cleanup branch: test, file structure, docstring and linting

Signed-off-by: kgao <kevin.leo.gao@gmail.com>

econml/tests/test_federated_learning.py

@@ -5,10 +5,11 @@
 import unittest
 from econml.dml import LinearDML
 from econml.inference import StatsModelsInference
-from econml.sklearn_extensions.federated_learning import FederatedLearner
+from econml.federated_learning import FederatedEstimator
 
 
 class FunctionRegressor:
+    """A simple model that ignores the data it is fitted on, always just using the specified function to predict"""
     def __init__(self, func):
         self.func = func
 
@@ -20,7 +21,20 @@ def predict(self, X):
 
 
 class TestFederatedLearning(unittest.TestCase):
+    """
+    A set of unit tests for the FederatedLearner class.
 
+    These tests check various scenarios of splitting, aggregation, and comparison
+    between FederatedLearner and individual LinearDML estimators.
+
+    Parameters
+    ----------
+    None
+
+    Returns
+    -------
+    None
+    """
     def test_splitting_works(self):
 
         num_samples = 1000
@@ -69,15 +83,19 @@ def test_splitting_works(self):
                                                 sample_weight=weights, freq_weight=freq_weights, sample_var=sample_var,
                                                 inference=StatsModelsInference(cov_type=cov_type))
                                     est_h1.fit(Y1, T1, X=X1, W=W1,
-                                               sample_weight=weights1, freq_weight=freq_weights1, sample_var=sample_var1,
+                                               sample_weight=weights1,
+                                               freq_weight=freq_weights1,
+                                               sample_var=sample_var1,
                                                inference=StatsModelsInference(cov_type=cov_type))
                                     est_h2.fit(Y2, T2, X=X2, W=W2,
-                                               sample_weight=weights2, freq_weight=freq_weights2, sample_var=sample_var2,
+                                               sample_weight=weights2,
+                                               freq_weight=freq_weights2,
+                                               sample_var=sample_var2,
                                                inference=StatsModelsInference(cov_type=cov_type))
 
-                                    est_fed1 = FederatedLearner([est_all])
+                                    est_fed1 = FederatedEstimator([est_all])
 
-                                    est_fed2 = FederatedLearner([est_h1, est_h2])
+                                    est_fed2 = FederatedEstimator([est_h1, est_h2])
 
                                     np.testing.assert_allclose(est_fed1.model_final_._param,
                                                                est_fed2.model_final_._param)