Revert to older commit to see where the bug is.

grmpy/check/check.py

@@ -46,7 +46,7 @@ def check_est_init_dict(dict_):
         raise UserError(msg)
 
 
-def check_sim_distribution(dict_):
+def check_sim_init_file(dict_):
     """This function checks if the specified initialization file meets the requirements
     for the simulation process.
     """

grmpy/estimate/estimate.py

@@ -18,6 +18,7 @@ def fit(init_file, semipar=False):
     or, alternatively, via the semiparametric method of
     local instrumental variables (LIV).
     """
+
     # Load the estimation file
     check_presence_init(init_file)
     dict_ = read(init_file, semipar)

grmpy/estimate/estimate_output.py

@@ -243,7 +243,6 @@ def simulate_estimation(rslt):
 
     # Distribute information
     seed = rslt["SIMULATION"]["seed"]
-
     # Determine parametrization and read in /simulate observables
     start, finish = process_results(rslt)
     data_frames = []

grmpy/estimate/estimate_par.py

@@ -1,31 +1,26 @@
 """
 The module provides auxiliary functions for the estimation process.
 """
+
 import copy
 
 import numpy as np
-from random import randint
 import statsmodels.api as sm
 from scipy.stats import t, norm
 from scipy.optimize import minimize
 from numpy.linalg import LinAlgError
 from statsmodels.tools.numdiff import approx_hess_cs
 from statsmodels.tools.sm_exceptions import PerfectSeparationError
 
-from grmpy.estimate.estimate_output import print_logfile
 from grmpy.estimate.estimate_output import write_comparison
+from grmpy.estimate.estimate_output import print_logfile
 
 from grmpy.check.check import UserError, check_start_values
 
 
 def par_fit(dict_, data):
     """The function estimates the coefficients of the simulated data set."""
-    # Set seed
-    if "SIMULATION" not in dict_ or "seed" not in dict_["SIMULATION"]:
-        seed_ = randint(0, 9999)
-        np.random.seed(seed_)
-    else:
-        np.random.seed(dict_["SIMULATION"]["seed"])
+    np.random.seed(dict_["SIMULATION"]["seed"])
 
     _, X1, X0, Z1, Z0, Y1, Y0 = process_data(data, dict_)
 
@@ -43,7 +38,6 @@ def par_fit(dict_, data):
     dict_["AUX"]["criteria"] = calculate_criteria(dict_, X1, X0, Z1, Z0, Y1, Y0, x0)
     dict_["AUX"]["starting_values"] = backward_transformation(x0)
     rslt_dict = bfgs_dict()
-
     if opts["maxiter"] == 0:
         rslt = adjust_output(None, dict_, x0, X1, X0, Z1, Z0, Y1, Y0, rslt_dict)
     else:
@@ -57,20 +51,16 @@ def par_fit(dict_, data):
         rslt = adjust_output(
             opt_rslt, dict_, opt_rslt["x"], X1, X0, Z1, Z0, Y1, Y0, rslt_dict
         )
-
     # Print Output files
     print_logfile(dict_, rslt)
 
-    if "SIMULATION" in dict_:
-        if "comparison" in dict_["ESTIMATION"].keys():
-            if dict_["ESTIMATION"]["comparison"] == 0:
-                pass
-            else:
-                write_comparison(data, rslt)
+    if "comparison" in dict_["ESTIMATION"].keys():
+        if dict_["ESTIMATION"]["comparison"] == 0:
+            pass
         else:
             write_comparison(data, rslt)
     else:
-        rslt.update({"ESTIMATION": {"seed": seed_}})
+        write_comparison(data, rslt)
 
     return rslt
 
@@ -370,17 +360,6 @@ def adjust_output(opt_rslt, init_dict, x0, X1, X0, Z1, Z0, Y1, Y0, dict_=None):
     rslt["CHOICE"]["p_values"] = np.array(rslt["AUX"]["p_values"][num_untreated:-4])
     rslt["CHOICE"]["t_values"] = np.array(rslt["AUX"]["t_values"][num_untreated:-4])
 
-    if "DIST" not in rslt:
-        rslt.update({"DIST": {"params": None}})
-        rslt.update({"DIST": {"starting_values": None}})
-        rslt.update({"DIST": {"order": None}})
-        rslt.update({"DIST": {"standard_errors": None}})
-        rslt.update({"DIST": {"confidence_intervals": None}})
-        rslt.update({"DIST": {"p_values": None}})
-        rslt.update({"DIST": {"t_values": None}})
-    else:
-        pass
-
     rslt["DIST"]["params"] = np.array(x[-4:])
     rslt["DIST"]["starting_values"] = np.array(init_dict["AUX"]["starting_values"][-4:])
 
@@ -391,7 +370,6 @@ def adjust_output(opt_rslt, init_dict, x0, X1, X0, Z1, Z0, Y1, Y0, dict_=None):
     )
     rslt["DIST"]["p_values"] = np.array(rslt["AUX"]["p_values"][-4:])
     rslt["DIST"]["t_values"] = np.array(rslt["AUX"]["t_values"][-4:])
-
     for subkey in [
         "num_covars_choice",
         "num_covars_treated",

grmpy/read/read.py

@@ -48,7 +48,7 @@ def check_append_constant(init_file, dict_, data, semipar=False):
     """
     if (
         "const" not in data
-        or np.array_equal(np.asarray(data.iloc[:, 0]), np.ones(len(data))) is False
+        and np.array_equal(np.asarray(data.iloc[:, 0]), np.ones(len(data))) is False
     ):
         dict_ = read(init_file, semipar, include_constant=True)
         data.insert(0, "const", 1.0)

grmpy/read/read_auxiliary.py

@@ -15,7 +15,7 @@ def create_attr_dict_est(init_dict, semipar=False, include_constant=False):
         else:
             pass
 
