def test_weightloss(self, mocker):
"""Tests that the weight is updated according to the weightloss method.
The mocker is used to give spesific values from random functions used
in the module.
"""
mocker.patch("numpy.random.normal", return_value=1)
h = Herb()
h.weightloss()
assert h.weight == 1 - h.eta * 1
h.weight = 100000
c = Carn()
c.weight = 100000
weight_list_herb = []
weight_list_carn = []
for _ in range(100):
weight_list_herb.append(h.weight)
weight_list_carn.append(c.weight)
h.weightloss()
c.weightloss()
assert all([
weight1 > weight2 for weight1, weight2 in zip(
weight_list_herb[:-1], weight_list_herb[1:])
]), "The weight does not decrease in the list when weightloss is used"
assert all([
weight1 > weight2 for weight1, weight2 in zip(
weight_list_carn[:-1], weight_list_carn[1:])
]), "The weight does not decrease in the list when weightloss is used"
def test_baseanimal_init_function_inherits_correctly_to_subclass(
self, herb, carn):
"""Tests that the attributes inherits correctly to the animal species.
This test function also tests if the class raises errors if the animals
are initiated with incorrect value ranges.
"""
assert herb.a == 0
assert carn.a == 0
assert carn.weight > 0
assert herb.weight > 0
assert 0 <= herb.fitness <= 1, "Fitness needs to be in interval [0,1]"
assert 0 <= carn.fitness <= 1, "Fitness needs to be in interval [0,1]"
with pytest.raises(ValueError):
Herb(age=-1)
with pytest.raises(ValueError):
Carn(age=-300)
with pytest.raises(ValueError):
Herb(weight=-1)
with pytest.raises(ValueError):
Carn(weight=-1)
with pytest.raises(ValueError):
Herb(age="hello")
with pytest.raises(ValueError):
Carn(weight="hi")
with pytest.raises(ValueError):
Herb(weight=[1, 2, 3])
def test_carnivore_continues_to_feed_if_it_has_appetite(self, mocker):
herb_list = [Herb() for _ in range(100)]
herb_list.sort(key=lambda x: x.fitness, reverse=True)
mocker.patch('random.uniform', return_value=0)
c = Carn()
c.F = 100000
c.fitness = 0.5
for herb in herb_list:
herb.fitness = 0
killed_herbs = len(c.feeding(herb_list))
assert killed_herbs == len(herb_list)
def test_migrate_all_animals_all_move(self, jungle, herb_list_big,
carn_list_big, tear_down_params):
"""Test to see that all animals migrate if they deside to migrate.
"""
j = jungle
j.herb_list = herb_list_big
j.carn_list = carn_list_big
neighbors = (Jungle(), Jungle(), Jungle(), Jungle())
Herb.set_parameters({"mu": 100, "F": 0})
Carn.set_parameters({"mu": 100, "F": 0})
j.migrate_all_animals(neighbors)
assert len(j.herb_move_from_list) == 1000
assert len(j.carn_move_from_list) == 1000
def test_will_birth(self, mocker):
"""Tests that the will_birth method works correctly.
The weights of herbivores and carnivores are set as needed to have the
correct return object from the will_birth method.
The mocker is used to give spesific values from random functions used
in the module.
"""
h = Herb()
c = Carn()
h.weight = 0
c.weight = 0
return_object_herb = h.will_birth(10)
return_object_carn = c.will_birth(10)
assert return_object_herb is None
assert return_object_carn is None
mocker.patch("random.uniform", return_value=0)
h = Herb()
c = Carn()
h.weight = 100
c.weight = 100
return_object_herb = h.will_birth(10000)
return_object_carn = c.will_birth(10000)
assert isinstance(return_object_herb, Herb)
assert isinstance(return_object_carn, Carn)
def set_animal_parameters(self, species, params):
"""
Set parameters for animal species.
