temp

VIDA-NYU · Jul 19, 2024 · a00caec · a00caec
1 parent 8e13f02
commit a00caec
Show file tree

Hide file tree

Showing 6 changed files with 14 additions and 13 deletions.
diff --git a/alpha_automl/automl_api.py b/alpha_automl/automl_api.py
@@ -8,7 +8,7 @@
 from alpha_automl.automl_manager import AutoMLManager
 from alpha_automl.scorer import make_scorer, make_splitter, make_str_metric, get_sign_sorting
 from alpha_automl.utils import make_d3m_pipelines, hide_logs, get_start_method, check_input_for_multiprocessing, \
-    setup_output_folder, SemiSupervisedSplitter, SemiSupervisedLabelEncoder, write_pipeline_code_as_pyfile
+    setup_output_folder, SemiSupervisedSplitter, SemiSupervisedLabelEncoder, write_pipeline_code_as_pyfile, sample_dataset
 from alpha_automl.visualization import plot_comparison_pipelines
 from alpha_automl.pipeline_serializer import PipelineSerializer
 
@@ -112,6 +112,8 @@ def fit(self, X, y):
         self.leaderboard = pd.DataFrame(leaderboard_data, columns=['ranking', 'pipeline', self.metric])
 
         best_pipeline_id = PIPELINE_PREFIX + '1'
+
+        X, y, _ = sample_dataset(X, y, 2000000, self.task_type)
         self._fit(X, y, best_pipeline_id)
 
     def predict(self, X):

diff --git a/alpha_automl/automl_manager.py b/alpha_automl/automl_manager.py
@@ -14,7 +14,7 @@
 INCLUDE_PRIMITIVES = []
 NEW_PRIMITIVES = {}
 SPLITTING_STRATEGY = 'holdout'
-SAMPLE_SIZE = 2000
+SAMPLE_SIZE = 100000
 MAX_RUNNING_PROCESSES = multiprocessing.cpu_count()
 
 logger = logging.getLogger(__name__)

diff --git a/alpha_automl/pipeline_search/agent_lab.py b/alpha_automl/pipeline_search/agent_lab.py
@@ -143,6 +143,7 @@ def dump_result_to_json(primitives, task_start, score, output_folder=None):
     if primitives in data.values():
         return
     data[score] = primitives
+    logger.info(f"{timestamp}-Pipeline found with score {score}: \n  {primitives}")
 
     # Write unique elements to output file
     with open(output_path, "w") as f:

diff --git a/alpha_automl/resource/base_grammar.bnf b/alpha_automl/resource/base_grammar.bnf
@@ -1,5 +1,5 @@
 S -> CLASSIFICATION_TASK | REGRESSION_TASK | CLUSTERING_TASK | TIME_SERIES_FORECAST_TASK | SEMISUPERVISED_TASK | NA_TASK
-CLASSIFICATION_TASK -> IMPUTER ENCODERS FEATURE_GENERATOR FEATURE_SCALER FEATURE_SELECTOR CLASSIFIER CLASSIFICATION_ENSEMBLER
+CLASSIFICATION_TASK -> IMPUTER ENCODERS FEATURE_GENERATOR FEATURE_SCALER FEATURE_SELECTOR CLASSIFIER
 REGRESSION_TASK -> IMPUTER ENCODERS FEATURE_GENERATOR FEATURE_SCALER FEATURE_SELECTOR REGRESSOR REGRESSION_ENSEMBLER
 CLUSTERING_TASK -> IMPUTER ENCODERS FEATURE_GENERATOR FEATURE_SCALER FEATURE_SELECTOR CLUSTERER
 TIME_SERIES_FORECAST_TASK -> IMPUTER TIME_SERIES_FORECASTER | REGRESSION_TASK

diff --git a/alpha_automl/resource/primitives_hierarchy.json b/alpha_automl/resource/primitives_hierarchy.json
@@ -20,7 +20,6 @@
         "alpha_automl.builtin_primitives.image_encoder.HogTransformer"
     ],
     "FEATURE_GENERATOR": [
-        "alpha_automl.wrapper_primitives.llm_feature_engine.LLMFeatureGenerator",
         "feature_engine.creation.math_features.MathFeatures-sum",
         "feature_engine.creation.math_features.MathFeatures-mean",
         "feature_engine.creation.math_features.MathFeatures-prod",
@@ -56,21 +55,12 @@
         "sklearn.ensemble.VotingRegressor"
     ],
     "CLASSIFIER": [
-        "sklearn.discriminant_analysis.LinearDiscriminantAnalysis",
-        "sklearn.discriminant_analysis.QuadraticDiscriminantAnalysis",
         "sklearn.ensemble.ExtraTreesClassifier",
         "sklearn.ensemble.GradientBoostingClassifier",
         "sklearn.ensemble.RandomForestClassifier",
-        "sklearn.naive_bayes.BernoulliNB",
-        "sklearn.naive_bayes.GaussianNB",
-        "sklearn.naive_bayes.MultinomialNB",
         "sklearn.neighbors.KNeighborsClassifier",
         "sklearn.linear_model.LogisticRegression",
-        "sklearn.linear_model.PassiveAggressiveClassifier",
         "sklearn.linear_model.SGDClassifier",
-        "sklearn.svm.LinearSVC",
-        "sklearn.svm.SVC",
-        "sklearn.tree.DecisionTreeClassifier",
         "xgboost.XGBClassifier",
         "lightgbm.LGBMClassifier",
         "catboost.CatBoostClassifier"

diff --git a/alpha_automl/wrapper_primitives/llm_feature_engine.py b/alpha_automl/wrapper_primitives/llm_feature_engine.py
@@ -29,6 +29,14 @@ def transform(self, X, y=None):
         exec(compile(parsed, filename="<ast>", mode="exec"), access_scope, loc)
         return np.array(X_cp)
 
+    def transform_df(self, X, y=None):
+        X_cp = copy.deepcopy(X)
+        loc = {}
+        access_scope = {"df": X_cp, "pd": pd, "np": np}
+        parsed = ast.parse(self.code)
+        exec(compile(parsed, filename="<ast>", mode="exec"), access_scope, loc)
+        return X_cp
+
 def get_prompt(
     df, description, iterative=1, data_description_unparsed=None, samples=None, **kwargs
 ):