def test_tuple_integer_limits(row_limits, col_limits, weights, prob, seed):
"""Verify that `mutation` creates a valid individual where the lower and
upper column limits and tuple and integer respectively."""
distributions = [Gamma, Normal, Poisson]
families = [Family(distribution) for distribution in distributions]
state = np.random.RandomState(seed)
individual = create_individual(row_limits, col_limits, families, weights,
state)
mutant = mutation(individual, prob, row_limits, col_limits, families,
weights)
_common_asserts(mutant, families)
dataframe, metadata = mutant
assert row_limits[0] <= dataframe.shape[0] <= row_limits[1]
assert sum(col_limits[0]) <= dataframe.shape[1] <= col_limits[1]
family_counts = {
family: sum(pdf.family is family for pdf in metadata)
for family in families
}
for i, count in enumerate(family_counts.values()):
assert count >= col_limits[0][i]
def test_get_fitness(row_limits, col_limits, weights, seed):
"""Create an individual and get its fitness. Then verify that the fitness
is of the correct data type and has been added to the cache."""
distributions = [Normal, Poisson, Uniform]
families = [edo.Family(dist) for dist in distributions]
random_state = np.random.RandomState(seed)
individual = create_individual(
row_limits, col_limits, families, weights, random_state
)
fit = get_fitness(individual, trivial_fitness).compute()
assert isinstance(fit, float)
assert individual.fitness == fit
def test_integer_limits(row_limits, col_limits, weights, prob, seed):
"""Verify that `mutation` creates a valid individual with all integer
column limits."""
distributions = [Gamma, Normal, Poisson]
families = [Family(distribution) for distribution in distributions]
state = np.random.RandomState(seed)
individual = create_individual(row_limits, col_limits, families, weights,
state)
mutant = mutation(individual, prob, row_limits, col_limits, families,
weights)
_common_asserts(mutant, families)
for i, limits in enumerate([row_limits, col_limits]):
assert limits[0] <= mutant.dataframe.shape[i] <= limits[1]
def test_integer_limits(row_limits, col_limits, weights, seed):
"""Create an individual with all-integer column limits and verify that it
is a namedtuple with a `pandas.DataFrame` field of a valid shape, and
metadata made up of instances from the classes in families."""
distributions = [Gamma, Normal, Poisson]
families = [Family(distribution) for distribution in distributions]
state = np.random.RandomState(seed)
individual = create_individual(
row_limits, col_limits, families, weights, state
)
_common_asserts(individual, state, families)
for i, limits in enumerate([row_limits, col_limits]):
assert limits[0] <= individual.dataframe.shape[i] <= limits[1]
def test_integer_limits(row_limits, col_limits, weights, prob, seed):
"""Verify that `crossover` produces a valid individual with all integer
column limits."""
distributions = [Gamma, Normal, Poisson]
families = [Family(distribution) for distribution in distributions]
random_state = np.random.RandomState(seed)
parents = [
create_individual(
row_limits, col_limits, families, weights, random_state
)
for _ in [0, 1]
]
individual = crossover(*parents, col_limits, families, random_state, prob)
_common_asserts(individual, *parents)
def test_tuple_limits(row_limits, col_limits, weights, prob, seed):
"""Verify that `crossover` produces a valid individual with all tuple
column limits."""
distributions = [Gamma, Normal, Poisson]
families = [Family(distribution) for distribution in distributions]
random_state = np.random.RandomState(seed)
parents = [
create_individual(
row_limits, col_limits, families, weights, random_state
)
for _ in [0, 1]
]
individual = crossover(*parents, col_limits, families, random_state, prob)
_common_asserts(individual, *parents)
for i, family in enumerate(families):
count = sum(pdf.family is family for pdf in individual.metadata)
assert col_limits[0][i] <= count <= col_limits[1][i]
def test_to_and_from_file(row_limits, col_limits, weights, seed):
""" Test that an individual can be saved to and created from file. """
path = Path(".testcache/individual")
distributions = [Gamma, Normal, Poisson]
families = [Family(distribution) for distribution in distributions]
state = np.random.RandomState(seed)
individual = create_individual(
row_limits, col_limits, families, weights, state
)
individual.to_file(path, ".testcache")
assert (path / "main.csv").exists()
assert (path / "main.meta").exists()
assert (path / "main.state").exists()
saved_individual = Individual.from_file(path, distributions, ".testcache")
assert np.allclose(
saved_individual.dataframe.values, individual.dataframe.values
)
for saved_pdf, pdf in zip(saved_individual.metadata, individual.metadata):
assert saved_pdf.family.name == pdf.family.name
assert saved_pdf.family.distribution is pdf.family.distribution
assert saved_pdf.to_dict() == pdf.to_dict()
for saved_part, state_part in zip(
saved_individual.random_state.get_state(), state.get_state()
):
try:
assert all(saved_part == state_part)
except TypeError:
assert saved_part == state_part
os.system("rm -r .testcache")
def test_tuple_integer_limits(row_limits, col_limits, weights, prob, seed):
"""Verify that `crossover` produces a valid individual where the lower and
upper column limits are tuple and integer respectively."""
