feat: SHAP support for symmetric CatBoost models (#1619)

* chore: remove redundant SHAP checks

* fix: convert_model is a staticmethod

* fix: Create GBTDAALRegressor if necessary

* Add SHAP kwargs to GBTDAALRegressor

* fix SHAP support for catboost symmetric trees

* Skip SHAP calculation for a month at a time

* Fix is_classification property

* Update expected warning message in test

* Add GBTDAALClassifier fallback

* fix version check

* Disable CatBoost SHAP checks for Jan

daal4py/mb/model_builders.py

@@ -283,8 +283,16 @@ def predict_proba(self, X):
 
 
 def convert_model(model):
-    gbm = GBTDAALModel()
-    gbm._convert_model(model)
+    try:
+        gbm = GBTDAALModel()
+        gbm._convert_model(model)
+    except TypeError as err:
+        if "Only GBTDAALRegressor can be created" in str(err):
+            gbm = d4p.sklearn.ensemble.GBTDAALRegressor.convert_model(model)
+        elif "Only GBTDAALClassifier can be created" in str(err):
+            gbm = d4p.sklearn.ensemble.GBTDAALClassifier.convert_model(model)
+        else:
+            raise
 
     gbm._is_regression = isinstance(gbm.daal_model_, d4p.gbt_regression_model)
 

daal4py/sklearn/ensemble/GBTDAAL.py

@@ -235,6 +235,7 @@ def predict_log_proba(self, X):
 
         return proba
 
+    @staticmethod
     def convert_model(model):
         gbm = GBTDAALClassifier()
         gbm._convert_model(model)
@@ -292,7 +293,7 @@ def fit(self, X, y):
         return self
 
     @run_with_n_jobs
-    def predict(self, X):
+    def predict(self, X, pred_contribs=False, pred_interactions=False):
         # Input validation
         if not self.allow_nan_:
             X = check_array(X, dtype=[np.single, np.double])
@@ -303,8 +304,9 @@ def predict(self, X):
         check_is_fitted(self, ["n_features_in_"])
 
         fptype = getFPType(X)
-        return self._predict_regression(X, fptype)
+        return self._predict_regression(X, fptype, pred_contribs, pred_interactions)
 
+    @staticmethod
     def convert_model(model):
         gbm = GBTDAALRegressor()
         gbm._convert_model(model)