已知有三个聚类好的包含三维坐标点的数组，在不调用库的情况下编写出求聚类指标CP,CH,DB,SC的代码

以下是一个Python实现的示例代码：

import numpy as np

def euclidean_distance(x, y):
    return np.sqrt(np.sum((x - y) ** 2))

def calculate_centroid(cluster):
    return np.mean(cluster, axis=0)

def calculate_sse(cluster):
    centroid = calculate_centroid(cluster)
    return np.sum([euclidean_distance(point, centroid) ** 2 for point in cluster])

def calculate_cp(clusters):
    sse_total = np.sum([calculate_sse(cluster) for cluster in clusters])
    num_points = np.sum([len(cluster) for cluster in clusters])
    sse_between = sse_total - np.sum([calculate_sse(cluster) / len(cluster) for cluster in clusters])
    return sse_between / sse_total * (num_points - len(clusters)) / (len(clusters) - 1)

def calculate_ch(clusters):
    num_clusters = len(clusters)
    centroids = [calculate_centroid(cluster) for cluster in clusters]
    centroid_distances = np.zeros((num_clusters, num_clusters))
    for i in range(num_clusters):
        for j in range(i+1, num_clusters):
            centroid_distances[i][j] = euclidean_distance(centroids[i], centroids[j])
            centroid_distances[j][i] = centroid_distances[i][j]
    max_intracluster_distance = np.max([np.max([euclidean_distance(point, cluster[i]) for point in cluster]) 
                                        for i, cluster in enumerate(clusters)])
    return max_intracluster_distance / np.mean(centroid_distances)

def calculate_db(clusters):
    num_clusters = len(clusters)
    centroids = [calculate_centroid(cluster) for cluster in clusters]
    centroid_distances = np.zeros((num_clusters, num_clusters))
    for i in range(num_clusters):
        for j in range(i+1, num_clusters):
            centroid_distances[i][j] = euclidean_distance(centroids[i], centroids[j])
            centroid_distances[j][i] = centroid_distances[i][j]
    cluster_separation = np.zeros(num_clusters)
    for i in range(num_clusters):
        sse_i = calculate_sse(clusters[i])
        max_intracluster_distance = np.max([euclidean_distance(point, centroids[i]) for point in clusters[i]])
        cluster_separation[i] = (sse_i / len(clusters[i]) + np.sum(centroid_distances[i])) / max_intracluster_distance
    return np.mean(cluster_separation)

def calculate_sc(clusters):
    num_points = np.sum([len(cluster) for cluster in clusters])
    distances = np.zeros((num_points, num_points))
    labels = np.zeros(num_points)
    index = 0
    for i, cluster in enumerate(clusters):
        for point in cluster:
            labels[index] = i
            for j in range(index+1, num_points):
                distances[index][j] = euclidean_distance(point, clusters[int(labels[j])][j-index-1])
                distances[j][index] = distances[index][j]
            index += 1
    a = np.zeros(num_points)
    for i in range(num_points):
        cluster_i = int(labels[i])
        a[i] = np.mean([distances[i][j] for j in range(num_points) if int(labels[j]) == cluster_i and i != j])
    b = np.zeros(num_points)
    for i in range(num_points):
        b[i] = np.min([np.mean([distances[i][j] for j in range(num_points) if int(labels[j]) == cluster_k and cluster_i != cluster_k]) 
                       for cluster_k in range(len(clusters)) if cluster_k != int(labels[i])])
    return np.mean((b-a) / np.maximum(a,b))

其中，`euclidean_distance`函数用于计算欧几里得距离，`calculate_centroid`函数用于计算聚类中心，`calculate_sse`函数用于计算聚类内误差平方和，`calculate_cp`函数用于计算聚类指标CP，`calculate_ch`函数用于计算聚类指标CH，`calculate_db`函数用于计算聚类指标DB，`calculate_sc`函数用于计算聚类指标SC。需要注意的是，这些函数都假设输入的聚类已经是三维坐标点的数组，因此不需要做额外的处理。