API Documentation

Functions

Convert to Data Series Functions

This module provides utilities for converting various data formats into time-series compatible structures. These utilities ensure the data is properly formatted and ready for analysis.

interpreTS.utils.data_conversion.convert_to_time_series(data)

Convert input data to a TimeSeriesData object.

Parameters:

data (pd.DataFrame, pd.Series, or np.ndarray) – The data to be converted into a TimeSeriesData object.

Returns:

An instance of TimeSeriesData wrapping the input data.

Return type:

TimeSeriesData

Raises:

• TypeError – If the input data is not of type pd.DataFrame, pd.Series, or np.ndarray.

• ValueError – If the input data is empty or has invalid dimensions.

Examples

>>> import pandas as pd
>>> data = pd.Series([1, 2, 3, 4, 5])
>>> ts_data = convert_to_time_series(data)

Data Manager

This module handles operations related to managing metadata, feature functions, and validation requirements necessary for extracting features from time-series data.

interpreTS.utils.data_manager.generate_feature_descriptions(self, extracted_features)

Generate textual descriptions for extracted features.

Parameters:

extracted_features (dict) – A dictionary where keys are feature names and values are their calculated values.

Returns:

A dictionary where keys are feature names and values are textual descriptions.

Return type:

dict

interpreTS.utils.data_manager.load_feature_functions()

interpreTS.utils.data_manager.load_metadata()

interpreTS.utils.data_manager.load_validation_requirements()

Data Validation

This module provides comprehensive functions to ensure that time-series data adheres to the expected format, dimensionality, and integrity. It includes checks for missing values, data type consistency, and other preprocessing requirements.

interpreTS.utils.data_validation.validate_time_series_data(data, feature_name=None, validation_requirements=None, **kwargs)

Validate the input time series data against dynamically provided requirements.

Parameters:

• data (pd.Series, pd.DataFrame, or np.ndarray) – The time series data to be validated.

• feature_name (str, optional) – The name of the feature to validate.

• validation_requirements (dict, optional) – A dictionary specifying the validation requirements for each feature.

• **kwargs (dict) – Additional validation parameters (overrides validation_requirements).

Returns:

True if the data is valid; raises an error otherwise.

Return type:

bool

Raises:

• TypeError – If data is not a pd.Series, pd.DataFrame, or np.ndarray.

• ValueError – If any validation requirement is not met.

Features Loader

This module enables dynamic loading and management of feature extraction functions available in the library. It provides an interface for accessing and utilizing predefined features.

class interpreTS.utils.feature_loader.FeatureLoader

Bases: object

static available_features()

Returns a list of all available features.

Returns:

List of feature names.

Return type:

list

generate_feature_options()

Generate a dictionary mapping human-readable feature names to their corresponding constants.

Returns:

A dictionary where keys are human-readable feature names (capitalized) and values are feature constants.

Return type:

dict

class interpreTS.utils.feature_loader.Features

Bases: object

ABOVE_9TH_DECILE = 'above_9th_decile'

ABSOLUTE_ENERGY = 'absolute_energy'

BELOW_1ST_DECILE = 'below_1st_decile'

BINARIZE_MEAN = 'binarize_mean'

CHANGE_IN_VARIANCE = 'change_in_variance'

CROSSING_POINTS = 'crossing_points'

DISTANCE_TO_LAST_TREND_CHANGE = 'distance_to_last_trend_change'

DOMINANT = 'dominant'

ENTROPY = 'entropy'

FLAT_SPOTS = 'flat_spots'

HETEROGENEITY = 'heterogeneity'

LENGTH = 'length'

LINEARITY = 'linearity'

MEAN = 'mean'

MEAN_CHANGE = 'mean_change'

MISSING_POINTS = 'missing_points'

OUTLIERS_IQR = 'outliers_iqr'

OUTLIERS_STD = 'outliers_std'

PEAK = 'peak'

SEASONALITY_STRENGTH = 'seasonality_strength'

SIGNIFICANT_CHANGES = 'significant_changes'

SPIKENESS = 'spikeness'

STABILITY = 'stability'

STD_1ST_DER = 'std_1st_der'

TREND_STRENGTH = 'trend_strength'

TROUGH = 'trough'

VARIABILITY_IN_SUB_PERIODS = 'variability_in_sub_periods'

VARIANCE = 'variance'

Task Manager

This module is responsible for orchestrating feature extraction tasks. It includes task generation, validation, execution, and parallelization using various computational backends.

class interpreTS.utils.task_manager.TaskManager(feature_functions, window_size, features, stride, feature_params, validation_requirements)

Bases: object

TaskManager handles feature extraction from time-series data using configurable parameters such as window size, stride, and specific feature functions.

feature_functions

A dictionary mapping feature names to their respective calculation functions.

Type:

dict

window_size

Size of the time window for feature extraction, can be a number or a time-based string.

Type:

int or str

features

Names of features to calculate.

Type:

list of str

stride

Stride size for sliding windows during feature extraction.

Type:

int or str

feature_params

Additional parameters for specific feature calculations.

Type:

dict

validation_requirements

Validation requirements for each feature.

Type:

dict

warning_registry

A set to keep track of warnings already issued during feature extraction.

Type:

set