def test_workplace_contact_distribution(do_show, do_save,
create_sample_pop_e2e,
get_fig_dir_by_module):
# calculate the workplace contacts count and plot
sp.logger.info(
"Test workplace contact distribution: workers in workplace such that worksize <= max_contacts must have"
"contacts equal to worksize-1, for workers in workplace with size > max_contacts, the distribution "
"should be closer to poisson distribution with mean = max_contacts ")
plotting_kwargs = sc.objdict(do_show=do_show,
do_save=do_save,
figdir=get_fig_dir_by_module)
contacts, contacts_by_id = cn.get_contact_counts_by_layer(
create_sample_pop_e2e.popdict, layer="w", with_layer_ids=1)
plotting_kwargs.append(
"title_prefix",
f"Total Workers = {len(contacts.get('wpid').get('all'))}")
plotting_kwargs.append("figname", f"workers_contact_count")
sp.plot_contact_counts(contacts, **plotting_kwargs)
plotting_kwargs.remove("title_prefix")
plotting_kwargs.remove("figname")
# check workplace with worksize <= max_contacts
max_contacts = create_sample_pop_e2e.max_contacts['W']
upperbound = st.poisson.interval(alpha=0.95, mu=max_contacts)[1]
group_size_contacts = {
f'all_worksize_contacts size > {max_contacts//2}':
[], # capture size > max_contacts//2
f'large_worksize_contacts size > {upperbound}':
[], # capture size > upperbound
f'medium_large_worksize_contacts size between {max_contacts}, {upperbound}':
[], # capture size between max_contacts and upperbound
f'small_medium_worksize_contacts size between {max_contacts//2}, {max_contacts}':
[], # capture size between max_contacts//2 and max_contacts
}
for k, v in contacts_by_id.items():
if len(v) <= max_contacts // 2:
assert len([i for i in v if i != len(v) - 1]) == 0, \
"Failed, not all contacts in {len(k)} are equal to {len(v)} : {v}"
else:
if len(v) > upperbound:
group_size_contacts[
f'large_worksize_contacts size > {upperbound}'] += v
elif len(v) >= max_contacts:
group_size_contacts[
f'medium_large_worksize_contacts size between {max_contacts}, {upperbound}'] += v
else:
group_size_contacts[
f'small_medium_worksize_contacts size between {max_contacts//2}, {max_contacts}'] += v
group_size_contacts[
f'all_worksize_contacts size > {max_contacts//2}'] += v
file_pattern = re.compile(r'([\s><=])')
for i in group_size_contacts:
plotting_kwargs["title_prefix"] = i
plotting_kwargs["figname"] = file_pattern.sub("_", i)
sp.check_truncated_poisson(testdata=group_size_contacts[i],
mu=max_contacts,
lowerbound=max_contacts // 2,
skipcheck=True if "small" in i else True,
**plotting_kwargs)
def test_plot_contact_counts(do_show=False, do_save=False):
sp.logger.info(
"Test plot_contact_counts method. Unit test --- for actual use with a sp.Pop object see e2etests/test_workplace_e2e.py."
)
# multiple subplots
contacts1 = {
"people_type1": {
"contact_type1": [2, 3, 4, 3, 4, 3, 4, 2, 1, 2],
"contact_type2": [10, 22, 11, 12, 9, 9, 8, 10, 11, 10]
},
"people_type2": {
"contact_type1": [1, 3, 4, 3, 5, 3, 4, 4, 1, 2],
"contact_type2": [10, 21, 11, 11, 10, 9, 8, 10, 11, 10]
}
}
# single subplot
contacts2 = {
"people_type": {
"contact_type": [2, 3, 4, 3, 4, 3, 4, 2, 1, 2]
}
}
params = sc.objdict(do_save=do_save,
do_show=do_show,
title_prefix='test=true')
fig1, ax1 = sp.plot_contact_counts(contact_counter=contacts1, **params)
assert isinstance(
fig1, mplt.figure.Figure), 'Check failed. Figure not generated.'
fig2, ax2 = sp.plot_contact_counts(contact_counter=contacts2,
varname="test",
**params)
assert isinstance(
fig2, mplt.figure.Figure), 'Check failed. Figure not generated.'
print('Check passed. Figures made.')
def test_average_class_size(average_class_size,
do_show,
do_save,
get_fig_dir,
quantiles=None):
"""
Test case to check average_class_size by taking average of student-student contacts
Args:
average_class_size: The average classroom size.
