Merge pull request #318 from AureumChaos/317-crowding-distance-calc-n…

…eeds-to-handle-single-individual-pops

317 crowding distance calc needs to handle single individual pops

.readthedocs.yaml

@@ -22,7 +22,6 @@ python:
   install:
     - requirements: docs/requirements.txt
 
-
 build:
   os: ubuntu-22.04
   tools:

leap_ec/multiobjective/ops.py

@@ -175,6 +175,12 @@ def per_rank_crowding_calc(ranked_population: list, is_maximizing) -> list:
     :param ranked_population: A population of entirely one rank
     :returns: population with crowding distance calculate for one rank
     """
+    if len(ranked_population) == 1:
+        # There is just one individual in this rank, so set their distance to 0
+        # since there are no neighbors.
+        ranked_population[0].distance = 0
+        return ranked_population
+
     # Presuming this is a population with homogeneous objectives, then the size of
     # the optimization directions array should be equal to the number of objectives.
     num_objectives = is_maximizing.shape[0]
@@ -212,7 +218,7 @@ def per_rank_crowding_calc(ranked_population: list, is_maximizing) -> list:
     sorted_pop = []
 
     for objective in range(num_objectives):
-        if objective_ranges[objective] == 0:
+        if objective_ranges[objective] == 0 or objective_ranges[objective] == np.nan:
             continue
 
         # sort by objective being mindful that maximization vs. minimization may
@@ -239,7 +245,8 @@ def per_rank_crowding_calc(ranked_population: list, is_maximizing) -> list:
                                             objective]) / objective_ranges[
                                             objective]
 
-    return sorted_pop
+    # return ranked_population not sorted since sometimes sorted_pop will be empty
+    return ranked_population
 
 @wrap_curry
 @listlist_op

tests/distributed/test_evaluate.py

@@ -3,7 +3,7 @@
 """
 import numpy as np
 import pytest
-from distributed import Client
+from distributed import Client, LocalCluster
 
 from leap_ec.binary_rep.problems import MaxOnes
 from leap_ec.distrib.evaluate import evaluate, is_viable
@@ -18,7 +18,7 @@ def test_good_eval():
         leap_ec.distrib.evaluate works for normal circumstances.
     """
     # set up a basic dask local cluster
-    with Client() as client:
+    with Client(LocalCluster(n_workers=1, threads_per_worker=1)) as client:
         # hand craft an individual that should evaluate fine
         # Let's try evaluating a single individual
         individual = Individual(np.array([1, 1]),
@@ -40,7 +40,7 @@ def test_broken_individual_eval():
         TODO implement this
     """
     # set up a basic dask local cluster
-    with Client() as client:
+    with Client(LocalCluster(n_workers=1, threads_per_worker=1)) as client:
         # hand craft an individual that will intentionally fail by not
         # assigning it a Problem class
         individual = DistributedIndividual(np.array([1, 1]),