Add causal analysis solution

README.md

@@ -45,10 +45,12 @@ For information on use cases and background material on causal inference and het
 
 # News
 
-**March 22, 2021:** Release v0.10.0, see release notes [here](https://github.com/Microsoft/EconML/releases/tag/v0.10.0)
+**May 8, 2021:** Release v0.11.0, see release notes [here](https://github.com/Microsoft/EconML/releases/tag/v0.11.0)
 
 <details><summary>Previous releases</summary>
 
+**March 22, 2021:** Release v0.10.0, see release notes [here](https://github.com/Microsoft/EconML/releases/tag/v0.10.0)
+
 **March 11, 2021:** Release v0.9.2, see release notes [here](https://github.com/Microsoft/EconML/releases/tag/v0.9.2)
 
 **March 3, 2021:** Release v0.9.1, see release notes [here](https://github.com/Microsoft/EconML/releases/tag/v0.9.1)

azure-pipelines-steps.yml

@@ -53,6 +53,11 @@ jobs:
     displayName: 'Install graphviz on Linux'
     condition: and(succeeded(), eq(variables['Agent.OS'], 'Linux'))
 
+  # Install OpenMP on Mac to support lightgbm
+  - script: 'brew install libomp'
+    displayName: 'Install OpenMP on Mac'
+    condition: and(succeeded(), eq(variables['Agent.OS'], 'Darwin'))
+
   # Install the package
   - script: 'python -m pip install --upgrade pip && pip install --upgrade setuptools wheel Cython && pip install ${{ parameters.package }}'
     displayName: 'Install dependencies'

econml/__init__.py

@@ -18,11 +18,12 @@
            'ortho_iv',
            'policy',
            'score',
+           'solutions',
            'sklearn_extensions',
            'tree',
            'two_stage_least_squares',
            'utilities',
            'dowhy',
            '__version__']
 
-__version__ = '0.10.0'
+__version__ = '0.11.0'

econml/_cate_estimator.py

@@ -10,7 +10,7 @@
 from warnings import warn
 from .inference import BootstrapInference
 from .utilities import (tensordot, ndim, reshape, shape, parse_final_model_params,
-                        inverse_onehot, Summary, get_input_columns)
+                        inverse_onehot, Summary, get_input_columns, check_input_arrays)
 from .inference import StatsModelsInference, StatsModelsInferenceDiscrete, LinearModelFinalInference,\
     LinearModelFinalInferenceDiscrete, NormalInferenceResults, GenericSingleTreatmentModelFinalInference,\
     GenericModelFinalInferenceDiscrete
@@ -809,6 +809,7 @@ def _postfit(self, Y, T, *args, **kwargs):
             self._set_transformed_treatment_names()
 
     def _expand_treatments(self, X=None, *Ts):
+        X, *Ts = check_input_arrays(X, *Ts)
         n_rows = 1 if X is None else shape(X)[0]
         outTs = []
         for T in Ts:

econml/dml/causal_forest.py

@@ -16,7 +16,8 @@
 from ..sklearn_extensions.model_selection import WeightedStratifiedKFold
 from ..inference import NormalInferenceResults
 from ..inference._inference import Inference
-from ..utilities import add_intercept, shape, check_inputs, _deprecate_positional, cross_product, Summary
+from ..utilities import (add_intercept, shape, check_inputs, check_input_arrays,
+                         _deprecate_positional, cross_product, Summary)
 from ..grf import CausalForest, MultiOutputGRF
 from .._cate_estimator import LinearCateEstimator
 from .._shap import _shap_explain_multitask_model_cate
@@ -643,6 +644,8 @@ def tune(self, Y, T, *, X=None, W=None,
             The tuned causal forest object. This is the same object (not a copy) as the original one, but where
             all parameters of the object have been set to the best performing parameters from the tuning grid.
         """
+        Y, T, X, W, sample_weight, groups = check_input_arrays(Y, T, X, W, sample_weight, groups)
+
         if params == 'auto':
             params = {'max_samples': [.3, .5],
                       'min_balancedness_tol': [.3, .5],
@@ -737,7 +740,6 @@ def fit(self, Y, T, X=None, W=None, *, sample_weight=None, groups=None,
         """
         if X is None:
             raise ValueError("This estimator does not support X=None!")
-        Y, T, X, W = check_inputs(Y, T, X, W=W, multi_output_T=True, multi_output_Y=True)
         return super().fit(Y, T, X=X, W=W,
                            sample_weight=sample_weight, groups=groups,
                            cache_values=cache_values,

econml/inference/_inference.py

@@ -867,6 +867,19 @@ def _expand_outputs(self, n_rows):
         """
         pass
 
+    def translate(self, offset):
+        """
+        Update the results in place by translating by an offset.
+
+        Parameters
+        ----------
+        offset: array-like
+            The offset by which to translate these results
+        """
+        # NOTE: use np.asarray(offset) becuase if offset is a pd.Series direct addition would make the sum
+        #       a Series as well, which would subsequently break summary_frame because flatten isn't supported
+        self.pred = self.pred + np.asarray(offset)
+
 
 class NormalInferenceResults(InferenceResults):
     """
@@ -1081,6 +1094,14 @@ def _expand_outputs(self, n_rows):
         return EmpiricalInferenceResults(self.d_t, self.d_y, pred, pred_dist, self.inf_type, self.fname_transformer,
                                          self.feature_names, self.output_names, self.treatment_names)
 
+    def translate(self, other):
+        # offset preds
+        super().translate(other)
+        # offset the distribution, too
+        self.pred_dist = self.pred_dist + np.asarray(other)
+
+    translate.__doc__ = InferenceResults.translate.__doc__
+
 
 class PopulationSummaryResults:
     """

econml/solutions/causal_analysis/__init__.py

@@ -0,0 +1,6 @@
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# Licensed under the MIT License.
+
+from ._causal_analysis import CausalAnalysis
+
+__all__ = ["CausalAnalysis"]