def test5():
"""The test checks if the estimation process works properly when maxiter is set to zero."""
for _ in range(10):
constr = dict()
constr['DETERMINISTIC'], constr['MAXITER'] = False, 0
generate_random_dict(constr)
simulate('test.grmpy.ini')
estimate('test.grmpy.ini')
def test5():
"""The test checks if the estimation process works properly when maxiter is set to
zero.
"""
for _ in range(10):
constr = constraints(probability=0.0, maxiter=0)
generate_random_dict(constr)
simulate('test.grmpy.ini')
estimate('test.grmpy.ini')
def test4():
"""The test checks if the estimation process works if the Powell algorithm is specified as
the optimizer option.
"""
for _ in range(5):
constr = constraints(probability=0.0,
agents=10000,
start='init',
optimizer='SCIPY-POWELL')
generate_random_dict(constr)
simulate('test.grmpy.ini')
estimate('test.grmpy.ini')
def test4():
"""The test checks if the estimation process works if the Powell algorithm is specified as
the optimizer option.
"""
for _ in range(5):
constr = dict()
constr['DETERMINISTIC'], constr['AGENTS'], constr[
'start'] = False, 10000, 'init'
constr['optimizer'] = 'SCIPY-Powell'
generate_random_dict(constr)
simulate('test.grmpy.ini')
estimate('test.grmpy.ini')
def test9():
"""This test ensures that the random initialization file generating process, the read in process
and the simulation process works if the constraints function allows for different number of co-
variates for each treatment state and the occurence of cost-benefit shifters."""
for i in range(5):
constr = dict()
constr['DETERMINISTIC'], constr['AGENT'], constr[
'STATE_DIFF'] = False, 1000, True
constr['OVERLAP'] = True
generate_random_dict(constr)
read('test.grmpy.ini')
simulate('test.grmpy.ini')
estimate('test.grmpy.ini')
cleanup()
def test6():
"""Additionally to test5 this test checks if the descriptives file provides the expected
output when maxiter is set to zero and the estimation process uses the initialization file
values as start values.
"""
for _ in range(5):
constr = constraints(probability=0.0,
maxiter=0,
agents=1000,
start='init')
generate_random_dict(constr)
simulate('test.grmpy.ini')
estimate('test.grmpy.ini')
dict_ = read_desc('descriptives.grmpy.txt')
for key_ in ['All', 'Treated', 'Untreated']:
np.testing.assert_equal(len(set(dict_[key_]['Number'])), 1)
np.testing.assert_array_equal(
dict_[key_]['Observed Sample'],
dict_[key_]['Simulated Sample (finish)'])
np.testing.assert_array_equal(
dict_[key_]['Simulated Sample (finish)'],
dict_[key_]['Simulated Sample (start)'])
cleanup()
def test6():
"""Additionally to test5 this test checks if the comparison file provides the expected
output when maxiter is set to zero and the estimation process uses the initialization file
values as start values.
"""
for _ in range(5):
constr = dict()
constr['DETERMINISTIC'], constr['MAXITER'], constr[
'AGENTS'] = False, 0, 10000
constr['START'], constr['SAME_SIZE'] = 'init', True
dict_ = generate_random_dict(constr)
dict_['DIST']['all'][1], dict_['DIST']['all'][5] = 0.0, 1.0
print_dict(dict_)
simulate('test.grmpy.ini')
estimate('test.grmpy.ini')
dict_ = read_desc('comparison.grmpy.txt')
for key_ in ['All', 'Treated', 'Untreated']:
np.testing.assert_equal(len(set(dict_[key_]['Number'])), 1)
np.testing.assert_almost_equal(
dict_[key_]['Observed Sample'],
dict_[key_]['Simulated Sample (finish)'], 0.001)
np.testing.assert_array_almost_equal(
dict_[key_]['Simulated Sample (finish)'],
dict_[key_]['Simulated Sample (start)'], 0.001)
def test11():
"""This test checks if the refactor auxiliary function returns an unchanged init file if the
maximum number of iterations is set to zero.
"""
for _ in range(10):
constr = dict()
constr['DETERMINISTIC'], constr['AGENTS'] = False, 1000
constr['MAXITER'], constr['START'] = 0, 'init'
generate_random_dict(constr)
init_dict = read('test.grmpy.ini')
df = simulate('test.grmpy.ini')
start = start_values(init_dict, df, 'init')
start = backward_transformation(start)
rslt = estimate('test.grmpy.ini')
np.testing.assert_equal(start, rslt['AUX']['x_internal'])
def test3():
"""The test checks if the criteria function value of the simulated and the 'estimated'
sample is equal if both samples include an identical number of individuals.
