def test_generate_larger_households(location='seattle_metro',
state_location='Washington',
country_location='usa'):
Nhomes_to_sample_smooth = 1000
household_size_distr = sp.get_household_size_distr(datadir, location,
state_location,
country_location)
hh_sizes = sp.generate_household_sizes(Nhomes_to_sample_smooth,
household_size_distr)
hha_brackets = sp.get_head_age_brackets(datadir,
country_location=country_location)
hha_by_size_counts = sp.get_head_age_by_size_distr(
datadir, country_location=country_location)
age_brackets_filepath = sp.get_census_age_brackets_path(
datadir, state_location, country_location)
age_brackets = sp.get_age_brackets_from_df(age_brackets_filepath)
age_by_brackets_dic = sp.get_age_by_brackets_dic(age_brackets)
contact_matrix_dic = sp.get_contact_matrix_dic(
datadir, sheet_name='United States of America')
single_year_age_distr = {}
for n in range(101):
single_year_age_distr[n] = float(1.0 / 101.0)
# generate households of size 3
size = 3
# first variable is the household size to be created, so here this means we want to create all households of size 3 and the hh_sizes variable tells us how many of size 3 will be created at index 3-1 (since hh_sizes is an array rather than a dictionary)
larger_households = sp.generate_larger_households(
size, hh_sizes, hha_by_size_counts, hha_brackets, age_brackets,
age_by_brackets_dic, contact_matrix_dic, single_year_age_distr)
assert larger_households is not None
print(larger_households)
def test_generate_household_sizes(location='seattle_metro', state_location='Washington', country_location='usa'):
sc.heading('Generate household sizes')
Nhomes_to_sample_smooth = 1000
household_size_distr = sp.get_household_size_distr(datadir, location, state_location, country_location)
hh_sizes = sp.generate_household_sizes(Nhomes_to_sample_smooth, household_size_distr)
assert len(hh_sizes) == 7
def test_get_totalpopsizes_from_household_sizes(location='seattle_metro', state_location='Washington',
country_location='usa'):
household_size_distr = sp.get_household_size_distr(datadir, location, state_location, country_location)
Nhomes_to_sample_smooth = 1000
hh_sizes = sp.generate_household_sizes(Nhomes_to_sample_smooth, household_size_distr)
sum_hh_sizes = sp.get_totalpopsize_from_household_sizes(hh_sizes)
assert sum_hh_sizes is not None
def test_generate_household_sizes(self):
"""
Test generate_larger_household_sizes method if hh_size =1, it expectes
method to return an empty array, otherwise an array of counts which the
total should match the the hh_size[1:].
Returns:
None
"""
size1 = sp.generate_household_sizes(hh_sizes=[])
self.assertEqual(len(size1), 0)
for i in range(2, 10):
size = np.random.randint(low=1, high=50, size=i)
with self.subTest(size=size):
print(f"hh_size:{size}")
result = sp.generate_household_sizes(hh_sizes=size)
print(f"actual hh_size:{collections.Counter(size)}")
self.assertEqual(sum(size), len(result))
def test_custom_household_size_distro_honored(self):
self.is_debugging = False
custom_distro = {
1: 0.25,
2: 0.075,
3: 0.10,
4: 0.30,
5: 0.05,
6: 0.05,
7: 0.175
}
hh_sizes = sp.generate_household_sizes(500, custom_distro)
hh_size_list = list(hh_sizes) # Comes as np.ndarray
fewest_houses = min(hh_size_list)
fewest_index = hh_size_list.index(fewest_houses)
most_houses = max(hh_size_list)
most_index = hh_size_list.index(most_houses)
highest_probability = max(custom_distro.values())
lowest_probability = min(custom_distro.values())
most_houses_probability = custom_distro[most_index +
1] # hh_distro is 1 indexed
fewest_houses_probability = custom_distro[fewest_index + 1]
self.assertEqual(
highest_probability,
most_houses_probability,
msg=
"The most common household size should be the size with the highest probability"
)
prob_bucket_list = list(custom_distro.values())
self.verify_portion_honored(probability_buckets=prob_bucket_list,
count_buckets=hh_size_list,
portion=0.25)
self.verify_portion_honored(probability_buckets=prob_bucket_list,
count_buckets=hh_size_list,
portion=0.2)
self.verify_portion_honored(probability_buckets=prob_bucket_list,
count_buckets=hh_size_list,
portion=0.1)
print(household_size_distr)
# Nhomes = 20000
Nhomes = 10000
# create_homes = True
create_homes = False
if create_homes:
household_size_distr = sp.get_household_size_distr(datadir, location,
state_location,
country_location,
use_bayesian)
print(household_size_distr)
Nhomes_to_sample_smooth = 100000
hh_sizes = sp.generate_household_sizes(Nhomes_to_sample_smooth,
household_size_distr)
totalpop = sp.get_totalpopsize_from_household_sizes(hh_sizes)
# hh_sizes = sp.generate_household_sizes(Nhomes,household_size_distr)
syn_ages, syn_sexes = sp.get_usa_age_sex_n(location, state_location,
totalpop)
syn_age_count = Counter(syn_ages)
syn_age_distr = sp.norm_dic(Counter(syn_ages))
N = Nhomes
hh_sizes = sp.generate_household_sizes_from_fixed_pop_size(
N, household_size_distr)
totalpop = sp.get_totalpopsize_from_household_sizes(hh_sizes)
print(totalpop, 'pop')
