def run_test(cls, obj, model_type, verbose=False):
if verbose:
print("Calling fit with `model_type = '%s'`..." % (model_type, ),
end="")
sys.stdout.flush()
with warnings.catch_warnings():
warnings.filterwarnings("ignore",
category=PendingDeprecationWarning)
warnings.filterwarnings("ignore", category=LineSearchWarning)
try:
fit(
X=obj.X,
Y=obj.Y,
model_type=model_type,
treated_units=obj.treated_units
if model_type in ("retrospective", "prospective",
"prospective-restricted") else None,
# KWARGS:
print_path=False,
progress=verbose,
grid_length=5,
min_iter=-1,
tol=1,
verbose=0,
)
if verbose:
print("DONE")
except LineSearchWarning:
pass
except PendingDeprecationWarning:
pass
except Exception as exc:
print("Failed with %s: %s" %
(exc.__class__.__name__, exc.message))
def run_test(cls,
obj,
model_type,
verbose=False,
w_pen_inner=False,
match_space_maker=None):
"""
main test runner
"""
if verbose:
print("Calling fit with `model_type = '%s'`..." % (model_type, ),
end="")
sys.stdout.flush()
# Catch the LineSearchWarning silently, but allow others
with warnings.catch_warnings():
warnings.filterwarnings("ignore",
category=PendingDeprecationWarning)
warnings.filterwarnings("ignore", category=LineSearchWarning)
try:
fit(
features=obj.X,
targets=obj.Y,
model_type=model_type,
treated_units=obj.treated_units
if model_type in ("retrospective", "prospective",
"prospective-restricted") else None,
# KWARGS:
print_path=False,
stopping_rule=1,
progress=verbose,
grid_length=5,
min_iter=-1,
tol=1,
verbose=0,
w_pen_inner=w_pen_inner,
match_space_maker=match_space_maker,
)
if verbose:
print("DONE")
except LineSearchWarning:
pass
except PendingDeprecationWarning:
pass
except Exception as exc: # pytlint: disable=broad-exception
print(
"Failed with {}({})/n{}\n=================================="
.format(exc.__class__.__name__,
getattr(exc, "message",
""), traceback.format_exc()))
raise exc
def test0s(self):
N1, N0_sim, N0_not = 1, 50, 50
N0 = N0_sim + N0_not
#N = N1 + N0
treated_units = range(N1)
#control_units = range(N1, N)
T0, T1 = 2, 1
#T = T0 + T1 # unused
te = 2
#configs = [[[1, 0, 0], [0, 1, 1]],
# [[0, 1, 0], [1, 0, 1]],
# [[1, 0, 2], [0, 1, 1]],
# [[0, 1, 2], [1, 0, 1]]]
#configs = [[[2, 1, 0], [1, 2, 1]],
# [[1, 2, 0], [2, 1, 1]],
# [[2, 1, 2], [1, 2, 1]],
# [[1, 2, 2], [2, 1, 1]]]
configs = [[[1, 2, 2], [2, 1, 1]],
[[1, 2, 3], [2, 1, 1]]]
for config in configs:
proto_sim = np.array(config[0], ndmin=2)
proto_not = np.array(config[1], ndmin=2)
proto_tr = proto_sim + np.hstack((np.zeros((1, T0)), np.full((1, T1), te)))
Y1 = np.matmul(np.ones((N1, 1)), proto_tr)
Y0_sim = np.matmul(np.ones((N0_sim, 1)), proto_sim)
Y0_not = np.matmul(np.ones((N0_not, 1)), proto_not)
Y = np.vstack((Y1, Y0_sim, Y0_not))
with warnings.catch_warnings():
warnings.simplefilter("ignore")
fit_res = fit(Y[:, :T0], Y[:, T0:], treated_units, model_type="retrospective",
stopping_rule=4, progress=False, verbose=0, print_path=False)
Y_sc = fit_res.predict(Y)
treated_diff = (Y - Y_sc)[0, :]
print("V: %s. Treated diff: %s" % (np.diag(fit_res.V), treated_diff))
def fitSparseSC_wrapper(Y_pre, Y_post, treated_units):
return (fit(X=Y_pre,
Y=Y_post,
treated_units=treated_units,
model_type="retrospective",
print_path=False,
progress=False,
verbose=0))
