implement ensemble analysis as script

gcrsef-gang · Feb 23, 2023 · 36e74dd · 36e74dd
1 parent b9ef667
commit 36e74dd
Show file tree

Hide file tree

Showing 9 changed files with 466 additions and 99 deletions.
diff --git a/rba/__main__.py b/rba/__main__.py
@@ -52,6 +52,13 @@
     draw_parser.add_argument("--partition_file", type=str, default=None)
     draw_parser.set_defaults(func=rba.visualization.visualize)
 
+    ensemble_parser = subparsers.add_parser("ensemble")
+    ensemble_parser.add_argument("--graph_file", type=str, default=os.path.join(package_dir, "data/2010/new_hampshire_geodata_merged.json"))
+    ensemble_parser.add_argument("--community_file", type=str, default=os.path.join(package_dir, "data/2010/new_hampshire_communities.json"))
+    ensemble_parser.add_argument("--vra_config_file", type=str, default=os.path.join(package_dir, "data/2010/vra_nh.json"))
+    ensemble_parser.add_argument("--district_file", type=str, default=os.path.join(package_dir, "data/2010/new_hampshire_districts.json"))
+    ensemble_parser.set_defaults(func=rba.ensemble.ensemble_analysis)
+
     optimize_parser = subparsers.add_parser("optimize")
     optimize_parser.add_argument("--graph_file", type=str, default=os.path.join(package_dir, "data/2010/new_hampshire_geodata_merged.json"))
     optimize_parser.add_argument("--communitygen_out_file", type=str, default=os.path.join(package_dir, "data/2010/new_hampshire_communities.json"))

diff --git a/rba/community_generation.py b/rba/community_generation.py
@@ -79,8 +79,11 @@ def compute_precinct_similarities(graph, verbose=False):
         pop_density_distance = abs(pop1 - pop2)
 
         similarity = 1 - np.average(
-            [race_distance, votes_distance, pop_density_distance],
-            weights=[SIMILARITY_WEIGHTS["race"], SIMILARITY_WEIGHTS["votes"], SIMILARITY_WEIGHTS["pop_density"]])
+            [race_distance, pop_density_distance],
+            weights=[SIMILARITY_WEIGHTS["race"], SIMILARITY_WEIGHTS["pop_density"]])
+        # similarity = 1 - np.average(
+        #     [race_distance, votes_distance, pop_density_distance],
+        #     weights=[SIMILARITY_WEIGHTS["race"], SIMILARITY_WEIGHTS["votes"], SIMILARITY_WEIGHTS["pop_density"]])
         # print(similarity, race_distance, votes_distance, pop_density_distance)
         graph.edges[node1, node2]["similarity"] = similarity
         race_distances.append(race_distance)

diff --git a/rba/constants.py b/rba/constants.py
@@ -4,7 +4,8 @@
 
 # Edges between nodes in the same county should be weighted less than those that cross, because
 # the maximum spanning tree should be made more likely to choose an edge crossing county lines.
-SAME_COUNTY_PENALTY = 0.1
+# SAME_COUNTY_PENALTY = 0.1
+SAME_COUNTY_PENALTY = 10
 
 POP_EQUALITY_THRESHOLD = 0.005