Improve abstract method coverage approach

Signed-off-by: Keith Battocchi <[email protected]>

Author: kbattocchi

econml/_cate_estimator.py

@@ -117,7 +117,7 @@ def fit(self, *args, inference=None, **kwargs):
         self
 
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     def _wrap_fit(m):
         @wraps(m)
@@ -160,7 +160,7 @@ def effect(self, X=None, *, T0, T1):
             Note that when Y is a vector rather than a 2-dimensional array, the corresponding
             singleton dimension will be collapsed (so this method will return a vector)
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     @abc.abstractmethod
     def marginal_effect(self, T, X=None):
@@ -185,7 +185,7 @@ def marginal_effect(self, T, X=None):
             the corresponding singleton dimensions in the output will be collapsed
             (e.g. if both are vectors, then the output of this method will also be a vector)
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     def ate(self, X=None, *, T0, T1):
         """
@@ -355,7 +355,7 @@ def effect_interval(self, X=None, *, T0=0, T1=1, alpha=0.05):
         lower, upper : tuple(type of :meth:`effect(X, T0, T1)<effect>`, type of :meth:`effect(X, T0, T1))<effect>` )
             The lower and the upper bounds of the confidence interval for each quantity.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @_defer_to_inference
     def effect_inference(self, X=None, *, T0=0, T1=1):
@@ -379,7 +379,7 @@ def effect_inference(self, X=None, *, T0=0, T1=1):
             can on demand calculate confidence interval, z statistic and p value. It can also output
             a dataframe summary of these inference results.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @_defer_to_inference
     def marginal_effect_interval(self, T, X=None, *, alpha=0.05):
@@ -403,7 +403,7 @@ def marginal_effect_interval(self, T, X=None, *, alpha=0.05):
                              type of :meth:`marginal_effect(T, X)<marginal_effect>` )
             The lower and the upper bounds of the confidence interval for each quantity.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @_defer_to_inference
     def marginal_effect_inference(self, T, X=None):
@@ -425,7 +425,7 @@ def marginal_effect_inference(self, T, X=None):
             can on demand calculate confidence interval, z statistic and p value. It can also output
             a dataframe summary of these inference results.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @_defer_to_inference
     def ate_interval(self, X=None, *, T0, T1, alpha=0.05):
@@ -450,7 +450,7 @@ def ate_interval(self, X=None, *, T0, T1, alpha=0.05):
         lower, upper : tuple(type of :meth:`ate(X, T0, T1)<ate>`, type of :meth:`ate(X, T0, T1))<ate>` )
             The lower and the upper bounds of the confidence interval for each quantity.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @_defer_to_inference
     def ate_inference(self, X=None, *, T0, T1):
@@ -474,7 +474,7 @@ def ate_inference(self, X=None, *, T0, T1):
             can on demand calculate confidence interval, z statistic and p value. It can also output
             a dataframe summary of these inference results.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @_defer_to_inference
     def marginal_ate_interval(self, T, X=None, *, alpha=0.05):
@@ -498,7 +498,7 @@ def marginal_ate_interval(self, T, X=None, *, alpha=0.05):
                              type of :meth:`marginal_ate(T, X)<marginal_ate>` )
             The lower and the upper bounds of the confidence interval for each quantity.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @_defer_to_inference
     def marginal_ate_inference(self, T, X=None):
@@ -520,7 +520,7 @@ def marginal_ate_inference(self, T, X=None):
             can on demand calculate confidence interval, z statistic and p value. It can also output
             a dataframe summary of these inference results.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @property
     def dowhy(self):
@@ -564,7 +564,7 @@ def const_marginal_effect(self, X=None):
             rather than a 2-dimensional array, the corresponding singleton dimensions in the output will be collapsed
             (e.g. if both are vectors, then the output of this method will also be a vector)
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     def effect(self, X=None, *, T0, T1):
         """
@@ -697,7 +697,7 @@ def const_marginal_effect_interval(self, X=None, *, alpha=0.05):
                              type of :meth:`const_marginal_effect(X)<const_marginal_effect>` )
             The lower and the upper bounds of the confidence interval for each quantity.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @BaseCateEstimator._defer_to_inference
     def const_marginal_effect_inference(self, X=None):
@@ -717,7 +717,7 @@ def const_marginal_effect_inference(self, X=None):
             can on demand calculate confidence interval, z statistic and p value. It can also output
             a dataframe summary of these inference results.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     def const_marginal_ate(self, X=None):
         """
@@ -759,7 +759,7 @@ def const_marginal_ate_interval(self, X=None, *, alpha=0.05):
                              type of :meth:`const_marginal_ate(X)<const_marginal_ate>` )
             The lower and the upper bounds of the confidence interval for each quantity.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @BaseCateEstimator._defer_to_inference
     def const_marginal_ate_inference(self, X=None):
@@ -779,20 +779,19 @@ def const_marginal_ate_inference(self, X=None):
             can on demand calculate confidence interval, z statistic and p value. It can also output
             a dataframe summary of these inference results.
         """
-        pass
 
     def marginal_ate(self, T, X=None):
         return np.mean(self.marginal_effect(T, X=X), axis=0)
     marginal_ate.__doc__ = BaseCateEstimator.marginal_ate.__doc__
 
