def test_parameter_sweep_plot_level_warnings(self, mgmt_data):
"""
This testcase executes a parameter sweep and captures any warnings
generated during the construction of the bokeh figure object.
:param mgmt_data: Use management department data
:type mgmt_data: dictionary
:return: Pass/Fail
:rtype: Boolean
"""
with warnings.catch_warnings(record=True) as w:
warnings.simplefilter("always")
modlist = list([Mod_Validate_Sweep(**mgmt_data)])
c = Comparison(modlist)
plot_settings = {'plottype': 'parameter sweep percentage',
'intervals': 'empirical',
'number_of_runs': 500,
# number simulations to average over
'target': 0.25,
'xlabel': 'Promotion Rate for Women',
'ylabel': 'Proportion Women',
'title': 'Sweep, Hire-Promote, Female Promotion Rate 1->2',
'model_legend_label': ['Model 1, Hire-Promote',
'Model '
'2, '
'Promote-Hire'],
'parameter_sweep_param': 'female_promotion_probability_1',
'parameter_ubound': 0.5,
'parameter_lbound': 0.05,
'number_of_steps': 15
}
c.plot_comparison_overall_chart(**plot_settings)
assert (len(w) == 0)
def test_parameter_sweep_function_overall_catch_warnings(self, mgmt_data):
"""
This testcase executes a parameter sweep and then checks for any
warnings generated while computing the plot figure. In the past there
were issues with column sizes not matching the in the plot functions.
:param mgmt_data: Use management department data
:type mgmt_data: dictionary
:return: Pass/Fail
:rtype: Boolean
"""
with warnings.catch_warnings(record=True) as w:
warnings.simplefilter("always")
modlist = list([Mod_Validate_Sweep(**mgmt_data)])
c = Comparison(modlist)
plot_settings = {
'plottype':
'parameter sweep percentage',
'intervals':
'empirical',
'number_of_runs':
3,
# number simulations to average over
'target':
0.25,
'xlabel':
'Hiring Rate for Women',
'ylabel':
'Proportion Women',
'title':
'Parameter Sweep Validation, uniform noise(0,'
'0.1)',
'model_legend_label':
['Model 1, Hire-Promote', 'Model '
'2, '
'Promote-Hire'],
'parameter_sweep_param':
'bf1',
'parameter_ubound':
0.6,
'parameter_lbound':
0.05,
'number_of_steps':
3
}
c.plot_comparison_overall_chart(**plot_settings)
assert (len(w) == 0)
def test_plot_overall_mf_numbers(mgmt_data):
modlist = list([Model3GenderDiversity(**mgmt_data)])
# modlist = list([Model2GenderDiversity(**mgmt_data),
# Mod_Stoch_FBPH(**mgmt_data)])
modlist[0].init_default_hiring_rate()
c = Comparison(modlist)
plot_settings = {
'plottype': 'male female numbers',
'intervals': 'empirical',
'number_of_runs': 100,
# number simulations to average over
'target': 0.25,
# target percentage of women in the department
# Main plot settings
'xlabel': 'Years',
'ylabel': 'Number of Professors',
'title': 'Male Female numbers plot',
'line_width': 2,
'transparency': 0.25,
'model_legend_label': ['Model 3', 'Mode 2, Promote-Hire'],
'legend_location': 'top_right',
'height_': height,
'width_': width,
'male_female_numbers_plot': True,
'mf_male_label': ['Male Model 3', 'Female Model 3'],
'mf_target_label': ['Target 3']
}
show(c.plot_comparison_overall_chart(**plot_settings))
def test_plot_overall_unfilled_vacancies(mgmt_data):
modlist = list([Model3GenderDiversity(**mgmt_data)])
# modlist = list([Model2GenderDiversity(**mgmt_data),
# Mod_Stoch_FBPH(**mgmt_data)])
modlist[0].init_default_hiring_rate()
c = Comparison(modlist)
plot_settings = {
'plottype': 'unfilled vacancies',
'intervals': 'empirical',
'number_of_runs': 100,
# number simulations to average over
'target': 0.25,
# target percentage of women in the department
# Main plot settings
'xlabel': 'Years',
'ylabel': 'Unfilled Vacancies',
'title': 'Unfilled Vacancies',
'line_width': 2,
'transparency': 0.25,
'model_legend_label': ['New Model', 'Mode 2, Promote-Hire'],
'legend_location': 'top_right',
'height_': height,
'width_': width,
}
show(c.plot_comparison_overall_chart(**plot_settings))
def test_bokeh_sweep_plot_overall(mgmt_data):
modlist = list([Model3GenderDiversity(**mgmt_data)])
# modlist = list([Model2GenderDiversity(**mgmt_data),
# Mod_Stoch_FBPH(**mgmt_data)])
modlist[0].init_default_hiring_rate()
c = Comparison(modlist)
plot_settings = {
'plottype': 'parameter sweep percentage',
'intervals': 'empirical',
'number_of_runs': 100,
# number simulations to average over
'target': 0.25,
'xlabel': 'Years',
'ylabel': 'Proportion Women',
'title': 'Parameter Sweep Gender Percentage',
'model_legend_label': ['New Model', 'Model '
'2, '
'Promote-Hire'],
'parameter_sweep_param': 'bf1',
'parameter_ubound': 0.6,
'parameter_lbound': 0.05,
'number_of_steps': 5
}
show(c.plot_comparison_overall_chart(**plot_settings))
def test_bokeh_comparison_plot_overall_one_model(mgmt_data):
modlist = list(
[ModelGenderDiversityGrowthForecastIncremental(**mgmt_data)])
# modlist = list([Model2GenderDiversity(**mgmt_data),
# Mod_Stoch_FBPH(**mgmt_data)])
modlist[0].init_default_hiring_rate()
modlist[0].init_growth_rate([0.02, 0.01, 0.10, 0.05])
# modlist[0].init_growth_rate([0.015])
c = Comparison(modlist)
# print(modlist[0].calculate_yearly_dept_size_targets())
plot_settings = {
'plottype': 'gender proportion',
'intervals': 'empirical',
'number_of_runs': 100,
# number simulations to average over
'target': 0.25,
# target percentage of women in the department
# Main plot settings
'xlabel': 'Years',
'ylabel': 'Proportion Women',
'title': 'Change in Proportion Women',
'line_width': 2,
'transparency': 0.25,
'model_legend_label': ['New Model', 'Mode 2, Promote-Hire'],
'legend_location': 'top_right',
'height_': height,
'width_': width,
'year_offset': 0
}
show(c.plot_comparison_overall_chart(**plot_settings))
def test_parameter_sweep_plot_level_warnings(self, mgmt_data):
"""
This testcase executes a parameter sweep and captures any warnings
generated during the construction of the bokeh figure object.
