Appease circleci

mggg · Apr 30, 2024 · d895b7a · d895b7a
1 parent be82f60
commit d895b7a
Show file tree

Hide file tree

Showing 4 changed files with 30 additions and 11 deletions.
diff --git a/tests/optimization/test_gingleator.py b/tests/optimization/test_gingleator.py
@@ -53,6 +53,7 @@ def gingleator_test_partition(four_by_five_grid_for_opt):
 
 
 def test_ginglator_needs_min_perc_or_min_pop_col(four_by_five_grid_for_opt):
+    random.seed(2024)
     initial_partition = Partition.from_random_assignment(
         graph=four_by_five_grid_for_opt,
         n_parts=4,
@@ -90,6 +91,7 @@ def test_ginglator_needs_min_perc_or_min_pop_col(four_by_five_grid_for_opt):
 
 
 def test_ginglator_warns_if_min_perc_and_min_pop_col_set(four_by_five_grid_for_opt):
+    random.seed(2024)
     initial_partition = Partition.from_random_assignment(
         graph=four_by_five_grid_for_opt,
         n_parts=4,
@@ -132,7 +134,8 @@ def test_ginglator_warns_if_min_perc_and_min_pop_col_set(four_by_five_grid_for_o
     )
 
 
-def test_ginglator_finds_best_partition(four_by_five_grid_for_opt):
+def test_gingleator_finds_best_partition(four_by_five_grid_for_opt):
+    random.seed(2024)
     initial_partition = Partition.from_random_assignment(
         graph=four_by_five_grid_for_opt,
         n_parts=4,
@@ -164,7 +167,7 @@ def test_ginglator_finds_best_partition(four_by_five_grid_for_opt):
         score_function=Gingleator.num_opportunity_dists,
     )
 
-    total_steps = 1000
+    total_steps = 5000
     burst_length = 10
 
     max_scores_sb = np.zeros(total_steps)

diff --git a/tests/optimization/test_single_metric.py b/tests/optimization/test_single_metric.py
@@ -20,6 +20,7 @@ def simple_cut_edge_count(partition):
 
 
 def test_single_metric_sb_attains_min_quickly(four_by_five_grid_for_opt):
+    random.seed(2024)
     initial_partition = Partition.from_random_assignment(
         graph=four_by_five_grid_for_opt,
         n_parts=4,
@@ -66,6 +67,7 @@ def test_single_metric_sb_attains_min_quickly(four_by_five_grid_for_opt):
 
 
 def test_single_metric_tilted_sb_attains_min_quickly(four_by_five_grid_for_opt):
+    random.seed(2024)
     initial_partition = Partition.from_random_assignment(
         graph=four_by_five_grid_for_opt,
         n_parts=4,
@@ -113,6 +115,7 @@ def test_single_metric_tilted_sb_attains_min_quickly(four_by_five_grid_for_opt):
 
 
 def test_single_metric_variable_len_sb_attains_min_quickly(four_by_five_grid_for_opt):
+    random.seed(2024)
     initial_partition = Partition.from_random_assignment(
         graph=four_by_five_grid_for_opt,
         n_parts=4,
@@ -163,6 +166,7 @@ def test_single_metric_variable_len_sb_attains_min_quickly(four_by_five_grid_for
 
 
 def test_single_metric_sa_jumpcycle_attains_min_quickly(four_by_five_grid_for_opt):
+    random.seed(2024)
     initial_partition = Partition.from_random_assignment(
         graph=four_by_five_grid_for_opt,
         n_parts=4,
@@ -209,6 +213,7 @@ def test_single_metric_sa_jumpcycle_attains_min_quickly(four_by_five_grid_for_op
 
 
 def test_single_metric_sa_lincycle_attains_min_quickly(four_by_five_grid_for_opt):
+    random.seed(2024)
     initial_partition = Partition.from_random_assignment(
         graph=four_by_five_grid_for_opt,
         n_parts=4,
@@ -256,6 +261,7 @@ def test_single_metric_sa_lincycle_attains_min_quickly(four_by_five_grid_for_opt
 def test_single_metric_sa_linear_jumpcycle_attains_min_quickly(
     four_by_five_grid_for_opt,
 ):
+    random.seed(2024)
     initial_partition = Partition.from_random_assignment(
         graph=four_by_five_grid_for_opt,
         n_parts=4,
@@ -301,6 +307,7 @@ def test_single_metric_sa_linear_jumpcycle_attains_min_quickly(
 
