Merge branch 'master' into kebatt/bootstrapFixes

Author: kbattocchi

README.md

@@ -91,10 +91,12 @@ Such questions arise frequently in customer segmentation (what is the effect of
 
 # News
 
-**March 6, 2020:** Release v0.7.0, see release notes [here](https://github.com/Microsoft/EconML/releases/tag/v0.7.0)
+**September 4, 2020:** Release v0.8.0b1, see release notes [here](https://github.com/Microsoft/EconML/releases/tag/v0.8.0b1)
 
 <details><summary>Previous releases</summary>
 
+**March 6, 2020:** Release v0.7.0, see release notes [here](https://github.com/Microsoft/EconML/releases/tag/v0.7.0)
+
 **February 18, 2020:** Release v0.7.0b1, see release notes [here](https://github.com/Microsoft/EconML/releases/tag/v0.7.0b1)
 
 **January 10, 2020:** Release v0.6.1, see release notes [here](https://github.com/Microsoft/EconML/releases/tag/v0.6.1)

econml/dml.py

@@ -50,7 +50,7 @@
 from sklearn.utils import check_random_state
 from .cate_estimator import (BaseCateEstimator, LinearCateEstimator,
                              TreatmentExpansionMixin, StatsModelsCateEstimatorMixin,
-                             DebiasedLassoCateEstimatorMixin)
+                             LinearModelFinalCateEstimatorMixin, DebiasedLassoCateEstimatorMixin)
 from .inference import StatsModelsInference, GenericSingleTreatmentModelFinalInference
 from ._rlearner import _RLearner
 from .sklearn_extensions.model_selection import WeightedStratifiedKFold
@@ -296,7 +296,7 @@ def cate_feature_names(self, input_feature_names=None):
             raise AttributeError("Featurizer does not have a method: get_feature_names!")
 
 
-class DMLCateEstimator(_BaseDMLCateEstimator):
+class DMLCateEstimator(LinearModelFinalCateEstimatorMixin, _BaseDMLCateEstimator):
     """
     The base class for parametric Double ML estimators. The estimator is a special
     case of an :class:`._RLearner` estimator, which in turn is a special case

econml/tests/test_dml.py

@@ -106,7 +106,16 @@ def make_random(n, is_discrete, d):
                                 all_infs = [None, 'statsmodels', BootstrapInference(1)]
 
                                 for est, multi, infs in\
-                                    [(LinearDMLCateEstimator(model_y=Lasso(),
+                                    [(DMLCateEstimator(model_y=Lasso(),
+                                                       model_t=model_t,
+                                                       model_final=Lasso(alpha=0.1, fit_intercept=False),
+                                                       featurizer=featurizer,
+                                                       fit_cate_intercept=fit_cate_intercept,
+                                                       discrete_treatment=is_discrete),
+                                      True,
+                                      [None] +
+                                      ([BootstrapInference(n_bootstrap_samples=20)] if not is_discrete else [])),
+                                     (LinearDMLCateEstimator(model_y=Lasso(),
                                                              model_t='auto',
                                                              featurizer=featurizer,
                                                              fit_cate_intercept=fit_cate_intercept,
@@ -167,8 +176,7 @@ def make_random(n, is_discrete, d):
                                             eff = est.effect(X, T0=T0, T1=T)
                                             self.assertEqual(shape(eff), effect_shape)
 
-                                            if isinstance(est, LinearDMLCateEstimator) or\
-                                                    isinstance(est, SparseLinearDMLCateEstimator):
+                                            if not isinstance(est, KernelDMLCateEstimator):
                                                 self.assertEqual(shape(est.coef_), coef_shape)
                                                 if fit_cate_intercept:
                                                     self.assertEqual(shape(est.intercept_), intercept_shape)
@@ -185,10 +193,7 @@ def make_random(n, is_discrete, d):
                                                                  (2,) + const_marginal_effect_shape)
                                                 self.assertEqual(shape(est.effect_interval(X, T0=T0, T1=T)),
                                                                  (2,) + effect_shape)
-                                                if (isinstance(est,
-                                                               LinearDMLCateEstimator) or
-                                                    isinstance(est,
-                                                               SparseLinearDMLCateEstimator)):
+                                                if not isinstance(est, KernelDMLCateEstimator):
                                                     self.assertEqual(shape(est.coef__interval()),
                                                                      (2,) + coef_shape)
                                                     if fit_cate_intercept:
@@ -263,10 +268,7 @@ def make_random(n, is_discrete, d):
                                                 marg_effect_inf.population_summary()._repr_html_()
 
                                                 # test coef__inference and intercept__inference
-                                                if (isinstance(est,
-                                                               LinearDMLCateEstimator) or
-                                                    isinstance(est,
-                                                               SparseLinearDMLCateEstimator)):
+                                                if not isinstance(est, KernelDMLCateEstimator):
                                                     if X is None:
                                                         cm = pytest.raises(AttributeError)
                                                     else:

setup.cfg

@@ -7,11 +7,11 @@ max-line-length=119
 convention=numpy
 
 [build_sphinx]
-version = 0.7.0
+version = 0.8.0b1
 
 [metadata]
 name = econml
-version = 0.7.0
+version = 0.8.0b1
 author = Microsoft Corporation
 description = This package contains several methods for calculating Conditional Average Treatment Effects
 long_description = file: README.md
@@ -59,6 +59,7 @@ tests_require =
     pytest-cov
     jupyter
     nbconvert < 6
+    traitlets < 5; python_version <= '3.5'
     seaborn
     lightgbm
     dowhy