-        init_dict = read_keys(init_dict, init_values)
+        init_dict = read_keys_semipar(init_dict, init_values)
 
     # semipar is False
     else:
@@ -24,7 +24,7 @@ def create_attr_dict_est(init_dict, semipar=False, include_constant=False):
         else:
             pass
 
-        init_dict = read_keys(init_dict, init_values)
+        init_dict = read_keys_par(init_dict, init_values)
 
     init_dict = provide_auxiliary_information(init_dict, init_values)
 
@@ -38,7 +38,7 @@ def create_attr_dict_sim(init_dict):
     init_dict["AUX"] = {"init_values"}
     init_values = []
 
-    init_dict = read_keys_sim(init_dict, init_values)
+    init_dict = read_keys_par(init_dict, init_values)
     init_dict = provide_auxiliary_information(init_dict, init_values)
 
     return init_dict
@@ -70,38 +70,38 @@ def add_constant(init_dict, semipar=False):
     return init_dict
 
 
-def read_keys(init_dict, init_values):
+def read_keys_par(init_dict, init_values):
     """This function reads the information provided by the
-    ["TREATED", "UNTREATED", "CHOICE"] keys for
-    semiparametric estimation.
+    ["TREATED", "UNTREATED", "CHOICE", "DIST"] keys for
+    the simulation and parametric estimation.
     """
-    for key in ["TREATED", "UNTREATED", "CHOICE"]:
+    for key in ["TREATED", "UNTREATED", "CHOICE", "DIST"]:
         if "params" in init_dict[key].keys():
             init_dict[key]["params"] = np.array(init_dict[key]["params"])
             init_values += list(init_dict[key]["params"])
         else:
             init_values += [0.0] * len(init_dict[key]["order"])
 
+    if np.all(init_dict["DIST"]["params"] == 0):
+        init_dict["DETERMINISTIC"] = True
+    else:
+        init_dict["DETERMINISTIC"] = False
+
     return init_dict
 
 
-def read_keys_sim(init_dict, init_values):
+def read_keys_semipar(init_dict, init_values):
     """This function reads the information provided by the
-    ["TREATED", "UNTREATED", "CHOICE", "DIST"] keys for
-    the simulation and parametric estimation.
+    ["TREATED", "UNTREATED", "CHOICE"] keys for
+    semiparametric estimation.
     """
-    for key in ["TREATED", "UNTREATED", "CHOICE", "DIST"]:
+    for key in ["TREATED", "UNTREATED", "CHOICE"]:
         if "params" in init_dict[key].keys():
             init_dict[key]["params"] = np.array(init_dict[key]["params"])
             init_values += list(init_dict[key]["params"])
         else:
             init_values += [0.0] * len(init_dict[key]["order"])
 
-    if np.all(init_dict["DIST"]["params"] == 0):
-        init_dict["DETERMINISTIC"] = True
-    else:
-        init_dict["DETERMINISTIC"] = False
-
     return init_dict
 
 

grmpy/simulate/simulate_auxiliary.py

@@ -29,15 +29,11 @@ def simulate_covariates(init_dict):
     X[labels[0]] = 1.0
 
     # Include binary variables
-    if "VARTYPES" in init_dict:
-        for variable in init_dict["VARTYPES"]:
-            if isinstance(init_dict["VARTYPES"][variable], list):
-                X[variable] = np.random.binomial(
-                    1, init_dict["VARTYPES"][variable][1], size=num_agents
-                )
-    else:
-        pass
-
+    for variable in init_dict["VARTYPES"]:
+        if isinstance(init_dict["VARTYPES"][variable], list):
+            X[variable] = np.random.binomial(
+                1, init_dict["VARTYPES"][variable][1], size=num_agents
+            )
     return X
 
 

grmpy/test/test_integration.py

@@ -8,7 +8,7 @@
 
 from grmpy.check.auxiliary import check_special_conf
 from grmpy.check.check import (
-    check_sim_distribution,
+    check_sim_init_file,
     check_sim_init_dict,
     check_start_values,
 )
@@ -155,7 +155,7 @@ def test7():
         constr["AGENTS"], constr["DETERMINISTIC"] = 1000, True
         generate_random_dict(constr)
         dict_ = read("test.grmpy.yml")
-        pytest.raises(UserError, check_sim_distribution, dict_)
+        pytest.raises(UserError, check_sim_init_file, dict_)
         pytest.raises(UserError, fit, "test.grmpy.yml")
 
         generate_random_dict(constr)
@@ -188,15 +188,15 @@ def test7():
     pytest.raises(UserError, simulate, fname_possd)
 
     dict_ = read(fname_zero)
-    pytest.raises(UserError, check_sim_distribution, dict_)
+    pytest.raises(UserError, check_sim_init_file, dict_)
     pytest.raises(UserError, fit, fname_zero)
 
     dict_ = read(fname_vzero)
-    pytest.raises(UserError, check_sim_distribution, dict_)
+    pytest.raises(UserError, check_sim_init_file, dict_)
     pytest.raises(UserError, fit, fname_vzero)
 
     dict_ = read(fname_noparams)
-    pytest.raises(UserError, check_sim_distribution, dict_)
+    pytest.raises(UserError, check_sim_init_file, dict_)
     pytest.raises(UserError, fit, fname_noparams)
 
     dict_ = read(fname_falsespec1)