:param mgmt_data: Use management department data
:type mgmt_data: dictionary
:return: Pass/Fail
:rtype: Boolean
"""
with warnings.catch_warnings(record=True) as w:
warnings.simplefilter("always")
modlist = list([Mod_Validate_Sweep(**mgmt_data)])
c = Comparison(modlist)
plot_settings = {
'plottype':
'parameter sweep percentage',
'intervals':
'empirical',
'number_of_runs':
500,
# number simulations to average over
'target':
0.25,
'xlabel':
'Promotion Rate for Women',
'ylabel':
'Proportion Women',
'title':
'Sweep, Hire-Promote, Female Promotion Rate 1->2',
'model_legend_label':
['Model 1, Hire-Promote', 'Model '
'2, '
'Promote-Hire'],
'parameter_sweep_param':
'female_promotion_probability_1',
'parameter_ubound':
0.5,
'parameter_lbound':
0.05,
'number_of_steps':
15
}
c.plot_comparison_overall_chart(**plot_settings)
assert (len(w) == 0)
def test_parameter_sweep_function_overall_catch_warnings(self, mgmt_data):
"""
This testcase executes a parameter sweep and then checks for any
warnings generated while computing the plot figure. In the past there
were issues with column sizes not matching the in the plot functions.
:param mgmt_data: Use management department data
:type mgmt_data: dictionary
:return: Pass/Fail
:rtype: Boolean
"""
with warnings.catch_warnings(record=True) as w:
warnings.simplefilter("always")
modlist = list([Mod_Validate_Sweep(**mgmt_data)])
c = Comparison(modlist)
plot_settings = {'plottype': 'parameter sweep percentage',
'intervals': 'empirical',
'number_of_runs': 3,
# number simulations to average over
'target': 0.25,
'xlabel': 'Hiring Rate for Women',
'ylabel': 'Proportion Women',
'title': 'Parameter Sweep Validation, uniform noise(0,'
'0.1)',
'model_legend_label': ['Model 1, Hire-Promote',
'Model '
'2, '
'Promote-Hire'],
'parameter_sweep_param': 'bf1',
'parameter_ubound': 0.6,
'parameter_lbound': 0.05,
'number_of_steps': 3
}
c.plot_comparison_overall_chart(**plot_settings)
assert(len(w) == 0)
def test_bokeh_comparison_plot_dept_size_overall(mgmt_data):
modlist = list([
Model3GenderDiversity(**mgmt_data),
ModelGenderDiversityLinearGrowth(**mgmt_data),
ModelGenderDiversityGrowthForecast(**mgmt_data)
])
modlist[0].init_default_hiring_rate()
modlist[1].init_default_hiring_rate()
modlist[1].init_growth_rate(0.01)
modlist[2].init_default_hiring_rate()
modlist[2].init_growth_rate([73, 78, 83, 88])
c = Comparison(modlist)
plot_settings = {
'plottype':
'department size',
'intervals':
'empirical',
'number_of_runs':
100,
# number simulations to average over
'target':
0.25,
# target percentage of women in the department
# Main plot settings
'xlabel':
'Years',
'ylabel':
'Department Size',
'title':
'Department Size',
'line_width':
2,
'transparency':
0.25,
'model_legend_label': [
'Model 3 No Growth', 'Model 3 Lin Growth(1%/year)',
'Model 3 Forecast(+5/5 year)'
],
'legend_location':
'top_right',
'height_':
height,
'width_':
width,
}
show(c.plot_comparison_overall_chart(**plot_settings))
def test_bokeh_comparison_plot_overall_multiple_models(mgmt_data):
modlist = list([
Model3GenderDiversity(**mgmt_data),
ModelGenderDiversityLinearGrowth(**mgmt_data),
ModelGenderDiversityGrowthForecast(**mgmt_data)
])
modlist[0].init_default_hiring_rate()
modlist[1].init_default_hiring_rate()
modlist[1].init_growth_rate(0.05)
modlist[2].init_default_hiring_rate()
modlist[2].init_growth_rate([73, 78, 83, 88])
c = Comparison(modlist)
plot_settings = {
'plottype':
'gender proportion',
'intervals':
'empirical',
'number_of_runs':
100,
# number simulations to average over
'target':
0.25,
# target percentage of women in the department
# Main plot settings
'xlabel':
'Years',
'ylabel':
'Proportion Women',
'title':
'Change in Proportion Women Overall',
'transparency':
0.25,
'model_legend_label':
['Model 3 No Growth', 'Model 3 Lin Growth', 'Model 3 Forecast']
}
show(c.plot_comparison_overall_chart(**plot_settings))