Gloucestershire and one in Oxfordshire (both
+            # upper tier authorities)
             'Cotswold',
             'West Oxfordshire',
         ],
     ),
+    (
+        [CHELTENHAM_AND_GLOUCESTER, BURFORD], [
+            # Area covers many lower-tier authorities
+            # in Gloucestershire but only one in Oxfordshire
+            'Gloucestershire',
+            'West Oxfordshire',
+        ],
+    ),
     (
         [SANTA_A], [
             # Does not overlap with the UK