import copy import sys import time import traceback from collections import OrderedDict, defaultdict import cloudpickle import numpy as np import pandas as pd from sklearn.model_selection import BaseCrossValidator from .pipeline_search_plots import PipelineSearchPlots from evalml.automl.automl_algorithm import IterativeAlgorithm from evalml.automl.callbacks import log_error_callback from evalml.automl.utils import ( get_default_primary_search_objective, make_data_splitter ) from evalml.data_checks import ( AutoMLDataChecks, DataChecks, DefaultDataChecks, EmptyDataChecks, HighVarianceCVDataCheck ) from evalml.exceptions import ( AutoMLSearchException, PipelineNotFoundError, PipelineScoreError ) from evalml.model_family import ModelFamily from evalml.objectives import ( get_core_objectives, get_non_core_objectives, get_objective ) from evalml.pipelines import ( BinaryClassificationPipeline, MeanBaselineRegressionPipeline, ModeBaselineBinaryPipeline, ModeBaselineMulticlassPipeline, PipelineBase, TimeSeriesBaselineBinaryPipeline, TimeSeriesBaselineMulticlassPipeline, TimeSeriesBaselineRegressionPipeline ) from evalml.pipelines.components.utils import get_estimators from evalml.pipelines.utils import get_generated_pipeline_class, make_pipeline from evalml.preprocessing import split_data from evalml.problem_types import ProblemTypes, handle_problem_types, is_binary from evalml.tuners import SKOptTuner from evalml.utils import convert_to_seconds, get_random_seed from evalml.utils.gen_utils import ( _convert_to_woodwork_structure, _convert_woodwork_types_wrapper ) from evalml.utils.logger import ( get_logger, log_subtitle, log_title, time_elapsed, update_pipeline ) logger = get_logger(__file__) [docs]class AutoMLSearch: """Automated Pipeline search.""" _MAX_NAME_LEN = 40 # Necessary for "Plotting" documentation, since Sphinx does not work well with instance attributes. plot = PipelineSearchPlots [docs] def __init__(self, X_train=None, y_train=None, problem_type=None, objective='auto', max_iterations=None, max_time=None, patience=None, tolerance=None, data_splitter=None, allowed_pipelines=None, allowed_model_families=None, start_iteration_callback=None, add_result_callback=None, error_callback=None, additional_objectives=None, random_state=0, n_jobs=-1, tuner_class=None, optimize_thresholds=False, ensembling=False, max_batches=None, problem_configuration=None, train_best_pipeline=True, pipeline_parameters=None, _pipelines_per_batch=5): """Automated pipeline search Arguments: X_train (pd.DataFrame, ww.DataTable): The input training data of shape [n_samples, n_features]. Required. y_train (pd.Series, ww.DataColumn): The target training data of length [n_samples]. Required for supervised learning tasks. problem_type (str or ProblemTypes): type of supervised learning problem. See evalml.problem_types.ProblemType.all_problem_types for a full list. objective (str, ObjectiveBase): The objective to optimize for. Used to propose and rank pipelines, but not for optimizing each pipeline during fit-time. When set to 'auto', chooses: - LogLossBinary for binary classification problems, - LogLossMulticlass for multiclass classification problems, and - R2 for regression problems. max_iterations (int): Maximum number of iterations to search. If max_iterations and max_time is not set, then max_iterations will default to max_iterations of 5. max_time (int, str): Maximum time to search for pipelines. This will not start a new pipeline search after the duration has elapsed. If it is an integer, then the time will be in seconds. For strings, time can be specified as seconds, minutes, or hours. patience (int): Number of iterations without improvement to stop search early. Must be positive. If None, early stopping is disabled. Defaults to None. tolerance (float): Minimum percentage difference to qualify as score improvement for early stopping. Only applicable if patience is not None. Defaults to None. allowed_pipelines (list(class)): A list of PipelineBase subclasses indicating the pipelines allowed in the search. The default of None indicates all pipelines for this problem type are allowed. Setting this field will cause allowed_model_families to be ignored. allowed_model_families (list(str, ModelFamily)): The model families to search. The default of None searches over all model families. Run evalml.pipelines.components.utils.allowed_model_families("binary") to see options. Change `binary` to `multiclass` or `regression` depending on the problem type. Note that if allowed_pipelines is provided, this parameter will be ignored. data_splitter (sklearn.model_selection.BaseCrossValidator): Data splitting method to use. Defaults to StratifiedKFold. tuner_class: The tuner class to use. Defaults to SKOptTuner. start_iteration_callback (callable): Function called before each pipeline training iteration. Callback function takes three positional parameters: The pipeline class, the pipeline parameters, and the AutoMLSearch object. add_result_callback (callable): Function called after each pipeline training iteration. Callback function takes three positional parameters: A dictionary containing the training results for the new pipeline, an untrained_pipeline