dfd23ba635f59f0dbc2c0cdf04445e5f6eda3c66,AutoSklearn/autosklearn.py,AutoSklearnClassifier,fit,#AutoSklearnClassifier#Any#Any#,84

Before Change

            self._preprocessor = components.preprocessing_components.\
                _preprocessors[preproc_name](random_state=random_state, **preproc_params)
            self._preprocessor.fit(X, Y)
            X = self._preprocessor.transform(X)

        self._estimator.fit(X, Y)
        return self

After Change

        // TODO: can this happen now that a configuration is specified at
        // instantiation time

        steps = []

        preprocessor = self.configuration["preprocessor"]
        if preprocessor.value != "None":
            preproc_name = preprocessor.value
            preproc_params = {}

            for instantiated_hyperparameter in self.configuration:
                if not instantiated_hyperparameter.hyperparameter.name \
                        .startswith(preproc_name):
                    continue
                if isinstance(instantiated_hyperparameter,
                              InactiveHyperparameter):
                    continue

                name_ = instantiated_hyperparameter.hyperparameter.name. \
                    split(":")[1]
                preproc_params[name_] = instantiated_hyperparameter.value

            preprocessor_object = components.preprocessing_components. \
                _preprocessors[preproc_name](random_state=self.random_state,
                                             **preproc_params)
            steps.append((preproc_name, preprocessor_object))

        // Extract Hyperparameters from the configuration object
        classifier_name = self.configuration["classifier"].value
        classifier_parameters = {}
        for instantiated_hyperparameter in self.configuration:
            if not instantiated_hyperparameter.hyperparameter.name.startswith(
                    classifier_name):
                continue
            if isinstance(instantiated_hyperparameter, InactiveHyperparameter):
                continue

            name_ = instantiated_hyperparameter.hyperparameter.name.\
                split(":")[1]
            classifier_parameters[name_] = instantiated_hyperparameter.value

        classifier_object = components.classification_components._classifiers\
            [classifier_name](random_state=self.random_state,
                              **classifier_parameters)
        steps.append((classifier_name, classifier_object))

        self._validate_input_X(X)
        self._validate_input_Y(Y)

        self._pipeline = Pipeline(steps)
        self._pipeline.fit(X, Y)
        return self

    def predict(self, X):

Instances