Revert "Revert "Revert "Add ensemble regressor primitives"""

This reverts commit 0a2eb3f.

alpha_automl/pipeline_synthesis/pipeline_builder.py

@@ -41,14 +41,14 @@ def change_default_hyperparams(primitive_object):
 
 def extract_estimators(pipeline_primitives, all_primitives):
     estimators = []
-    estimator_name, estimator_obj = pipeline_primitives.pop()
-    current_primitive_type = all_primitives[estimator_name]['type']
+    classifier_name, classifier_obj = pipeline_primitives.pop()
+    current_primitive_type = all_primitives[classifier_name]['type']
     counter = 0
 
-    while current_primitive_type == 'CLASSIFIER' or current_primitive_type == 'REGRESSOR':
-        estimators.append((f'{estimator_name}-{counter}', estimator_obj))
-        estimator_name, estimator_obj = pipeline_primitives.pop()
-        current_primitive_type = all_primitives[estimator_name]['type']
+    while current_primitive_type == 'CLASSIFIER':
+        estimators.append((f'{classifier_name}-{counter}', classifier_obj))
+        classifier_name, classifier_obj = pipeline_primitives.pop()
+        current_primitive_type = all_primitives[classifier_name]['type']
         counter += 1
 
     return estimators
@@ -101,10 +101,10 @@ def make_primitive_objects(self, primitives):
             if primitive_type == 'SEMISUPERVISED_SELFTRAINER':
                 classifier_obj = pipeline_primitives.pop()[1]
                 primitive_object = create_object(primitive_name, {'base_estimator': classifier_obj})
-            elif primitive_type == 'CLASSIFICATION_SINGLE_ENSEMBLER' or primitive_type == 'REGRESSION_SINGLE_ENSEMBLER':
+            elif primitive_type == 'SINGLE_ENSEMBLER':
                 classifier_obj = pipeline_primitives.pop()[1]
                 primitive_object = create_object(primitive_name, {'estimator': classifier_obj})
-            elif primitive_type == 'CLASSIFICATION_MULTI_ENSEMBLER' or primitive_type == 'REGRESSION_MULTI_ENSEMBLER':
+            elif primitive_type == 'MULTI_ENSEMBLER':
                 estimators = extract_estimators(pipeline_primitives, self.all_primitives)
                 primitive_object = create_object(primitive_name, {'estimators': estimators})
             elif self.all_primitives[primitive_name]['origin'] == NATIVE_PRIMITIVE:  # It's an installed primitive

alpha_automl/pipeline_synthesis/setup_search.py

@@ -26,13 +26,7 @@
         "SEMISUPERVISED": 5,
         "NA": 6,
     },
-    "DATA_TYPES": {
-        "TABULAR": 1, 
-        "TEXT": 2, 
-        "IMAGE": 3, 
-        "VIDEO": 4, 
-        "MULTIMODAL": 5
-    },
+    "DATA_TYPES": {"TABULAR": 1, "TEXT": 2, "IMAGE": 3, "VIDEO": 4, "MULTIMODAL": 5},
     "PIPELINE_SIZE": 10,
 }
 
@@ -45,10 +39,12 @@ def search_pipelines(X, y, scoring, splitting_strategy, task_name, time_bound, a
     task_start = datetime.now()
 
     def evaluate_pipeline(primitives):
-        has_repeated_estimators = check_repeated_classifiers(primitives, all_primitives, ensemble_pipelines_hash)
+        has_repeated_classifiers = check_repeated_classifiers(
+            primitives, all_primitives, ensemble_pipelines_hash
+        )
 
-        if has_repeated_estimators:
-            logger.info("Repeated estimators detected in ensembles, ignoring pipeline")
+        if has_repeated_classifiers:
+            logger.info("Repeated classifiers detected in ensembles, ignoring pipeline")
             return None
 
         pipeline = builder.make_pipeline(primitives)
@@ -117,34 +113,37 @@ def update_config(task_name, metric, grammar, metadata):
     return config
 
 
-def check_repeated_classifiers(pipeline_primitives, all_primitives, ensemble_pipelines_hash):
-    # We should rename this function to check_repeated_estimators, but loading checkpoints raise errors
-    # Verify if the estimators are repeated in the ensembles (regardless of the order)
-    estimators = []
+def check_repeated_classifiers(
+    pipeline_primitives, all_primitives, ensemble_pipelines_hash
+):
+    # Verify if the classifiers are repeated in the ensembles (regardless of the order)
+    classifiers = []
     pipeline_hash = ""
     has_ensemble_primitive = False
-    has_repeated_estimators = False
+    has_repeated_classifiers = False
 
     for primitive_name in pipeline_primitives:
         primitive_type = all_primitives[primitive_name]["type"]
 
-        if primitive_type == "CLASSIFIER" or primitive_type == "REGRESSOR":
-            estimators.append(primitive_name)
-        elif primitive_type == "CLASSIFICATION_MULTI_ENSEMBLER" or primitive_type == "REGRESSION_MULTI_ENSEMBLER":
+        if primitive_type == "CLASSIFIER":
+            classifiers.append(primitive_name)
+        elif primitive_type == "MULTI_ENSEMBLER":
             has_ensemble_primitive = True
             pipeline_hash += primitive_name
-            if len(estimators) != len(set(estimators)):  # All estimators should be different
-                has_repeated_estimators = True
+            if len(classifiers) != len(
+                set(classifiers)
+            ):  # All classifiers should be different
+                has_repeated_classifiers = True
         else:
             pipeline_hash += primitive_name
 
     if not has_ensemble_primitive:
         return False
 
-    if has_repeated_estimators:
+    if has_repeated_classifiers:
         return True
 