containing the parameters used during training, and the AutoMLSearch object. error_callback (callable): Function called when `search()` errors and raises an Exception. Callback function takes three positional parameters: the Exception raised, the traceback, and the AutoMLSearch object. Must also accepts kwargs, so AutoMLSearch is able to pass along other appropriate parameters by default. Defaults to None, which will call `log_error_callback`. additional_objectives (list): Custom set of objectives to score on. Will override default objectives for problem type if not empty. random_state (int): Seed for the random number generator. Defaults to 0. n_jobs (int or None): Non-negative integer describing level of parallelism used for pipelines. None and 1 are equivalent. If set to -1, all CPUs are used. For n_jobs below -1, (n_cpus + 1 + n_jobs) are used. ensembling (boolean): If True, runs ensembling in a separate batch after every allowed pipeline class has been iterated over. If the number of unique pipelines to search over per batch is one, ensembling will not run. Defaults to False. max_batches (int): The maximum number of batches of pipelines to search. Parameters max_time, and max_iterations have precedence over stopping the search. problem_configuration (dict, None): Additional parameters needed to configure the search. For example, in time series problems, values should be passed in for the gap and max_delay variables. train_best_pipeline (boolean): Whether or not to train the best pipeline before returning it. Defaults to True _pipelines_per_batch (int): The number of pipelines to train for every batch after the first one. The first batch will train a baseline pipline + one of each pipeline family allowed in the search. """ if X_train is None: raise ValueError('Must specify training data as a 2d array using the X_train argument') if y_train is None: raise ValueError('Must specify training data target values as a 1d vector using the y_train argument') try: self.problem_type = handle_problem_types(problem_type) except ValueError: raise ValueError('choose one of (binary, multiclass, regression) as problem_type') self.tuner_class = tuner_class or SKOptTuner self.start_iteration_callback = start_iteration_callback self.add_result_callback = add_result_callback self.error_callback = error_callback or log_error_callback self.data_splitter = data_splitter self.optimize_thresholds = optimize_thresholds self.ensembling = ensembling if objective == 'auto': objective = get_default_primary_search_objective(self.problem_type.value) objective = get_objective(objective, return_instance=False) self.objective = self._validate_objective(objective) if self.data_splitter is not None and not issubclass(self.data_splitter.__class__, BaseCrossValidator): raise ValueError("Not a valid data splitter") if not objective.is_defined_for_problem_type(self.problem_type): raise ValueError("Given objective {} is not compatible with a {} problem.".format(self.objective.name, self.problem_type.value)) if additional_objectives is None: additional_objectives = get_core_objectives(self.problem_type) # if our main objective is part of default set of objectives for problem_type, remove it existing_main_objective = next((obj for obj in additional_objectives if obj.name == self.objective.name), None) if existing_main_objective is not None: additional_objectives.remove(existing_main_objective) else: additional_objectives = [get_objective(o) for o in additional_objectives] additional_objectives = [self._validate_objective(obj) for obj in additional_objectives] self.additional_objectives = additional_objectives self.objective_name_to_class = {o.name: o for o in [self.objective] + self.additional_objectives} if not isinstance(max_time, (int, float, str, type(None))): raise TypeError(f"Parameter max_time must be a float, int, string or None. Received {type(max_time)} with value {str(max_time)}..") if isinstance(max_time, (int, float)) and max_time < 0: raise ValueError(f"Parameter max_time must be None or non-negative. Received {max_time}.") if max_batches is not None and max_batches < 0: raise ValueError(f"Parameter max_batches must be None or non-negative. Received {max_batches}.") if max_iterations is not None and max_iterations < 0: raise ValueError(f"Parameter max_iterations must be None or non-negative. Received {max_iterations}.") self.max_time = convert_to_seconds(max_time) if isinstance(max_time, str) else max_time self.max_iterations = max_iterations self.max_batches = max_batches self._pipelines_per_batch = _pipelines_per_batch if not self.max_iterations and not self.max_time and not self.max_batches: self.max_batches = 1 logger.info("Using default limit of max_batches=1.