 
 def test_single_metric_sa_logitcycle_attains_min_quickly(four_by_five_grid_for_opt):
+    random.seed(2024)
     initial_partition = Partition.from_random_assignment(
         graph=four_by_five_grid_for_opt,
         n_parts=4,
@@ -348,6 +355,7 @@ def test_single_metric_sa_logitcycle_attains_min_quickly(four_by_five_grid_for_o
 def test_single_metric_sa_logit_jumpcycle_attains_min_quickly(
     four_by_five_grid_for_opt,
 ):
+    random.seed(2024)
     initial_partition = Partition.from_random_assignment(
         graph=four_by_five_grid_for_opt,
         n_parts=4,
@@ -400,6 +408,7 @@ def test_single_metric_sa_logit_jumpcycle_attains_min_quickly(
 def test_single_metric_tilted_runs_attains_min_quickly_with_p_eq_0p1(
     four_by_five_grid_for_opt,
 ):
+    random.seed(2024)
     initial_partition = Partition.from_random_assignment(
         graph=four_by_five_grid_for_opt,
         n_parts=4,
@@ -446,6 +455,8 @@ def test_single_metric_tilted_runs_attains_min_quickly_with_p_eq_0p1(
 
 
 def test_single_metric_sb_finds_hard_max(four_by_five_grid_for_opt):
+    random.seed(2024)
+
     def opt_fn(partition):
         mx = 10
         count = sum(1 for x in partition["opt_value_sum"].values() if x == mx)
@@ -494,6 +505,8 @@ def opt_fn(partition):
 
 
 def test_single_metric_sa_finds_hard_max(four_by_five_grid_for_opt):
+    random.seed(2024)
+
     def opt_fn(partition):
         mx = 10
         count = sum(1 for x in partition["opt_value_sum"].values() if x == mx)
@@ -525,13 +538,13 @@ def opt_fn(partition):
         maximize=True,
     )
 
-    total_steps = 10000
+    total_steps = 20000
 
     max_scores_anneal = np.zeros(total_steps)
     for i, part in enumerate(
         optimizer.simulated_annealing(
             total_steps,
-            optimizer.linear_jumpcycle_beta_function(10, 10, 10),
+            optimizer.linear_jumpcycle_beta_function(100, 100, 100),
             beta_magnitude=1,
         )
     ):
@@ -541,6 +554,8 @@ def opt_fn(partition):
 
 
 def test_single_metric_tilted_runs_finds_hard_max(four_by_five_grid_for_opt):
+    random.seed(2024)
+
     def opt_fn(partition):
         mx = 10
         count = sum(1 for x in partition["opt_value_sum"].values() if x == mx)
@@ -572,12 +587,12 @@ def opt_fn(partition):
         maximize=True,
     )
 
-    total_steps = 5000
+    total_steps = 10000
 
-    min_scores_tilt = np.zeros(total_steps)
+    max_scores_tilt = np.zeros(total_steps)
     for i, part in enumerate(
         optimizer.tilted_run(num_steps=total_steps, p=0.1, with_progress_bar=True)
     ):
-        min_scores_tilt[i] = optimizer.best_score
+        max_scores_tilt[i] = optimizer.best_score
 
-    assert np.max(min_scores_tilt) == 2
+    assert np.max(max_scores_tilt) == 2
diff --git a/tests/test_reproducibility.py b/tests/test_reproducibility.py
@@ -62,7 +62,7 @@ def test_repeatable(three_by_three_grid):
     assert flips == expected_flips
 
 
-# @pytest.mark.slow
+@pytest.mark.slow
 def test_pa_freeze():
     from gerrychain import (
         GeographicPartition,

diff --git a/tests/updaters/test_cut_edges.py b/tests/updaters/test_cut_edges.py
@@ -4,7 +4,7 @@
 
 from gerrychain import MarkovChain, Partition, proposals
 from gerrychain.accept import always_accept
-from gerrychain.constraints import no_vanishing_districts, contiguous
+from gerrychain.constraints import no_vanishing_districts, single_flip_contiguous
 from gerrychain.grid import Grid
 from gerrychain.updaters import cut_edges, cut_edges_by_part
 import random
@@ -142,11 +142,12 @@ def test_implementation_of_cut_edges_matches_naive_method(three_by_three_grid):
     ],
 )
 def test_cut_edges_matches_naive_cut_edges_at_every_step(proposal, number_of_steps):
+    random.seed(2024)
     partition = Grid((10, 10), with_diagonals=True)
 
     chain = MarkovChain(
         proposal,
-        [contiguous, no_vanishing_districts],
+        [single_flip_contiguous, no_vanishing_districts],
         always_accept,
         partition,
         number_of_steps,