Skip to content

API Reference

Introduction

This document provides a detailed reference for the ete_ts library API. It covers installation, basic usage, and a complete guide to all available functions for time-series feature extraction.

Installation

Install the library from PyPI using pip:

pip install ete_ts

Basic Usage

To get started, import the library and a numerical package like NumPy. Then you can call any feature-extraction function with your time-series data.

import ete_ts
import numpy as np

# Create a sample time-series
my_series = np.array([1, 2, 4, 8, 12, 18, 25, 33, 45, 58])

# Calculate a feature
slope = ete_ts.linear_regression_slope(my_series)

print(f"The slope of the series is: {slope}")

Exceptions

The library uses standard Python exceptions to report errors. The most common are: * ValueError: Raised when a function receives an argument of the correct type but an inappropriate value (e.g., a time-series that is too short for a given calculation). * TypeError: Raised when a function receives an argument of the wrong type (e.g., a list where a NumPy array is expected).

Functions Reference

trend_strength(series, period=1, seasonal=7, robust=False)

strength = ete_ts.trend_strength(my_series)

trend_changes(series, model="l2", min_size=2, jump=5, params=None, custom_cost=None)

changes = ete_ts.trend_changes(my_series)

linear_regression_slope(series)

slope = ete_ts.linear_regression_slope(my_series)

linear_regression_r2(series)

r2_score = ete_ts.linear_regression_r2(my_series)

forecastability(series, sf, method="welch", nperseg=None, normalize=False)

entropy = ete_ts.forecastability(my_series, sf=1.0)

fluctuation(series)

fluc = ete_ts.fluctuation(my_series)

window_fluctuation(series)

win_fluc = ete_ts.window_fluctuation(my_series)

seasonal_strength(series, period=1, seasonal=7, robust=False)

season_str = ete_ts.seasonal_strength(my_series)

ac_relevance(series)

relevance = ete_ts.ac_relevance(my_series)

st_variation(series)

variation = ete_ts.st_variation(my_series)

diff_series(series)

diff_acf = ete_ts.diff_series(my_series)

complexity(series)

comp = ete_ts.complexity(my_series)

rec_concentration(series)

concentration = ete_ts.rec_concentration(my_series)

centroid(series, fs=1)

spec_centroid = ete_ts.centroid(my_series, fs=1)

info()

ete_ts.info()

all_metrics(series, fs=1)

outputs = ete_ts.all_metrics(my_series, fs=1)