     @BaseCateEstimator._defer_to_inference
     def marginal_ate_interval(self, T, X=None, *, alpha=0.05):
-        pass
+        raise NotImplementedError("Defer to inference")
     marginal_ate_interval.__doc__ = BaseCateEstimator.marginal_ate_interval.__doc__
 
     @BaseCateEstimator._defer_to_inference
     def marginal_ate_inference(self, T, X=None):
-        pass
+        raise NotImplementedError("Defer to inference")
     marginal_ate_inference.__doc__ = BaseCateEstimator.marginal_ate_inference.__doc__
 
     def shap_values(self, X, *, feature_names=None, treatment_names=None, output_names=None, background_samples=100):
@@ -1003,7 +1002,7 @@ def coef__interval(self, *, alpha=0.05):
         lb, ub: tuple(type of :meth:`coef_()<coef_>`, type of :meth:`coef_()<coef_>`)
             The lower and upper bounds of the confidence interval for each quantity.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @BaseCateEstimator._defer_to_inference
     def coef__inference(self):
@@ -1015,7 +1014,7 @@ def coef__inference(self):
         InferenceResults: object
             The inference of the coefficients in the final linear model
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @BaseCateEstimator._defer_to_inference
     def intercept__interval(self, *, alpha=0.05):
@@ -1033,7 +1032,7 @@ def intercept__interval(self, *, alpha=0.05):
         lower, upper: tuple(type of :meth:`intercept_()<intercept_>`, type of :meth:`intercept_()<intercept_>`)
             The lower and upper bounds of the confidence interval.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @BaseCateEstimator._defer_to_inference
     def intercept__inference(self):
@@ -1045,7 +1044,7 @@ def intercept__inference(self):
         InferenceResults: object
             The inference of the intercept in the final linear model
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     def summary(self, alpha=0.05, value=0, decimals=3, feature_names=None, treatment_names=None, output_names=None):
         """ The summary of coefficient and intercept in the linear model of the constant marginal treatment
@@ -1274,7 +1273,7 @@ def coef__interval(self, T, *, alpha=0.05):
         lower, upper: tuple(type of :meth:`coef_(T)<coef_>`, type of :meth:`coef_(T)<coef_>`)
             The lower and upper bounds of the confidence interval for each quantity.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @BaseCateEstimator._defer_to_inference
     def coef__inference(self, T):
@@ -1291,7 +1290,7 @@ def coef__inference(self, T):
         InferenceResults: object
             The inference of the coefficients in the final linear model
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @BaseCateEstimator._defer_to_inference
     def intercept__interval(self, T, *, alpha=0.05):
@@ -1311,7 +1310,7 @@ def intercept__interval(self, T, *, alpha=0.05):
         lower, upper: tuple(type of :meth:`intercept_(T)<intercept_>`, type of :meth:`intercept_(T)<intercept_>`)
             The lower and upper bounds of the confidence interval.
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     @BaseCateEstimator._defer_to_inference
     def intercept__inference(self, T):
@@ -1329,7 +1328,7 @@ def intercept__inference(self, T):
             The inference of the intercept in the final linear model
 
         """
-        pass
+        raise NotImplementedError("Defer to inference")
 
     def summary(self, T, *, alpha=0.05, value=0, decimals=3,
                 feature_names=None, treatment_names=None, output_names=None):

econml/_ortho_learner.py

@@ -469,7 +469,7 @@ def _gen_ortho_learner_model_nuisance(self):
             If the estimator also provides a score method with the same arguments as fit, it will be used to
             calculate scores during training.
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     @abstractmethod
     def _gen_ortho_learner_model_final(self):
@@ -492,7 +492,7 @@ def _gen_ortho_learner_model_final(self):
             then the input ``T`` to both above calls will be the one-hot encoding of the original input ``T``,
             excluding the first column of the one-hot.
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     def _check_input_dims(self, Y, T, X=None, W=None, Z=None, *other_arrays):
         assert shape(Y)[0] == shape(T)[0], "Dimension mis-match!"

