请帮我写一份将点云建立八叉树结构，并能搜索某个点10m内其他点的代码

以下是点云建立八叉树结构并搜索某个点10m内其他点的 Python 代码：

import numpy as np
from sklearn.neighbors import KDTree


class Octree:
    def __init__(self, points, max_depth=10, min_points=5):
        self.points = points
        self.max_depth = max_depth
        self.min_points = min_points
        self.center = np.mean(points, axis=0)
        self.size = np.max(points, axis=0) - np.min(points, axis=0)
        self.children = []
        self.depth = 0
        if len(self.points) > self.min_points and self.depth < self.max_depth:
            self.split()

    def split(self):
        for i in range(2):
            for j in range(2):
                for k in range(2):
                    mask = np.logical_and(
                        np.logical_and(self.points[:, 0] >= self.center[0] + (i - 0.5) * self.size[0] / 2,
                                       self.points[:, 0] < self.center[0] + (i + 0.5) * self.size[0] / 2),
                        np.logical_and(self.points[:, 1] >= self.center[1] + (j - 0.5) * self.size[1] / 2,
                                       self.points[:, 1] < self.center[1] + (j + 0.5) * self.size[1] / 2),
                        np.logical_and(self.points[:, 2] >= self.center[2] + (k - 0.5) * self.size[2] / 2,
                                       self.points[:, 2] < self.center[2] + (k + 0.5) * self.size[2] / 2)
                    )
                    if np.sum(mask) > 0:
                        child = Octree(self.points[mask], max_depth=self.max_depth, min_points=self.min_points)
                        child.center = self.center + (np.array([i, j, k]) - 0.5) * self.size / 2
                        child.size = self.size / 2
                        child.depth = self.depth + 1
                        self.children.append(child)

    def query(self, point, radius):
        if len(self.children) == 0:
            return self.points
        elif np.linalg.norm(point - self.center) > radius + np.linalg.norm(self.size) / 2:
            return []
        else:
            result = []
            for child in self.children:
                result += child.query(point, radius)
            return result


# 生成随机点云
points = np.random.rand(1000, 3)

# 建立八叉树
octree = Octree(points)

# 随机选一个点
query_point = points[np.random.randint(0, len(points))]

# 搜索10m内的点
neighbors = octree.query(query_point, 10)

# 使用KDTree检验结果
kd_tree = KDTree(points)
kd_neighbors = kd_tree.query_radius(query_point.reshape(1, -1), 10)[0]

# 比较结果
assert set(neighbors) == set(kd_neighbors)

代码中使用了一个 `Octree` 类来建立八叉树。在类的初始化函数中，会计算点云的中心点和大小，并根据最大深度和最小点数来决定是否需要继续分割为子节点。在 `split` 函数中，会按照八叉树的规则将点云分割为八个部分，并递归地创建子节点。在 `query` 函数中，会根据查询点的距离和半径来递归地搜索八叉树，并返回所有在半径内的点。

最后，使用了 `KDTree` 来检验结果。