\n") if patience and (not isinstance(patience, int) or patience < 0): raise ValueError("patience value must be a positive integer. Received {} instead".format(patience)) if tolerance and (tolerance > 1.0 or tolerance < 0.0): raise ValueError("tolerance value must be a float between 0.0 and 1.0 inclusive. Received {} instead".format(tolerance)) self.patience = patience self.tolerance = tolerance or 0.0 self._results = { 'pipeline_results': {}, 'search_order': [], 'errors': [] } self.random_state = get_random_seed(random_state) self.n_jobs = n_jobs self.plot = None try: self.plot = PipelineSearchPlots(self) except ImportError: logger.warning("Unable to import plotly; skipping pipeline search plotting\n") self._data_check_results = None self.allowed_pipelines = allowed_pipelines self.allowed_model_families = allowed_model_families self._automl_algorithm = None self._start = None self._baseline_cv_scores = {} self.show_batch_output = False self._validate_problem_type() self.problem_configuration = self._validate_problem_configuration(problem_configuration) self._train_best_pipeline = train_best_pipeline self._best_pipeline = None self._searched = False # make everything ww objects self.X_train = _convert_to_woodwork_structure(X_train) self.y_train = _convert_to_woodwork_structure(y_train) default_data_splitter = make_data_splitter(self.X_train, self.y_train, self.problem_type, self.problem_configuration, n_splits=3, shuffle=True, random_state=self.random_state) self.data_splitter = self.data_splitter or default_data_splitter self.pipeline_parameters = pipeline_parameters if pipeline_parameters is not None else {} def _validate_objective(self, objective): non_core_objectives = get_non_core_objectives() if isinstance(objective, type): if objective in non_core_objectives: raise ValueError(f"{objective.name.lower()} is not allowed in AutoML! " "Use evalml.objectives.utils.get_core_objective_names() " "to get all objective names allowed in automl.") return objective() return objective @property def data_check_results(self): """If there are data checks, return any error messages that are found""" return self._data_check_results def __str__(self): def _print_list(obj_list): lines = sorted(['\t{}'.format(o.name) for o in obj_list]) return '\n'.join(lines) def _get_funct_name(function): if callable(function): return function.__name__ else: return None search_desc = ( f"{handle_problem_types(self.problem_type).name} Search\n\n" f"Parameters: \n{'='*20}\n" f"Objective: {get_objective(self.objective).name}\n" f"Max Time: {self.max_time}\n" f"Max Iterations: {self.max_iterations}\n" f"Max Batches: {self.max_batches}\n" f"Allowed Pipelines: \n{_print_list(self.allowed_pipelines or [])}\n" f"Patience: {self.patience}\n" f"Tolerance: {self.tolerance}\n" f"Data Splitting: {self.data_splitter}\n" f"Tuner: {self.tuner_class.__name__}\n" f"Start Iteration Callback: {_get_funct_name(self.start_iteration_callback)}\n" f"Add Result Callback: {_get_funct_name(self.add_result_callback)}\n" f"Additional Objectives: {_print_list(self.additional_objectives or [])}\n" f"Random State: {self.random_state}\n" f"n_jobs: {self.n_jobs}\n" f"Optimize Thresholds: {self.optimize_thresholds}\n" ) rankings_desc = "" if not self.rankings.empty: rankings_str = self.rankings.drop(['parameters'], axis='columns').to_string() rankings_desc = f"\nSearch Results: \n{'='*20}\n{rankings_str}" return search_desc + rankings_desc def _validate_problem_configuration(self, problem_configuration=None): if self.problem_type in [ProblemTypes.TIME_SERIES_REGRESSION]: required_parameters = {'gap', 'max_delay'} if not problem_configuration or not all(p in problem_configuration for p in required_parameters): raise ValueError("user_parameters must be a dict containing values for at least the gap and max_delay " f"parameters. Received {problem_configuration}.") return problem_configuration or {} def _validate_data_checks(self, data_checks): """Validate data_checks parameter. Arguments: data_checks (DataChecks, list(Datacheck), str, None): Input to validate. If not of the right type, raise an exception. Returns: An instance of DataChecks used to perform checks before search. """ if isinstance(data_checks, DataChecks): return data_checks elif isinstance(data_checks, list): return AutoMLDataChecks(data_checks) elif isinstance(data_checks, str): if data_checks == "auto": return DefaultDataChecks(problem_type=self.problem_type, objective=self.objective, n_splits=self.data_splitter.get_n_splits()) elif data_checks == "disabled": return EmptyDataChecks() else: raise ValueError("If data_checks is a string, it must be either 'auto' or 'disabled'. " f"Received '{data_checks}'.") elif data_checks is None: return EmptyDataChecks() else: return DataChecks(data_checks) def _handle_keyboard_interrupt(self, pipeline, current_batch_pipelines): """Presents a prompt to the user asking if they want to stop the search. Arguments: pipeline (PipelineBase): Current pipeline in the search. current_batch_pipelines (list): Other pipelines in the batch. Returns: list: Next pipelines to search in the batch. If the user decides to stop the search, an empty list will be returned. """ leading_char = "\n" start_of_loop = time.time() while True: choice = input(leading_char + "Do you really want to exit search (y/n)? ").strip().lower() if choice == "y": logger.info("Exiting AutoMLSearch.") return [] elif choice == "n": # So that the time in this loop does not count towards the time budget (if set) time_in_loop = time.time() - start_of_loop self._start += time_in_loop return [pipeline] + current_batch_pipelines else: leading_char = "" [docs] def search(self, data_checks="auto", show_iteration_plot=True): """Find