-    pipeline_hash += "".join(sorted(estimators))
+    pipeline_hash += "".join(sorted(classifiers))
 
     if pipeline_hash in ensemble_pipelines_hash:
         return True

alpha_automl/resource/base_grammar.bnf

@@ -1,13 +1,12 @@
 S -> CLASSIFICATION_TASK | REGRESSION_TASK | CLUSTERING_TASK | TIME_SERIES_FORECAST_TASK | SEMISUPERVISED_TASK
-CLASSIFICATION_TASK -> IMPUTER ENCODERS FEATURE_SCALER FEATURE_SELECTOR CLASSIFIER CLASSIFICATION_ENSEMBLER
-REGRESSION_TASK -> IMPUTER ENCODERS FEATURE_SELECTOR FEATURE_SCALER REGRESSOR REGRESSION_ENSEMBLER
+CLASSIFICATION_TASK -> IMPUTER ENCODERS FEATURE_SCALER FEATURE_SELECTOR CLASSIFIER ENSEMBLER
+REGRESSION_TASK -> IMPUTER ENCODERS FEATURE_SELECTOR FEATURE_SCALER REGRESSOR
 CLUSTERING_TASK -> IMPUTER ENCODERS FEATURE_SCALER FEATURE_SELECTOR CLUSTERER
 TIME_SERIES_FORECAST_TASK -> IMPUTER TIME_SERIES_FORECASTER | REGRESSION_TASK
 SEMISUPERVISED_TASK -> IMPUTER ENCODERS FEATURE_SCALER SEMISUPERVISED_CLASSIFIER
 NA_TASK -> CLASSIFICATION_TASK | REGRESSION_TASK | SEMISUPERVISED_TASK
 ENCODERS -> TEXT_ENCODER DATETIME_ENCODER CATEGORICAL_ENCODER IMAGE_ENCODER
-CLASSIFICATION_ENSEMBLER -> CLASSIFICATION_SINGLE_ENSEMBLER | CLASSIFIER CLASSIFIER CLASSIFICATION_MULTI_ENSEMBLER | E
-REGRESSION_ENSEMBLER -> REGRESSION_SINGLE_ENSEMBLER | REGRESSOR REGRESSOR REGRESSION_MULTI_ENSEMBLER | E
+ENSEMBLER -> SINGLE_ENSEMBLER | CLASSIFIER CLASSIFIER MULTI_ENSEMBLER | E
 SEMISUPERVISED_CLASSIFIER -> CLASSIFIER SEMISUPERVISED_SELFTRAINER | SEMISUPERVISED_LABELPROPAGATOR
 IMPUTER -> 'primitive_terminal'
 FEATURE_SCALER -> 'primitive_terminal' | 'E'
@@ -16,10 +15,8 @@ TEXT_ENCODER -> 'primitive_terminal'
 CATEGORICAL_ENCODER -> 'primitive_terminal'
 DATETIME_ENCODER -> 'primitive_terminal'
 IMAGE_ENCODER -> 'primitive_terminal'
-CLASSIFICATION_SINGLE_ENSEMBLER -> 'primitive_terminal'
-CLASSIFICATION_MULTI_ENSEMBLER -> 'primitive_terminal'
-REGRESSION_SINGLE_ENSEMBLER -> 'primitive_terminal'
-REGRESSION_MULTI_ENSEMBLER -> 'primitive_terminal'
+SINGLE_ENSEMBLER -> 'primitive_terminal'
+MULTI_ENSEMBLER -> 'primitive_terminal'
 CLASSIFIER -> 'primitive_terminal'
 REGRESSOR -> 'primitive_terminal'
 CLUSTERER -> 'primitive_terminal'

alpha_automl/resource/primitives_hierarchy.json

@@ -92,20 +92,12 @@
         "alpha_automl.builtin_primitives.semisupervised_classifier.SkLabelSpreading",
         "alpha_automl.builtin_primitives.semisupervised_classifier.SkLabelPropagation"
     ],
-    "CLASSIFICATION_SINGLE_ENSEMBLER": [
+    "SINGLE_ENSEMBLER": [
         "sklearn.ensemble.AdaBoostClassifier",
         "sklearn.ensemble.BaggingClassifier"
     ],
-    "CLASSIFICATION_MULTI_ENSEMBLER": [
+    "MULTI_ENSEMBLER": [
         "sklearn.ensemble.StackingClassifier",
         "sklearn.ensemble.VotingClassifier"
-    ],
-    "REGRESSION_SINGLE_ENSEMBLER": [
-        "sklearn.ensemble.AdaBoostRegressor",
-        "sklearn.ensemble.BaggingRegressor"
-    ],
-    "REGRESSION_MULTI_ENSEMBLER": [
-        "sklearn.ensemble.StackingRegressor",
-        "sklearn.ensemble.VotingRegressor"
     ]
 }