distributions = [Gamma, Normal, Poisson]
families = [Family(distribution) for distribution in distributions]
random_state = np.random.RandomState(seed)
parents = [
create_individual(
row_limits, col_limits, families, weights, random_state
)
for _ in [0, 1]
]
individual = crossover(*parents, col_limits, families, random_state, prob)
_common_asserts(individual, *parents)
for family, lower_limit in zip(families, col_limits[0]):
count = sum(pdf.family is family for pdf in individual.metadata)
assert count >= lower_limit
def test_tuple_integer_limits(row_limits, col_limits, weights, seed):
"""Create an individual with tuple lower limits and integer upper limits on
the columns. Verify the individual is valid and of a reasonable shape and
does not exceed the lower bounds."""
distributions = [Gamma, Normal, Poisson]
families = [Family(distribution) for distribution in distributions]
state = np.random.RandomState(seed)
individual = create_individual(
row_limits, col_limits, families, weights, state
)
_common_asserts(individual, state, families)
dataframe, metadata = individual
assert row_limits[0] <= dataframe.shape[0] <= row_limits[1]
assert sum(col_limits[0]) <= dataframe.shape[1] <= col_limits[1]
for family, lower_limit in zip(families, col_limits[0]):
count = sum([isinstance(pdf, family.distribution) for pdf in metadata])
assert count >= lower_limit
def test_tuple_limits(row_limits, col_limits, weights, seed):
"""Create an individual with tuple column limits. Verify the individual is
valid and of a reasonable shape and does not exceed either of the column
bounds."""
distributions = [Gamma, Normal, Poisson]
families = [Family(distribution) for distribution in distributions]
state = np.random.RandomState(seed)
individual = create_individual(
row_limits, col_limits, families, weights, state
)
_common_asserts(individual, state, families)
dataframe, metadata = individual
assert row_limits[0] <= dataframe.shape[0] <= row_limits[1]
assert sum(col_limits[0]) <= dataframe.shape[1] <= sum(col_limits[1])
for i, family in enumerate(families):
count = sum([isinstance(pdf, family.distribution) for pdf in metadata])
assert col_limits[0][i] <= count <= col_limits[1][i]
""" .. Mutation example. """
import numpy as np
from edo.individual import create_individual
from edo.operators import mutation
from edo.pdfs import Poisson
np.random.seed(0)
row_limits, col_limits = [1, 3], [1, 5]
pdfs = [Poisson]
individual = create_individual(row_limits, col_limits, pdfs)
mutation_prob = 0.5
mutant = mutation(individual, mutation_prob, row_limits, col_limits, pdfs)
for name, ind in zip(["individual", "mutant"], [individual, mutant]):
df, meta = ind
df.to_csv(f"../discussion/operators/{name}.csv")
with open(f"../discussion/operators/{name}.rst", "w") as ind_file:
string = ".. :orphan:\n\n"
string += "And their metadata is::\n\n ["
for col in meta:
string += str(col) + ", "
string = string[:-2]
string += "]\n "
""" .. Crossover example. """
import numpy as np
from edo.individual import create_individual
from edo.operators import crossover
from edo.pdfs import Poisson
np.random.seed(0)
row_limits, col_limits = [1, 3], [1, 5]
pdfs = [Poisson]
parents = [create_individual(row_limits, col_limits, pdfs) for _ in range(2)]
crossover_prob = 0.5
offspring = crossover(*parents, col_limits, pdfs, crossover_prob)
for i, parent in enumerate(parents):
parent.dataframe.to_csv(f"../discussion/operators/parent_{i+1}.csv")
with open("../discussion/operators/parents.rst", "w") as parent_file:
string = ".. :orphan:\n\n"
string += "And their metadata is::\n\n "
for parent in parents:
string += "["
for col in parent.metadata:
string += str(col) + ", "
string = string[:-2]
string += "]\n "