:param species: String, name of animal species
:param params: Dict with valid parameter specification for species
"""
if species == "Herbivore":
h = Herb()
h.set_parameters(params)
elif species == "Carnivore":
c = Carn()
c.set_parameters(params)
else:
raise ValueError(f'Got non existing species {species} ')
def test_death(self, mocker):
h = Herb()
c = Carn()
h.fitness = 0
c.fitness = 0
assert h.death() is True
assert c.death() is True
h = Herb()
c = Carn()
mocker.patch('random.uniform', return_value=1)
assert h.death() is False
assert c.death() is False
h = Herb()
c = Carn()
mocker.patch('random.uniform', return_value=0)
assert h.death() is True
assert c.death() is True
def test_carnivore_continues_to_feed_if_it_has_appetite(self, mocker):
"""Tests that the carnivores continues to feed.
Tests that the carnivore feeds until there are no more herbivores on
the nature square, given that its appetite and fitness is high.
Mocker is used to ensure that the carnivore eats for certain.
"""
herb_list = [Herb() for _ in range(100)]
herb_list.sort(key=lambda x: x.fitness, reverse=True)
mocker.patch("random.uniform", return_value=0)
c = Carn()
c.F = 100000
c.fitness = 0.5
for herb in herb_list:
herb.fitness = 0.1
killed_herbs = len(c.feeding(herb_list))
assert killed_herbs == len(herb_list)
def test_BaseAnimal_init_function_inherits_correctly_to_subclass(
self, herb, carn):
assert herb.a == 0
assert carn.a == 0
assert carn.weight > 0
assert herb.weight > 0
assert 0 <= herb.fitness <= 1, 'Fitness needs to be in interval [0,1]'
assert 0 <= carn.fitness <= 1, 'Fitness needs to be in interval [0,1]'
with pytest.raises(ValueError):
h = Herb(age=-1)
with pytest.raises(ValueError):
c = Carn(age=-300)
with pytest.raises(ValueError):
h = Herb(weight=-1)
with pytest.raises(ValueError):
c = Carn(weight=-1)
with pytest.raises(ValueError):
h = Herb(age="hello")
with pytest.raises(ValueError):
c = Carn(weight='hi')
with pytest.raises(ValueError):
h = Herb(weight=[1, 2, 3])
def test_migrate_all_animals_equal_prob(self, jungle, herb_list_big,
carn_list_big, tear_down_params):
"""
Test to see that animal is equally likely to migrate to any square if
move propensity is equal for all squares.
"""
j = jungle
j.herb_list = herb_list_big
j.carn_list = carn_list_big
neighbors = (Jungle(), Jungle(), Jungle(), Jungle())
Herb.set_parameters({"mu": 100, "F": 0})
Carn.set_parameters({"mu": 100, "F": 0})
j.migrate_all_animals(neighbors)
num_moved = np.array((
len(neighbors[0].herb_move_to_list),
len(neighbors[1].herb_move_to_list),
len(neighbors[2].herb_move_to_list),
len(neighbors[3].herb_move_to_list),
))
num_expected = np.array((250, 250, 250, 250))
_, pvalue = chisquare(num_moved, num_expected)
assert pvalue > 0.001
def test_weightloss(self, mocker):
mocker.patch('numpy.random.normal', return_value=1)
h = Herb()
h.weightloss()
assert h.weight == 1 - h.eta * 1
h.weight = 100000
c = Carn()
c.weight = 100000
weight_list_herb = []
weight_list_carn = []
for _ in range(100):
weight_list_herb.append(h.weight)
weight_list_carn.append(c.weight)
h.weightloss()
c.weightloss()
assert all([
weight1 > weight2 for weight1, weight2 in zip(
weight_list_herb[:-1], weight_list_herb[1:])
])
assert all([
weight1 > weight2 for weight1, weight2 in zip(
weight_list_carn[:-1], weight_list_carn[1:])
])
def test_death(self, mocker):
"""Tests that the death method works correctly.
The fitness of herbivores and carnivores are set as needed to have the
correct return object from the death method.
The mocker is used to give spesific values from random functions used
in the module.