the best pipeline for the data set. Arguments: data_checks (DataChecks, list(Datacheck), str, None): A collection of data checks to run before automl search. If data checks produce any errors, an exception will be thrown before the search begins. If "disabled" or None, `no` data checks will be done. If set to "auto", DefaultDataChecks will be done. Default value is set to "auto". feature_types (list, optional): list of feature types, either numerical or categorical. Categorical features will automatically be encoded show_iteration_plot (boolean, True): Shows an iteration vs. score plot in Jupyter notebook. Disabled by default in non-Jupyter enviroments. """ if self._searched: logger.info("AutoMLSearch.search() has already been run and will not run again on the same instance. Re-initialize AutoMLSearch to search again.") return # don't show iteration plot outside of a jupyter notebook if show_iteration_plot: try: get_ipython except NameError: show_iteration_plot = False text_column_vals = self.X_train.select('natural_language') text_columns = list(text_column_vals.to_dataframe().columns) if len(text_columns) == 0: text_columns = None data_checks = self._validate_data_checks(data_checks) self._data_check_results = data_checks.validate(_convert_woodwork_types_wrapper(self.X_train.to_dataframe()), _convert_woodwork_types_wrapper(self.y_train.to_series())) for result in self._data_check_results["warnings"]: logger.warning(result["message"]) for result in self._data_check_results["errors"]: logger.error(result["message"]) if self._data_check_results["errors"]: raise ValueError("Data checks raised some warnings and/or errors. Please see `self.data_check_results` for more information or pass data_checks='disabled' to search() to disable data checking.") if self.allowed_pipelines is None: logger.info("Generating pipelines to search over...") allowed_estimators = get_estimators(self.problem_type, self.allowed_model_families) logger.debug(f"allowed_estimators set to {[estimator.name for estimator in allowed_estimators]}") self.allowed_pipelines = [make_pipeline(self.X_train, self.y_train, estimator, self.problem_type, custom_hyperparameters=self.pipeline_parameters, text_columns=text_columns) for estimator in allowed_estimators] if self.allowed_pipelines == []: raise ValueError("No allowed pipelines to search") run_ensembling = self.ensembling if run_ensembling and len(self.allowed_pipelines) == 1: logger.warning("Ensembling is set to True, but the number of unique pipelines is one, so ensembling will not run.") run_ensembling = False if run_ensembling and self.max_iterations is not None: # Baseline + first batch + each pipeline iteration + 1 first_ensembling_iteration = (1 + len(self.allowed_pipelines) + len(self.allowed_pipelines) * self._pipelines_per_batch + 1) if self.max_iterations < first_ensembling_iteration: run_ensembling = False logger.warning(f"Ensembling is set to True, but max_iterations is too small, so ensembling will not run. Set max_iterations >= {first_ensembling_iteration} to run ensembling.") else: logger.info(f"Ensembling will run at the {first_ensembling_iteration} iteration and every {len(self.allowed_pipelines) * self._pipelines_per_batch} iterations after that.") if self.max_batches and self.max_iterations is None: self.show_batch_output = True if run_ensembling: ensemble_nth_batch = len(self.allowed_pipelines) + 1 num_ensemble_batches = (self.max_batches - 1) // ensemble_nth_batch if num_ensemble_batches == 0: logger.warning(f"Ensembling is set to True, but max_batches is too small, so ensembling will not run. Set max_batches >= {ensemble_nth_batch + 1} to run ensembling.") else: logger.info(f"Ensembling will run every {ensemble_nth_batch} batches.") self.max_iterations = (1 + len(self.allowed_pipelines) + self._pipelines_per_batch * (self.max_batches - 1 - num_ensemble_batches) + num_ensemble_batches) else: self.max_iterations = 1 + len(self.allowed_pipelines) + (self._pipelines_per_batch * (self.max_batches - 1)) self.allowed_model_families = list(set([p.model_family for p in (self.allowed_pipelines)])) logger.debug(f"allowed_pipelines set to {[pipeline.name for pipeline in self.allowed_pipelines]}") logger.debug(f"allowed_model_families set to {self.allowed_model_families}") if len(self.problem_configuration): pipeline_params = {**{'pipeline': self.problem_configuration}, **self.pipeline_parameters} else: pipeline_params = self.pipeline_parameters self._automl_algorithm = IterativeAlgorithm( max_iterations=self.max_iterations, allowed_pipelines=self.allowed_pipelines, tuner_class=self.tuner_class, text_columns=text_columns, random_state=self.random_state, n_jobs=self.n_jobs, number_features=self.X_train.shape[1], pipelines_per_batch=self._pipelines_per_batch, ensembling=run_ensembling, pipeline_params=pipeline_params ) log_title(logger, "Beginning pipeline search") logger.info("Optimizing for %s. " % self.objective.name) logger.info("{} score is better.\n".format('Greater' if self.objective.greater_is_better else 'Lower')) if self.max_batches is not None: logger.info(f"Searching up to {self.max_batches} batches for a total of {self.max_iterations} pipelines. ") elif self.max_iterations is not None: logger.info("Searching up to %s pipelines. " % self.max_iterations) if self.max_time is not None: logger.info("Will stop searching for new pipelines after %d seconds.