Remove deprecated functionality

Author: kbattocchi

doc/spec/comparison.rst

@@ -7,7 +7,7 @@ Detailed estimator comparison
 | Estimator                                   | | Treatment  | | Requires   | | Delivers Conf. | | Linear    | | Linear        | | Mulitple | | Multiple   | | High-Dimensional |
 |                                             | | Type       | | Instrument | | Intervals      | | Treatment | | Heterogeneity | | Outcomes | | Treatments | | Features         |
 +=============================================+==============+==============+==================+=============+=================+============+==============+====================+
-| :class:`.NonparametricTwoStageLeastSquares` | Any          | Yes          |                  | Yes         | Assumed         | Yes        | Yes          |                    |
+| :class:`.SieveTSLS`                         | Any          | Yes          |                  | Yes         | Assumed         | Yes        | Yes          |                    |
 +---------------------------------------------+--------------+--------------+------------------+-------------+-----------------+------------+--------------+--------------------+
 | :class:`.DeepIV`                            | Any          | Yes          |                  |             |                 | Yes        | Yes          |                    |
 +---------------------------------------------+--------------+--------------+------------------+-------------+-----------------+------------+--------------+--------------------+

doc/spec/estimation/dr.rst

@@ -306,10 +306,10 @@ Below we give a brief description of each of these classes:
             point = est.effect(X, T0=T0, T1=T1)
             lb, ub = est.effect_interval(X, T0=T0, T1=T1, alpha=0.05)
 
-          This method is related to the :class:`.DiscreteTreatmentOrthoForest` and you can check [Oprescu2019]_ for more technical details;
+          This method is related to the :class:`.DROrthoForest` and you can check [Oprescu2019]_ for more technical details;
           the main difference being how the nuisance models are being constructed for the CATE estimation at some
           target :math:`X=x`. Check out :ref:`Forest Estimators User Guide <orthoforestuserguide>` for more information on forest based CATE models and other
-          alternatives to the :class:`.ForestDML`.
+          alternatives to the :class:`.CausalForestDML`.
 
 
 Usage FAQs

econml/__init__.py

@@ -6,22 +6,17 @@
            'cate_interpreter',
            'causal_forest',
            'data',
-           'deepiv',
            'dml',
            'dr',
-           'drlearner',
            'inference',
            'iv',
            'metalearners',
-           'ortho_forest',
            'orf',
-           'ortho_iv',
            'policy',
            'score',
            'solutions',
            'sklearn_extensions',
            'tree',
-           'two_stage_least_squares',
            'utilities',
            'dowhy',
            '__version__']

econml/bootstrap.py

@@ -35,13 +35,12 @@
 """
 
 from .dml import (DML, LinearDML, SparseLinearDML,
-                  KernelDML, NonParamDML, ForestDML)
+                  KernelDML, NonParamDML)
 from .causal_forest import CausalForestDML
 
 __all__ = ["DML",
            "LinearDML",
            "SparseLinearDML",
            "KernelDML",
            "NonParamDML",
-           "ForestDML",
            "CausalForestDML"]