changefinder算法，请给我Python代码

Changepoint Detection (CPD) 或者 ChangeFinder算法是一种用于识别数据序列中突然变化点的技术。常用的CPD算法有Prophet、Kneedle等，其中Prophet是Facebook开源的一种强大工具，而Kneedle算法相对简单，基于动态程序。这里是使用Python的Kneedle算法的一个简化版示例：

from ruptures import detect changepoints
from ruptures.utils import data
import numpy as np

def knn_changepoint_detection(data, min_size=2, jump=1):
    """
    Detect changepoints using the Kneedle algorithm.
    
    :param data: a list or an array of observations
    :param min_size: minimum size of a segment
    :param jump: maximum gap between two consecutive segments
    :return: a list of detected change points
    """
    signal, n_samples = data, len(data)
    distances = [np.sum(np.abs(signal[i:] - signal[:-i])) for i in range(1, n_samples)]
    sorted_idx = np.argsort(distances)
    
    # Knitrode procedure
    change_points = []
    last = None
    for k in sorted_idx:
        if (k - last >= jump) and (len(signal[k:]) >= min_size):
            change_points.append(k)
            last = k
            
    return change_points

# Example usage
data = data()  # replace this with your time series data
change_points = knn_changepoint_detection(data)

在这个例子中，你需要安装ruptures库，可以使用pip install ruptures命令。这个函数返回一个列表，其中包含了所有发现的突变点。