\n" % self.max_time) logger.info("Allowed model families: %s\n" % ", ".join([model.value for model in self.allowed_model_families])) search_iteration_plot = None if self.plot: search_iteration_plot = self.plot.search_iteration_plot(interactive_plot=show_iteration_plot) self._start = time.time() should_terminate = self._add_baseline_pipelines() if should_terminate: return current_batch_pipelines = [] current_batch_pipeline_scores = [] while self._check_stopping_condition(self._start): try: if current_batch_pipeline_scores and np.isnan(np.array(current_batch_pipeline_scores, dtype=float)).all(): raise AutoMLSearchException(f"All pipelines in the current AutoML batch produced a score of np.nan on the primary objective {self.objective}.") current_batch_pipelines = self._automl_algorithm.next_batch() current_batch_pipeline_scores = [] except StopIteration: logger.info('AutoML Algorithm out of recommendations, ending') break current_batch_size = len(current_batch_pipelines) current_batch_pipeline_scores = self._evaluate_pipelines(current_batch_pipelines, search_iteration_plot=search_iteration_plot) # Different size indicates early stopping if len(current_batch_pipeline_scores) != current_batch_size: break elapsed_time = time_elapsed(self._start) desc = f"\nSearch finished after {elapsed_time}" desc = desc.ljust(self._MAX_NAME_LEN) logger.info(desc) best_pipeline = self.rankings.iloc[0] best_pipeline_name = best_pipeline["pipeline_name"] self._find_best_pipeline() logger.info(f"Best pipeline: {best_pipeline_name}") logger.info(f"Best pipeline {self.objective.name}: {best_pipeline['score']:3f}") self._searched = True def _find_best_pipeline(self): """Finds the best pipeline in the rankings If self._best_pipeline already exists, check to make sure it is different from the current best pipeline before training and thresholding""" best_pipeline = self.rankings.iloc[0] if not (self._best_pipeline and self._best_pipeline == self.get_pipeline(best_pipeline['id'])): self._best_pipeline = self.get_pipeline(best_pipeline['id']) if self._train_best_pipeline: X_threshold_tuning = None y_threshold_tuning = None X_train, y_train = self.X_train, self.y_train if self.optimize_thresholds and self.objective.is_defined_for_problem_type(ProblemTypes.BINARY) and self.objective.can_optimize_threshold and is_binary(self.problem_type): X_train, X_threshold_tuning, y_train, y_threshold_tuning = split_data(X_train, y_train, self.problem_type, test_size=0.2, random_state=self.random_state) self._best_pipeline.fit(X_train, y_train) self._best_pipeline = self._tune_binary_threshold(self._best_pipeline, X_threshold_tuning, y_threshold_tuning) def _tune_binary_threshold(self, pipeline, X_threshold_tuning, y_threshold_tuning): """Tunes the threshold of a binary pipeline to the X and y thresholding data Arguments: pipeline (Pipeline): Pipeline instance to threshold X_threshold_tuning (ww.DataTable): X data to tune pipeline to y_threshold_tuning (ww.DataColumn): Target data to tune pipeline to Returns: Trained pipeline instance """ if self.objective.is_defined_for_problem_type(ProblemTypes.BINARY) and is_binary(self.problem_type): pipeline.threshold = 0.5 if X_threshold_tuning: y_predict_proba = pipeline.predict_proba(X_threshold_tuning) y_predict_proba = y_predict_proba.iloc[:, 1] pipeline.threshold = self.objective.optimize_threshold(y_predict_proba, y_threshold_tuning, X=X_threshold_tuning) return pipeline def _check_stopping_condition(self, start): should_continue = True num_pipelines = len(self._results['pipeline_results']) # check max_time and max_iterations elapsed = time.time() - start if self.max_time and elapsed >= self.max_time: return False elif self.max_iterations and num_pipelines >= self.max_iterations: return False # check for early stopping if self.patience is None: return True first_id = self._results['search_order'][0] best_score = self._results['pipeline_results'][first_id]['score'] num_without_improvement = 0 for id in self._results['search_order'][1:]: curr_score = self._results['pipeline_results'][id]['score'] significant_change = abs((curr_score - best_score) / best_score) > self.tolerance score_improved = curr_score > best_score if self.objective.greater_is_better else curr_score < best_score if score_improved and significant_change: best_score = curr_score num_without_improvement = 0 else: num_without_improvement += 1 if num_without_improvement >= self.patience: logger.info("\n\n{} iterations without improvement. Stopping search early...".format(self.patience)) return False return should_continue def _validate_problem_type(self): for obj in self.additional_objectives: if not obj.is_defined_for_problem_type(self.problem_type): raise ValueError("Additional objective {} is not compatible with a {} problem.".format(obj.name, self.problem_type.value)) for pipeline in self.allowed_pipelines or []: if pipeline.problem_type != self.problem_type: raise ValueError("Given pipeline {} is not compatible with problem_type {}.".format(pipeline.name, self.problem_type.value)) def _add_baseline_pipelines(self): """Fits a baseline pipeline to the data. This is the first pipeline fit during search. Returns: bool - If the user ends the search early, will return True and searching will immediately finish. Else, will return False and more pipelines will be searched. """ if self.problem_type == ProblemTypes.BINARY: baseline = ModeBaselineBinaryPipeline(parameters={}) elif self.problem_type == ProblemTypes.MULTICLASS: baseline = ModeBaselineMulticlassPipeline(parameters={}) elif self.problem_type == ProblemTypes.REGRESSION: baseline = MeanBaselineRegressionPipeline(parameters={}) else: pipeline_class = {ProblemTypes.TIME_SERIES_REGRESSION: TimeSeriesBaselineRegressionPipeline, ProblemTypes.TIME_SERIES_MULTICLASS: TimeSeriesBaselineMulticlassPipeline, ProblemTypes.TIME_SERIES_BINARY: TimeSeriesBaselineBinaryPipeline}[self.problem_type] gap = self.problem_configuration['gap'] max_delay = self.problem_configuration['max_delay'] baseline = pipeline_class(parameters={"pipeline": {"gap": gap, "max_delay": max_delay}, "Time Series Baseline Estimator": {"gap": gap, "max_delay": max_delay}}) pipelines = [baseline] scores = self._evaluate_pipelines(pipelines, baseline=True) if scores == []: return True return False @staticmethod def _get_mean_cv_scores_for_all_objectives(cv_data, objective_name_to_class): scores = defaultdict(int) n_folds = len(cv_data) for fold_data in cv_data: for field, value in fold_data['all_objective_scores'].items(): # The 'all_objective_scores' field contains scores for all objectives # but also fields like "# Training" and "# Testing", so we want to exclude them since # they are not scores if field in objective_name_to_class: scores[field] += value return {objective: float(score) / n_folds for objective, score in scores.items()} def _compute_cv_scores(self, pipeline): start = time.time() cv_data = [] logger.info("\tStarting cross validation") X_pd = _convert_woodwork_types_wrapper(self.X_train.to_dataframe()) y_pd = _convert_woodwork_types_wrapper(self.y_train.to_series()) for i, (train, valid) in enumerate(self.data_splitter.split(X_pd, y_pd)): if pipeline.model_family == ModelFamily.ENSEMBLE and i > 0: # Stacked ensembles do CV internally, so we do not run CV here for performance reasons. logger.debug(f"Skipping fold {i} because CV for stacked ensembles is not supported.") break logger.debug(f"\t\tTraining and scoring on fold {i}") X_train, X_valid = self.X_train.iloc[train], self.X_train.iloc[valid] y_train, y_valid = self.y_train.iloc[train], self.y_train.iloc[valid] if self.problem_type in [ProblemTypes.BINARY, ProblemTypes.MULTICLASS]: diff_train = set(np.setdiff1d(self.y_train.to_series(), y_train.to_series())) diff_valid = set(np.setdiff1d(self.y_train.to_series(), y_valid.to_series())) diff_string = f"Missing target values in the training set after data split: {diff_train}. " if diff_train else "" diff_string += f"Missing target values in the validation set after data split: {diff_valid}." if diff_valid else "" if diff_string: raise Exception(diff_string) objectives_to_score = [self.objective] + self.additional_objectives cv_pipeline = None try: X_threshold_tuning = None y_threshold_tuning = None if self.optimize_thresholds and self.objective.is_defined_for_problem_type(ProblemTypes.BINARY) and self.objective.can_optimize_threshold and is_binary(self.problem_type): X_train, X_threshold_tuning, y_train, y_threshold_tuning = split_data(X_train, y_train, self.problem_type, test_size=0.2, random_state=self.random_state) cv_pipeline = pipeline.clone() logger.debug(f"\t\t\tFold {i}: starting training") cv_pipeline.fit(X_train, y_train) logger.debug(f"\t\t\tFold {i}: finished training") cv_pipeline = self._tune_binary_threshold(cv_pipeline, X_threshold_tuning, y_threshold_tuning) if X_threshold_tuning: logger.debug(f"\t\t\tFold {i}: Optimal threshold found ({cv_pipeline.threshold:.3f})") logger.debug(f"\t\t\tFold {i}: Scoring trained pipeline") scores = cv_pipeline.score(X_valid, y_valid, objectives=objectives_to_score) logger.debug(f"\t\t\tFold {i}: {self.objective.name} score: {scores[self.objective.name]:.3f}") score = scores[self.objective.name] except Exception as e: if self.error_callback is not None: self.error_callback(exception=e, traceback=traceback.format_tb(sys.exc_info()[2]), automl=self, fold_num=i, pipeline=pipeline) if isinstance(e, PipelineScoreError): nan_scores = {objective: np.nan for objective in e.exceptions} scores = {**nan_scores, **e.scored_successfully} scores = OrderedDict({o.name: scores[o.name] for o in [self.objective] + self.additional_objectives}) score = scores[self.objective.name] else: score = np.nan scores = OrderedDict(zip([n.name for n in self.additional_objectives], [np.nan] * len(self.additional_objectives))) ordered_scores = OrderedDict() ordered_scores.update({self.objective.name: score}) ordered_scores.update(scores) ordered_scores.update({"# Training": y_train.shape[0]}) ordered_scores.update({"# Validation": y_valid.shape[0]}) evaluation_entry = {"all_objective_scores": ordered_scores, "score": score, 'binary_classification_threshold': None} if isinstance(cv_pipeline, BinaryClassificationPipeline) and