def test_get_fitness_value(self):
ga = GASolver(max_generations, population_size, antibody_cnt, nc_cnt,
c_cnt, markers, cell_cnt, mu_list, sigma_list)
val = ga.get_fitness_value([1, 2, 3, 4])
self.assertEqual(val, 0, "fitness not updated yet")
ga.fitness["1-2-3-4"] = 0.5
val = ga.get_fitness_value([1, 2, 3, 4])
self.assertAlmostEqual(val, 0.5, "fitness updated")
def test_survive_n_fittest(self):
ga = GASolver(max_generations, population_size, antibody_cnt, nc_cnt,
c_cnt, markers, cell_cnt, mu_list, sigma_list)
ga.create_random_population()
fit_cnt = ga.survive_n_fittest(0, 2)
self.assertNotEqual(ga.population[1][0].tolist(), [0, 0, 0, 0])
self.assertNotEqual(ga.population[1][1].tolist(), [0, 0, 0, 0])
self.assertEqual(ga.population[1][fit_cnt + 1].tolist(), [0, 0, 0, 0])
self.assertEqual(ga.population[1][fit_cnt + 2].tolist(), [0, 0, 0, 0])
def test_get_fitness_key(self):
ga = GASolver(max_generations, population_size, antibody_cnt, nc_cnt,
c_cnt, markers, cell_cnt, mu_list, sigma_list)
key = ga.get_ab_key(np.array([1, 2, 3, 4]))
self.assertEqual(key, "1-2-3-4")
key = ga.get_ab_key(np.array([0, 20, 15, 14]))
self.assertEqual(key, "0-20-15-14")
key = ga.get_ab_key(np.array([0, 20, 15, 14, 3, 5, 7, 7]))
self.assertEqual(key, "0-20-15-14-3-5-7-7")
def test_is_child_unique(self):
ga = GASolver(max_generations, population_size, antibody_cnt, nc_cnt,
c_cnt, markers, cell_cnt, mu_list, sigma_list)
arr = np.array([1, 2, 3, 4])
res = ga.is_child_unique(0, arr)
self.assertTrue(res)
arr = np.array([1, 2, 3, 4])
ga.population[0][8] = arr
res = ga.is_child_unique(0, arr)
self.assertFalse(res)
def test_mutate(self):
ga = GASolver(max_generations, population_size, antibody_cnt, nc_cnt,
c_cnt, markers, cell_cnt, mu_list, sigma_list)
child = np.array([1, 2, 3, 4])
mutant = ga.mutate(child)
self.assertTrue(mutant[0] != mutant[1] and mutant[1] != mutant[2]
and mutant[2] != mutant[3])
self.assertTrue(mutant[0] < mutant[1] and mutant[1] < mutant[2]
and mutant[2] < mutant[3])
self.assertTrue(mutant[0] != child[0] or mutant[1] != child[1]
or mutant[2] != child[2] or mutant[3] != child[3])
def test_is_child_diverse(self):
ga = GASolver(max_generations, population_size, antibody_cnt, nc_cnt,
c_cnt, markers, cell_cnt, mu_list, sigma_list)
arr = np.array([1, 2, 3, 4])
res = ga._is_child_diverse(arr)
self.assertTrue(res)
arr = np.array([1, 2, 3, 4, 1])
res = ga._is_child_diverse(arr)
self.assertFalse(res)
arr = np.array([1, 2, 3, 4, 3, 1])
res = ga._is_child_diverse(arr)
self.assertFalse(res)
def test_create_random_population(self):
ga = GASolver(max_generations, population_size, antibody_cnt, nc_cnt,
c_cnt, markers, cell_cnt, mu_list, sigma_list)
ga.create_random_population()
for i in range(population_size - 1):
self.assertNotEqual(ga.population[0][i].tolist(),
ga.population[0][i + 1].tolist(),
"ab groups are different")
for j in range(len(ga.population[0][i]) -
1): # which is 4 -- 4 antibodies
self.assertNotEqual(ga.population[0][i][j],
ga.population[0][i][j + 1],
"ab's are different in each group")
self.assertLess(ga.population[0][i][j], antibody_cnt,
"antibody ids are correct")
self.assertGreaterEqual(ga.population[0][i][j], 0,
"antibody ids are correct")
def test_cross_over(self):
ga = GASolver(max_generations, population_size, antibody_cnt, nc_cnt,
c_cnt, markers, cell_cnt, mu_list, sigma_list)
group1 = np.array([1, 2, 3, 4])
group2 = np.array([1, 6, 4, 8])
child = ga.cross_over(group1, group2)
self.assertTrue(child.tolist() != group1.tolist()
and child.tolist() != group2.tolist())
self.assertTrue(child[0] != child[1] and child[1] != child[2]
and child[2] != child[3])
self.assertTrue(child[0] < child[1] and child[1] < child[2]
and child[2] < child[3])
def test_init(self):
ga = GASolver(max_generations, population_size, antibody_cnt, nc_cnt,
c_cnt, markers, cell_cnt, mu_list, sigma_list)
self.assertEqual(ga.max_generations, max_generations)
self.assertEqual(ga.population_size, population_size)
self.assertEqual(ga.antibody_cnt, antibody_cnt)
self.assertEqual(ga.nc_cnt, nc_cnt)
self.assertEqual(ga.c_cnt, c_cnt)
self.assertEqual(ga.markers_list, markers)
self.assertEqual(ga.total_population, population_size)
def test_find_max_fitness_and_child(self):
ga = GASolver(max_generations, population_size, antibody_cnt, nc_cnt,
c_cnt, markers, cell_cnt, mu_list, sigma_list)
ga.create_random_population()
# don't include measured. All the population is measured so return -1
val = ga.find_max_fitness_and_child(0, False)
self.assertEqual(val['fitness'], -1)
self.assertEqual(val['child'].tolist(), ga.population[0][0].tolist())
# include measured
val = ga.find_max_fitness_and_child(0, True)
self.assertGreaterEqual(val['fitness'], 0)
ga.cross_over_generation(0, ga.population_size / 2)
# don't include measured. Some are unmeasured so must be > 0
val = ga.find_max_fitness_and_child(0, False)
self.assertGreaterEqual(val['fitness'], 0)
val = ga.find_max_fitness_and_child(0, True)
self.assertGreaterEqual(val['fitness'], 0)
def test_mutate_generation(self):
ga = GASolver(max_generations, population_size, antibody_cnt, nc_cnt,
c_cnt, markers, cell_cnt, mu_list, sigma_list)
np.random.rand(1)[0]
ga.create_random_population()
pop1 = np.copy(ga.population[0])
ga.mutate_generation(0, ga.population_size / 2, ga.population_size)
pop2 = ga.population[1]
self.assertNotEqual(pop1.tolist(), pop2.tolist())
def test_cross_over_generation(self):
ga = GASolver(max_generations, population_size, antibody_cnt, nc_cnt,
c_cnt, markers, cell_cnt, mu_list, sigma_list)
ga.create_random_population()
ga.survive_n_fittest(0, population_size / 2)
start_ind = population_size / 2
new_start_ind = ga.cross_over_generation(1, start_ind)
self.assertEqual(ga.total_population, population_size * 5 / 4)
self.assertEqual(new_start_ind, start_ind * 3 / 2)
# half is zero
self.assertTrue([0, 0, 0, 0] not in ga.population[0].tolist())
x = [el for el in ga.population[1] if el.tolist() == [0, 0, 0, 0]]
self.assertEqual(len(x), population_size / 4)