This package contains an implementation of a vantage-point tree data structure.
Simply install through pip:
pip install vptree
Example usage:
import numpy as np
import vptree
# Define distance function.
def euclidean(p1, p2):
return np.sqrt(np.sum(np.power(p2 - p1, 2)))
# Generate some random points.
points = np.random.randn(20000, 10)
query = [.5] * 10
# Build tree in O(n log n) time complexity.
tree = vptree.VPTree(points, euclidean)
# Query single point.
tree.get_nearest_neighbor(query)
# Query n-points.
tree.get_n_nearest_neighbors(query, 10)
# Get all points within certain distance.
tree.get_all_in_range(query, 3.14)