"""
h = Herb()
c = Carn()
h.fitness = 0
c.fitness = 0
assert h.death() is True
assert c.death() is True
h = Herb()
c = Carn()
mocker.patch("random.uniform", return_value=1)
assert h.death() is False
assert c.death() is False
h = Herb()
c = Carn()
mocker.patch("random.uniform", return_value=0)
assert h.death() is True
assert c.death() is True
def test_will_birth(self, mocker):
h = Herb()
c = Carn()
h.weight = 0
c.weight = 0
return_object_herb = h.will_birth(10)
return_object_carn = c.will_birth(10)
assert return_object_herb is None
assert return_object_carn is None
mocker.patch('random.uniform', return_value=0)
h = Herb()
c = Carn()
h.weight = 100
c.weight = 100
return_object_herb = h.will_birth(10000)
return_object_carn = c.will_birth(10000)
assert isinstance(return_object_herb, Herb)
assert isinstance(return_object_carn, Carn)
def test_migrate(self, mocker):
mocker.patch('random.uniform', return_value=0)
h = Herb()
c = Carn()
assert h.migrate() is True
assert c.migrate() is True
mocker.patch('random.uniform', return_value=1)
h = Herb()
c = Carn()
assert h.migrate() is False
assert c.migrate() is False
def test_weight_follows_normal_distribution(self):
#np.random.seed(123) # need this?
# Using D.agostinos K^2 test which is accessed from the normaltest in scipy
n_trials = 10000
some_herb_list = [Herb() for _ in range(n_trials)]
some_carn_list = [Carn() for _ in range(n_trials)]
weight_data_herb = [h.weight for h in some_herb_list]
weight_data_carn = [c.weight for c in some_carn_list]
stat, p_value1 = normaltest(weight_data_herb)
stat, p_value2 = normaltest(weight_data_carn)
# Significance level 0.01.
alpha = 0.01
# Nullhypothesis is that the weight follows a normal distribution.
# Weight follows a normal distribution if the test is passed.
# If it doesnt pass, we reject H0 on a 0.01 significance level. This
# means that it is more likely that the data doesnt follow a normal
# distribution. If it passes it means that its probable that it follows
# a normal distribution but we cant say for sure.
assert p_value1 > alpha
assert p_value2 > alpha
def test_migrate(self, mocker):
"""Tests that the migrate method works properly.
The mocker is used to give spesific values from random functions used
in the module.
"""
mocker.patch("random.uniform", return_value=0)
h = Herb()
c = Carn()
assert h.migrate() is True
assert c.migrate() is True
mocker.patch("random.uniform", return_value=1)
h = Herb()
c = Carn()
assert h.migrate() is False
assert c.migrate() is False
def test_weight_follows_normal_distribution(self):
"""A statistical test for the distribution of animal weights.
The test determines whether it is probable that the drawn weights of
the animals follows a normal distribution.
The test uses the normaltest from scipy, which is based on D.agostinos
K^2 test. The nullhypothesis for the test is that the weight follows a
normal distribution. If the assertion fails we reject the
nullhypothesis on the given significance level, which is defined
as alpha.
"""
n_trials = 10000
some_herb_list = [Herb() for _ in range(n_trials)]
some_carn_list = [Carn() for _ in range(n_trials)]
weight_data_herb = [h.weight for h in some_herb_list]
weight_data_carn = [c.weight for c in some_carn_list]
stat, p_value1 = normaltest(weight_data_herb)
stat, p_value2 = normaltest(weight_data_carn)
alpha = 0.001
assert p_value1 > alpha, ("Herbivore weight probably "
"doesn't follow a normal distribution")
assert p_value2 > alpha, ("Carnivore weight probably "
"doesn't follow a normal distribution")
def carn(self):
return Carn()
def carn_list_big(self):
"""Creates a fixture that returns a list of 1000 carnivores.
"""
return [Carn() for _ in range(1000)]
def carn_list_gen(self):
"""Creates a fixture which returns a list of 100 carnivores.
"""
return [Carn() for _ in range(100)]
def tear_down_params(self):
"""Creates a tear_down fixture that resets the parameters.
