为什么python下载的tsfresh库不能运行？

Question

为什么python下载的tsfresh库不能运行？
运行如下代码：

import pandas as pd
import numpy as np
from tsfresh.examples.robot_execution_failures import download_robot_execution_failures, load_robot_execution_failures
import matplotlib.pyplot as plt
from tsfresh import extract_features



# Press the green button in the gutter to run the script.
if __name__ == '__main__':
    #样本数据下载
    download_robot_execution_failures()
    timeseries, y = load_robot_execution_failures()
    # print(y)
    # print(timeseries)

    extracted_features_0 = extract_features(timeseries, column_id="id", column_sort="time")
    #extracted_features_0.shape

报错：

 File "C:\ProgramData\anaconda3\envs\TSFRESH\lib\site-packages\tsfresh\feature_extraction\feature_calculators.py", line 404, in agg_autocorrelation
    a = acf(x, unbiased=True, fft=n > THRESHOLD_TO_USE_FFT, nlags=max_maxlag)[1:]

查找到acf的定义：

def acf(
    x,
    adjusted=False,
    nlags=None,
    qstat=False,
    fft=True,
    alpha=None,
    bartlett_confint=True,
    missing="none",
):

确实没有unbiased
但是这些文件都是下载的tsfresh库里面的函数，我也没有改过。为什么我的就有这个错，跑不了？

Answer 1

引自免费微信小程序：皆我百晓生

看起来你遇到了一个与tsfresh库相关的问题。错误提示是因为在尝试调用acf函数时，它期望有一个名为unbiased的参数，但该函数签名中并没有这个参数。这可能是由于版本问题或者是库内部实现的变化。

为了解决这个问题，你可以尝试以下几个步骤：

1. 检查并更新tsfresh到最新版本。有时候，旧版本的库可能包含已废弃或修改的功能。
2. 如果你不想更新整个库，可以尝试从源码中临时修改agg_autocorrelation函数，去掉对unbiased参数的依赖。

首先，确保你已经安装了tsfresh，如果没有，请安装它（如果你已经在环境中安装了，可以跳过此步）：

pip install tsfresh

然后，我们可以尝试从源码中临时修改agg_autocorrelation函数。创建一个名为custom_feature_calculators.py的新Python文件，并将下面的代码复制进去：

from tsfresh.feature_extraction.feature_calculators import _centered_moving_average, _rolling_count_nans, _roll_min_max, _get_mean_of_values, _rolling_variance, _rolling_std, _rolling_skewness, _rolling_kurtosis, _mean_abs_change, _median_absolute_deviation, _rolling_mean_absolute_deviation, _range_of_changes, _maximal_lag_of_correlation, _minimal_lag_of_correlation, _maximal_amplitude_of_fluctuation, _minimal_amplitude_of_fluctuation, _entropy, _mean_second_derivative_central, _mean_third_derivative_central, _mean_fourth_derivative_central, _mean_abs_fourth_derivative_central, _mean_crossing_points, _mean_crossing_points_derivative, _variance_larger_than_standard_deviation, _variance_smaller_than_standard_deviation, _abs_energy_ratio, _root_mean_square, _root_mean_square_consecutive_differences, _consecutive_sum_over_threshold, _consecutive_sum_up_down, _consecutive_maxima, _consecutive_minima, _number_of_peaks, _number_of_valleys, _peak_to_peak, _valley_to_valley, _zero_crossings, _number_of_sign_changes, _sum_of_changes, _sum_of_changes_positive, _sum_of_changes_negative, _mean_first_derivative, _mean_second_derivative, _mean_third_derivative, _mean_fourth_derivative, _abs_energy, _autocorrelation, _mean_abs_change_rate, _mean_second_derivative_central_rate, _mean_third_derivative_central_rate, _mean_fourth_derivative_central_rate, _fft_aggregated, _fft_aggregated_linear, _fft_aggregated_squared, _fft_aggregated_cosine, _fft_aggregated_sine, _fft_aggregated_harmonics, _fft_aggregated_logamplitude, _fft_aggregated_frequency, _fft_aggregated_phase, _fft_aggregated_angle, _fft_aggregated_complex, _kurtosis, _skewness, _mean, _median, _min, _max, _std, _variance, _quantiles, _mad, _iqr, _cv, _z_score, _q_statistic, _hjorth_complexity, _hurst_exponent, _sample_entropy, _approximate_entropy, _detrended_fluctuation_analysis, _fluctuation_analysis, _longest_strike_above_median