econml/_tree_exporter.py

@@ -612,7 +612,7 @@ def _make_dot_exporter(self, *, out_file, feature_names, treatment_names, max_de
             Number of digits of precision for floating point in the values of
             impurity, threshold and value attributes of each node.
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     @abc.abstractmethod
     def _make_mpl_exporter(self, *, title=None, feature_names=None, treatment_names=None, max_depth=None,
@@ -650,7 +650,7 @@ def _make_mpl_exporter(self, *, title=None, feature_names=None, treatment_names=
         fontsize : int
             Fontsize for text
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     def export_graphviz(self, out_file=None, feature_names=None, treatment_names=None,
                         max_depth=None,

econml/cate_interpreter/_interpreters.py

@@ -31,7 +31,7 @@ def interpret(self, cate_estimator, X):
             must be compatible shape-wise with the features used to fit
             the estimator
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
 
 class SingleTreeCateInterpreter(_SingleTreeInterpreter):

econml/dml/_rlearner.py

@@ -295,7 +295,7 @@ def _gen_model_y(self):
                 model_y.fit(X, W, Y, sample_weight=sample_weight)
                 model_y.predict(X, W)
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     @abstractmethod
     def _gen_model_t(self):
@@ -310,7 +310,7 @@ def _gen_model_t(self):
                 model_t.fit(X, W, T, sample_weight=sample_weight)
                 model_t.predict(X, W)
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     @abstractmethod
     def _gen_rlearner_model_final(self):
@@ -326,7 +326,7 @@ def _gen_rlearner_model_final(self):
                                 sample_weight=sample_weight, freq_weight=freq_weight, sample_var=sample_var)
                 model_final.predict(X)
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     def _gen_ortho_learner_model_nuisance(self):
         return _ModelNuisance(self._gen_model_y(), self._gen_model_t())

econml/grf/_base_grf.py

@@ -103,6 +103,7 @@ def __init__(self,
         self.warm_start = warm_start
         self.max_samples = max_samples
 
+    @abstractmethod
     def _get_alpha_and_pointJ(self, X, T, y, **kwargs):
         """ This function must be implemented by child class and given input variables
         X, T, y and any auxiliary variables passed as keyword only, should be calculating
@@ -116,8 +117,9 @@ def _get_alpha_and_pointJ(self, X, T, y, **kwargs):
         J : array of shape (n_samples, n_outputs * n_outputs)
             The J matrix part of the moment equation, flattened in Fortran-contiguous format.
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
+    @abstractmethod
     def _get_n_outputs_decomposition(self, X, T, y, **kwargs):
         """ This function must be implemented by child class and given input variables
         X, T, y and any auxiliary variables passed as keyword only, should return a tuple
@@ -131,7 +133,7 @@ def _get_n_outputs_decomposition(self, X, T, y, **kwargs):
         n_relevant_outputs : int
             The length of the prefix of parameters that we care about (remainder are nuisance)
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     def apply(self, X):
         """

econml/inference/_inference.py

@@ -33,7 +33,7 @@ def fit(self, estimator, *args, **kwargs):
 
         This is called after the estimator's fit.
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     def ate_interval(self, X=None, *, T0=0, T1=1, alpha=0.05):
         return self.effect_inference(X=X, T0=T0, T1=T1).population_summary(alpha=alpha).conf_int_mean()
@@ -738,7 +738,7 @@ def stderr(self):
             the corresponding singleton dimensions in the output will be collapsed
             (e.g. if both are vectors, then the output of this method will also be a vector)
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     @property
     def var(self):
@@ -776,7 +776,7 @@ def conf_int(self, alpha=0.05):
             the corresponding singleton dimensions in the output will be collapsed
             (e.g. if both are vectors, then the output of this method will also be a vector)
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     @abc.abstractmethod
     def pvalue(self, value=0):
@@ -796,7 +796,7 @@ def pvalue(self, value=0):
             the corresponding singleton dimensions in the output will be collapsed
             (e.g. if both are vectors, then the output of this method will also be a vector)
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     def zstat(self, value=0):
         """
@@ -963,7 +963,7 @@ def _expand_outputs(self, n_rows):
         results: InferenceResults
             The expanded results
         """
-        pass
+        raise NotImplementedError("Abstract method")
 
     def translate(self, offset):
         """

setup.cfg

@@ -100,3 +100,7 @@ concurrency =
 source = econml
 omit = econml/tests/*
 relative_files = True
+
+[coverage:report]
+exclude_lines =
+    raise NotImplementedError\("(Abstract method|Defer to inference)"\)