cv_pipeline.threshold is not None: evaluation_entry['binary_classification_threshold'] = cv_pipeline.threshold cv_data.append(evaluation_entry) training_time = time.time() - start cv_scores = pd.Series([fold['score'] for fold in cv_data]) cv_score_mean = cv_scores.mean() logger.info(f"\tFinished cross validation - mean {self.objective.name}: {cv_score_mean:.3f}") return {'cv_data': cv_data, 'training_time': training_time, 'cv_scores': cv_scores, 'cv_score_mean': cv_score_mean} def _add_result(self, trained_pipeline, parameters, training_time, cv_data, cv_scores): cv_score = cv_scores.mean() percent_better_than_baseline = {} mean_cv_all_objectives = self._get_mean_cv_scores_for_all_objectives(cv_data, self.objective_name_to_class) for obj_name in mean_cv_all_objectives: objective_class = self.objective_name_to_class[obj_name] # In the event add_to_rankings is called before search _baseline_cv_scores will be empty so we will return # nan for the base score. percent_better = objective_class.calculate_percent_difference(mean_cv_all_objectives[obj_name], self._baseline_cv_scores.get(obj_name, np.nan)) percent_better_than_baseline[obj_name] = percent_better pipeline_name = trained_pipeline.name pipeline_summary = trained_pipeline.summary pipeline_id = len(self._results['pipeline_results']) high_variance_cv_check = HighVarianceCVDataCheck(threshold=0.2) high_variance_cv_check_results = high_variance_cv_check.validate(pipeline_name=pipeline_name, cv_scores=cv_scores) high_variance_cv = False if high_variance_cv_check_results["warnings"]: logger.warning(high_variance_cv_check_results["warnings"][0]["message"]) high_variance_cv = True self._results['pipeline_results'][pipeline_id] = { "id": pipeline_id, "pipeline_name": pipeline_name, "pipeline_class": type(trained_pipeline), "pipeline_summary": pipeline_summary, "parameters": parameters, "score": cv_score, "high_variance_cv": high_variance_cv, "training_time": training_time, "cv_data": cv_data, "percent_better_than_baseline_all_objectives": percent_better_than_baseline, "percent_better_than_baseline": percent_better_than_baseline[self.objective.name], "validation_score": cv_scores[0] } self._results['search_order'].append(pipeline_id) if self.add_result_callback: self.add_result_callback(self._results['pipeline_results'][pipeline_id], trained_pipeline, self) def _evaluate_pipelines(self, current_pipeline_batch, baseline=False, search_iteration_plot=None): current_batch_pipeline_scores = [] add_single_pipeline = False if isinstance(current_pipeline_batch, PipelineBase): current_pipeline_batch = [current_pipeline_batch] add_single_pipeline = True while len(current_pipeline_batch) > 0 and (add_single_pipeline or baseline or self._check_stopping_condition(self._start)): pipeline = current_pipeline_batch.pop() try: parameters = pipeline.parameters logger.debug('Evaluating pipeline {}'.format(pipeline.name)) logger.debug('Pipeline parameters: {}'.format(parameters)) if self.start_iteration_callback: self.start_iteration_callback(pipeline.__class__, parameters, self) desc = f"{pipeline.name}" if len(desc) > self._MAX_NAME_LEN: desc = desc[:self._MAX_NAME_LEN - 3] + "..." desc = desc.ljust(self._MAX_NAME_LEN) if not add_single_pipeline: update_pipeline(logger, desc, len(self._results['pipeline_results']) + 1, self.max_iterations, self._start, 1 if baseline else self._automl_algorithm.batch_number, self.show_batch_output) evaluation_results = self._compute_cv_scores(pipeline) parameters = pipeline.parameters if baseline: self._baseline_cv_scores = self._get_mean_cv_scores_for_all_objectives(evaluation_results["cv_data"], self.objective_name_to_class) logger.debug('Adding results for pipeline {}\nparameters {}\nevaluation_results {}'.format(pipeline.name, parameters, evaluation_results)) self._add_result(trained_pipeline=pipeline, parameters=parameters, training_time=evaluation_results['training_time'], cv_data=evaluation_results['cv_data'], cv_scores=evaluation_results['cv_scores']) logger.debug('Adding results complete') score = evaluation_results['cv_score_mean'] score_to_minimize = -score if self.objective.greater_is_better else score current_batch_pipeline_scores.append(score_to_minimize) if not baseline and not add_single_pipeline: self._automl_algorithm.add_result(score_to_minimize, pipeline) if search_iteration_plot: search_iteration_plot.update() if add_single_pipeline: add_single_pipeline = False except KeyboardInterrupt: current_pipeline_batch = self._handle_keyboard_interrupt(pipeline, current_pipeline_batch) if current_pipeline_batch == []: return current_batch_pipeline_scores return current_batch_pipeline_scores [docs] def get_pipeline(self, pipeline_id): """Given the ID of a pipeline training result, returns an untrained instance of the specified pipeline initialized with the parameters used to train that pipeline during automl search. Arguments: pipeline_id (int): pipeline to retrieve Returns: PipelineBase: untrained pipeline instance associated with the provided ID """ pipeline_results = self.results['pipeline_results'].get(pipeline_id) if pipeline_results is None: raise PipelineNotFoundError("Pipeline not found in automl results") pipeline_class = pipeline_results.get('pipeline_class') parameters = pipeline_results.get('parameters') if pipeline_class is None or parameters is None: raise PipelineNotFoundError("Pipeline class or parameters not found in automl results") pipeline = get_generated_pipeline_class(self.problem_type) pipeline.custom_hyperparameters = pipeline_class.custom_hyperparameters pipeline.custom_name = pipeline_class.name pipeline.component_graph = pipeline_class.component_graph return pipeline(parameters, random_state=self.random_state) [docs] def describe_pipeline(self, pipeline_id, return_dict=False): """Describe a pipeline Arguments: pipeline_id (int): pipeline to describe return_dict (bool): If True, return dictionary of information about pipeline. Defaults to False. Returns: Description of specified pipeline. Includes information such as type of pipeline components, problem, training time, cross validation, etc. """ if pipeline_id not in self._results['pipeline_results']: raise PipelineNotFoundError("Pipeline not found") pipeline = self.get_pipeline(pipeline_id) pipeline_results = self._results['pipeline_results'][pipeline_id] pipeline.describe() log_subtitle(logger, "Training") logger.info("Training for {} problems.".format(pipeline.problem_type)) if self.optimize_thresholds and self.objective.is_defined_for_problem_type(ProblemTypes.BINARY) and self.objective.can_optimize_threshold: logger.info("Objective to optimize binary classification pipeline thresholds for: {}".format(self.objective)) logger.info("Total training time (including CV): %.1f seconds" % pipeline_results["training_time"]) log_subtitle(logger, "Cross Validation", underline="-") all_objective_scores = [fold["all_objective_scores"] for fold in pipeline_results["cv_data"]] all_objective_scores = pd.DataFrame(all_objective_scores) for c in all_objective_scores: if c in ["# Training", "# Validation"]: all_objective_scores[c] = all_objective_scores[c].astype("object") continue mean = all_objective_scores[c].mean(axis=0) std = all_objective_scores[c].std(axis=0) all_objective_scores.loc["mean", c] = mean all_objective_scores.loc["std", c] = std all_objective_scores.loc["coef of var", c] = std / mean if abs(mean) > 0 else np.inf all_objective_scores = all_objective_scores.fillna("-") with pd.option_context('display.float_format', '{:.3f}'.format, 'expand_frame_repr', False): logger.info(all_objective_scores) if return_dict: return pipeline_results [docs] def add_to_rankings(self, pipeline): """Fits and evaluates a given pipeline then adds the results to the automl rankings with the requirement that automl search has been run. Arguments: pipeline (PipelineBase): pipeline to train and evaluate. """ pipeline_rows = self.full_rankings[self.full_rankings['pipeline_name'] == pipeline.name] for parameter in pipeline_rows['parameters']: if pipeline.parameters == parameter: return self._evaluate_pipelines(pipeline) self._find_best_pipeline() @property def results(self): """Class that allows access to a copy of the results from `automl_search`. Returns: dict containing `pipeline_results`: a dict with results from each pipeline, and `search_order`: a list describing the order the pipelines were searched. """ return copy.deepcopy(self._results) @property def rankings(self): """Returns a pandas.DataFrame with scoring results from the highest-scoring set of parameters used with each pipeline.""" return self.full_rankings.drop_duplicates(subset="pipeline_name", keep="first") @property def full_rankings(self): """Returns a pandas.DataFrame with scoring results from all pipelines searched""" ascending = True if self.objective.greater_is_better: ascending = False full_rankings_cols = ["id", "pipeline_name", "score", "validation_score", "percent_better_than_baseline", "high_variance_cv", "parameters"] if not self._results['pipeline_results']: return pd.DataFrame(columns=full_rankings_cols) rankings_df = pd.DataFrame(self._results['pipeline_results'].values()) rankings_df = rankings_df[full_rankings_cols] rankings_df.sort_values("score", ascending=ascending, inplace=True) rankings_df.reset_index(drop=True, inplace=True) return rankings_df @property def best_pipeline(self): """Returns a trained instance of the best pipeline and parameters found during automl search. If `train_best_pipeline` is set to False, returns an untrained pipeline instance. Returns: PipelineBase: A trained instance of the best pipeline and parameters found during automl search. If `train_best_pipeline` is set to False, returns an untrained pipeline instance. """ if not self._best_pipeline: raise PipelineNotFoundError("automl search must be run before selecting `best_pipeline`.") return self._best_pipeline [docs] def save(self, file_path, pickle_protocol=cloudpickle.DEFAULT_PROTOCOL): """Saves AutoML object at file path Arguments: file_path (str): location to save file pickle_protocol (int): the pickle data stream format. Returns: None """ with open(file_path, 'wb') as f: cloudpickle.dump(self, f, protocol=pickle_protocol) [docs] @staticmethod def load(file_path): """Loads AutoML object at file path Arguments: file_path (str): location to find file to load Returns: AutoSearchBase object """ with open(file_path, 'rb') as f: return cloudpickle.load(f)