class TestReturns(object):
""" Class for testing correct parameters are returned from firststage"""
def __init__(self):
self.opts = {'tables': 0, 'reaction': 'simple', 'entrance': 1,
'verbose': 1}
self.bbl = BBL(self.opts['tables'], self.opts['reaction'],
self.opts['entrance'], self.opts['verbose'])
self.bbl._firststage()
self.fs_params = self.bbl.fs_params
def test_demand(self):
tvars = [
'tuition_hat', 'MedianLSAT', 'UndergraduatemedianGPA',
'OverallRank', 'treat', 'RankTreat', 'AL', 'AR', 'WY'
]
expected = np.array([
-0.927284, 19.286605, -203.475186, -1.167404, -22.917197,
0.354199, -1749.431023, -1799.828207, -1875.056317
])
print("Actual")
print(self.fs_params['demand'][tvars])
assert np.allclose(self.fs_params['demand'][tvars], expected)
def test_react(self):
pass
def test_entry(self):
pass
def test_rank(self):
pass
def __init__(self):
opts = {'tables': 0, 'reaction': 'simple', 'entrance': 1, 'verbose': 1}
bbl = BBL(opts)
data = bbl.data
bbl._firststage()
self.fs_params = bbl.fs_params
path = dirname(dirname(dirname(__file__)))
path = join(path, 'Results', 'SecondStage', 'SSEst.npy')
ss_params = np.load(path)
self.countersim = lps.CounterfactualSim(data, self.fs_params,
ss_params, opts)
def __init__(self):
self.opts = {'tables': 0, 'reaction': 'simple', 'entrance': 1,
'verbose': 1}
self.bbl = BBL(self.opts['tables'], self.opts['reaction'],
self.opts['entrance'], self.opts['verbose'])
self.bbl._firststage()
self.fs_params = self.bbl.fs_params
class Second(object):
""" Base class for testing Second Stage classes/methods """
def __init__(self):
self.opts = {'tables': 0, 'reaction': 'simple', 'entrance': 1,
'verbose': 1}
self.bbl = BBL(self.opts)
self.data = self.bbl.data
self.fs_params = self.first_stage()
self.sim = ss.Simulation(
fs_params=self.fs_params, ss_params=[],
constants={'n_periods': 2, 'n_sims': 2},
opts=self.opts
)
def first_stage(self):
""" Get first-stage coefficients """
self.bbl._firststage()
return self.bbl.fs_params
def __init__(self):
self.opts = {'tables': 0, 'reaction': 'simple', 'entrance': 1,
'verbose': 1}
self.bbl = BBL(self.opts)
self.data = self.bbl.data
self.fs_params = self.first_stage()
self.sim = ss.Simulation(
fs_params=self.fs_params, ss_params=[],
constants={'n_periods': 2, 'n_sims': 2},
opts=self.opts
)
def driver(opts):
"""
Replicate paper results
Args:
opts: Dict with members:
data: 1 to format data
tables: 1 to recreate latex tables and figures
react: 'simple' to use median quality reaction variables
'full' to use 25, 50, and 75 quantiles
entrance: 1 to estimate with entrance/exit decision
verbose: 1 to print estimation and simulation output to screen
during runtime
nose: 1 to run tests before program
"""
path = dirname(realpath(__file__))
if opts['pylint']:
run_pylint()
return 0
if opts['nose']:
run_nosetests()
return 0
if opts['clean']:
#clear_results()
opts['generate'] = 1 # SS results need to be regenerated now
reset_results = ResetResults(path)
reset_results.reset()
if opts['tables']:
initial_tables(path)
if opts['data']:
data_format()
my_bbl = BBL(opts)
if not my_bbl.data_check():
my_bbl.estimate()
my_bbl.simulate()
if opts['tables']:
post_tables(path)
if opts['latex']:
typeset_latex(path)
if opts['beamer']:
typeset_beamer(path)
print("* ------------------ *")
print("* RUN COMPLETE *")
print("* ------------------ *")