"""
yield None
Herb().set_default_parameters_for_species()
Carn().set_default_parameters_for_species()
def test_carnivore_weight_and_fitness_updates_after_feeding(self, mocker):
# usikker på denne testen.
a = Herb()
mocker.patch("random.uniform", return_value=0)
c = Carn()
c.F = 10000
c.fitness = 1
herb_list = [a]
herb_list.sort(key=lambda x: x.fitness, reverse=True)
herb_weight_list = [h.weight for h in herb_list]
carn_weight = c.weight + c.beta * herb_weight_list[0]
c.feeding(herb_list)
assert c.weight == pytest.approx(carn_weight)
assert c.fitness == pytest.approx(
1 / (1 + m.exp(c.phi_age * (c.a - c.a_half))) * 1 /
(1 + m.exp(-c.phi_age * (c.weight - c.w_half))))
c = Carn()
c.fitness = 1
c.F = 0.001
herb_list[0].weight = 5
carn_weight = c.weight + c.beta * c.F
c.feeding(herb_list)
assert c.weight == pytest.approx(carn_weight)
def tear_down_params(self):
yield None
Herb().set_default_parameters_for_species()
Carn().set_default_parameters_for_species()
def test_carnivore_weight_and_fitness_updates_after_feeding(self, mocker):
"""Tests that the carnivore weight and fitness updates after feeding.
Mocker is used to ensure that the carnivore eats.
"""
a = Herb()
mocker.patch("random.uniform", return_value=0)
c = Carn()
c.F = 10000
c.fitness = 1
herb_list = [a]
herb_list.sort(key=lambda x: x.fitness, reverse=True)
herb_weight_list = [h.weight for h in herb_list]
carn_weight = c.weight + c.beta * herb_weight_list[0]
c.feeding(herb_list)
assert c.weight == pytest.approx(carn_weight)
assert c.fitness == pytest.approx(
1 / (1 + m.exp(c.phi_age * (c.a - c.a_half))) * 1 /
(1 + m.exp(-c.phi_age * (c.weight - c.w_half))))
c = Carn()
c.fitness = 1
c.F = 0.001
herb_list[0].weight = 5
carn_weight = c.weight + c.beta * c.F
c.feeding(herb_list)
assert c.weight == pytest.approx(carn_weight), "Weight doesn't update"
def carn_params(self):
return Carn().DEFAULT_PARAMETERS
def carn(self):
"""Creates a fixture of the carnivore class instance.
"""
return Carn()
def carn_params(self):
"""Creates a fixture of the carnivore default parameters.
"""
return Carn().DEFAULT_PARAMETERS
def carn_list(self):
return [Carn() for _ in range(100)]
def test_fitness_update(self, mocker):
h = Herb()
c = Carn()
h.weight = -3
c.weight = -2
h.fitness_update()
c.fitness_update()
assert h.fitness == 0
assert c.fitness == 0
mocker.patch('numpy.random.normal', return_value=0)
h = Herb()
c = Carn()
h.fitness_update()
c.fitness_update()
assert h.fitness == 0
assert c.fitness == 0
mocker.patch('numpy.random.normal', return_value=1)
h = Herb()
c = Carn()
h.fitness_update()
c.fitness_update()
assert h.fitness == pytest.approx(
1 / (1 + m.exp(h.phi_age * (h.a - h.a_half))) * 1 /
(1 + m.exp(-h.phi_weight * (1 - h.w_half))))
assert c.fitness == pytest.approx(
1 / (1 + m.exp(c.phi_age * (c.a - c.a_half))) * 1 /
(1 + m.exp(-c.phi_weight * (1 - c.w_half))))
def test_fitness_update(self, mocker):
"""Tests that the fitness gets updated for given values of weight.
The mocker is used to give spesific values from random functions used
in the module.
"""
h = Herb()
c = Carn()
h.weight = -3
c.weight = -2
h.fitness_update()
c.fitness_update()
assert h.fitness == 0
assert c.fitness == 0
mocker.patch("numpy.random.normal", return_value=0)
h = Herb()
c = Carn()
h.fitness_update()
c.fitness_update()
assert h.fitness == 0
assert c.fitness == 0
mocker.patch("numpy.random.normal", return_value=1)
h = Herb()
c = Carn()
h.fitness_update()
c.fitness_update()
assert h.fitness == pytest.approx(
1 / (1 + m.exp(h.phi_age * (h.a - h.a_half))) * 1 /
(1 + m.exp(-h.phi_weight * (1 - h.w_half))))
assert c.fitness == pytest.approx(
1 / (1 + m.exp(c.phi_age * (c.a - c.a_half))) * 1 /
(1 + m.exp(-c.phi_weight * (1 - c.w_half))))