Returns:
None
"""
testpars = dict(
average_class_size=average_class_size,
# average_student_teacher_ratio = average_class_size, # DM: note that this parameter will overide the average class size parameter when school mixing types are something other than random or undefined (which defaults to random) --- method refactor work for schools will clarify these relationships
)
pop = sp.Pop(**pars, **testpars)
plotting_kwargs = sc.objdict(do_show=do_show,
do_save=do_save,
figdir=get_fig_dir)
contacts = sp.get_contact_counts_by_layer(pop.popdict, with_layer_ids=1)[0]
plotting_kwargs.append("title_prefix",
f"Average Class Size = {average_class_size}")
plotting_kwargs.append("figname",
f"contact_average_class_size_{average_class_size}")
sp.plot_contact_counts(contacts, **plotting_kwargs)
counts = []
if not pop.school_pars.with_school_types:
counts.extend(contacts['sc_student']['all'])
counts.extend(contacts['sc_teacher']['all'])
counts.extend(contacts['sc_staff']['all'])
elif pop.school_pars.with_school_types and pop.school_pars.school_mixing_type == 'age_and_class_clustered':
counts.extend(contacts['sc_student']['sc_student'])
sp.check_poisson(actual=counts,
expected=average_class_size,
label='average_class_size',
check='dist')
# visual check with scipy.stats.probplot -- temporary, just to show that the null hypothesis should pass here for the distribution
fig, ax = plt.subplots(1, 1)
res = stats.probplot(counts,
dist=stats.poisson,
sparams=(average_class_size, ),
plot=ax)
if do_show:
plt.show()
plt.close()
return
def test_average_teacher_teacher_degree(average_teacher_teacher_degree,
do_show,
do_save,
get_fig_dir,
threshold=2):
"""
Test case for average_teacher_teacher_degree by taking average of teachers' contacts per teacher
Args:
average_teacher_teacher_degree: The average number of contacts per teacher with other teachers
Returns:
None
"""
testpars = dict(
average_teacher_teacher_degree=average_teacher_teacher_degree,
with_school_types=1,
school_mixing_type={
'pk':
'age_and_class_clustered', # average_teacher_teacher_degree will not be used in school_mixing_type == 'random' scenario
'es': 'age_and_class_clustered',
'ms': 'age_and_class_clustered',
'hs': 'age_clustered',
'uv': 'age_clustered'
})
pop = sp.Pop(**pars, **testpars)
plotting_kwargs = sc.objdict(do_show=do_show,
do_save=do_save,
figdir=get_fig_dir)
contacts = sp.get_contact_counts_by_layer(pop.popdict, with_layer_ids=1)[0]
plotting_kwargs.append(
"title_prefix",
f"Average Teacher-Teacher Degree = {average_teacher_teacher_degree}")
plotting_kwargs.append(
"figname",
f"contact_average_teacher_teacher_degree_{average_teacher_teacher_degree}"
)
sp.plot_contact_counts(contacts, **plotting_kwargs)
counts = contacts['sc_teacher']['sc_teacher']
sp.check_normal(actual=counts,
expected=average_teacher_teacher_degree,
label='teacher degree',
check='mean')
return
def test_average_additional_staff_degree(average_additional_staff_degree,
do_show,
do_save,
get_fig_dir,
threshold=2):
"""
Test case to check average_additional_staff_degree by taking average of all contacts per staff
Args:
average_additional_staff_degree: The average number of contacts per additional non teaching staff in schools
Returns:
None
"""
# note this must be greater than default average_student_all_staff_ratio (20)
testpars = dict(
average_additional_staff_degree=average_additional_staff_degree,
with_school_types=1,
)
pop = sp.Pop(**pars, **testpars)
plotting_kwargs = sc.objdict(do_show=do_show,
do_save=do_save,
figdir=get_fig_dir)
contacts = sp.get_contact_counts_by_layer(pop.popdict, with_layer_ids=1)[0]
plotting_kwargs.append(
"title_prefix",
f"Average Additional Staff Degree = {average_additional_staff_degree}")
plotting_kwargs.append(
"figname",
f"contact_average_additional_staff_degree_{average_additional_staff_degree}"
)
sp.plot_contact_counts(contacts, **plotting_kwargs)
counts = contacts['sc_staff']['all']
sp.check_normal(actual=counts,
expected=average_additional_staff_degree,
label='staff degree',
check='mean')
return