"""
for _ in range(5):
constr = dict()
constr['DETERMINISTIC'], constr['AGENTS'], constr[
'START'] = False, 1000, 'init'
constr['OPTIMIZER'], constr['SAME_SIZE'] = 'SCIPY-BFGS', True
generate_random_dict(constr)
df1 = simulate('test.grmpy.ini')
rslt = estimate('test.grmpy.ini')
init_dict = read('test.grmpy.ini')
df2 = simulate_estimation(init_dict, rslt)
start = start_values(init_dict, df1, 'init')
criteria = []
for data in [df1, df2]:
criteria += [calculate_criteria(init_dict, data, start)]
np.testing.assert_allclose(criteria[1], criteria[0], rtol=0.1)
def monte_carlo(file, grid_points):
"""This function estimates the ATE for a sample with different correlation structures between U1
and V. Two different strategies for (OLS,LATE) are implemented.
"""
# Define a dictionary with a key for each estimation strategy
effects = {}
for key_ in ['grmpy', 'ols', 'true']:
effects[key_] = []
# Loop over different correlations between V and U_1
for rho in np.linspace(0.00, 0.99, grid_points):
# Readjust the initialization file values to add correlation
model_spec = read(file)
sim_spec = read('reliability.grmpy.ini')
X = [sim_spec['varnames'][j - 1] for j in sim_spec['TREATED']['order']]
update_correlation_structure(model_spec, rho)
# Simulate a Data set and specify exogeneous and endogeneous variables
df_mc = create_data()
endog, exog, exog_ols = df_mc['wage'], df_mc[X], df_mc[['state'] + X]
# Calculate true average treatment effect
ATE = np.mean(df_mc['wage1'] - df_mc['wage0'])
effects['true'] += [ATE]
# Estimate via grmpy
rslt = estimate('reliability.grmpy.ini')
beta_diff = rslt['TREATED']['all'] - rslt['UNTREATED']['all']
stat = np.dot(np.mean(exog), beta_diff)
effects['grmpy'] += [stat]
# Estimate via OLS
ols = sm.OLS(endog, exog_ols).fit()
stat = ols.params[0]
effects['ols'] += [stat]
return effects
def test3():
"""The test checks if the criteria function value of the simulated and the 'estimated'
sample is equal if both samples include an identical number of individuals.
"""
for _ in range(5):
constr = constraints(probability=0.0,
agents=10000,
start='init',
optimizer='SCIPY-BFGS')
dict_ = generate_random_dict(constr)
print_dict(dict_)
df1 = simulate('test.grmpy.ini')
rslt = estimate('test.grmpy.ini')
init_dict = read('test.grmpy.ini')
df2 = simulate_estimation(init_dict, rslt, df1)
start = start_values(init_dict, df1, 'init')
criteria = []
for data in [df1, df2]:
criteria += [calculate_criteria(init_dict, data, start)]
np.testing.assert_allclose(criteria[1], criteria[0], rtol=0.1)
def test12():
"""This test ensures that the tutorial configuration works as intended."""
fname = TEST_RESOURCES_DIR + '/tutorial.grmpy.ini'
simulate(fname)
estimate(fname)
x_neg = [-i for i in x]
x += x_neg
x = np.array(x)
# Create auxiliary parameters
part1 = np.dot(x, np.dot(param_cov, x))
part2 = np.dot(dist_gradients, np.dot(dist_cov, dist_gradients))
# Prepare two lists for storing the values
mte_up = []
mte_d = []
# Combine all auxiliary parameters and calculate the confidence intervals
for counter, i in enumerate(quantiles):
value = part2 * (norm.ppf(i))**2
aux = np.sqrt(part1 + value) / 4
mte_up += [mte[counter] + norm.ppf(0.95) * aux]
mte_d += [mte[counter] - norm.ppf(0.95) * aux]
return mte_up, mte_d
if __name__ == '__main__':
init_dict = read('replication.grmpy.ini')
# Estimate the coefficients
rslt = estimate('replication.grmpy.ini')
# Calculate and plot the marginal treatment effect
data = pd.read_pickle('aer-replication-mock.pkl')
mte = plot_est_mte(rslt, init_dict, data)
#!/usr/bin/env python
""" This module contains a first stab at a reliability test."""
import grmpy
import sys
from grmpy.simulate.simulate import simulate
from grmpy.estimate.estimate import estimate
simulate('test.grmpy.ini')
option = 'true_values'
estimate('test.grmpy.ini', option)
"""This module contains a tutorial illustrating the basic capabilities of the grmpy package.""" import os from grmpy.simulate.simulate import simulate from grmpy.estimate.estimate import estimate f = os.path.dirname(__file__) + '/tutorial.grmpy.ini' simulate(f) rslt = estimate(f)