/vptree

Python implemenation of a vantage-point tree data structure for nearest neighbor searching.

Primary LanguagePythonMIT LicenseMIT

VP-Tree

This package contains an implementation of a vantage-point tree data structure.

Installation

Simply install through pip:

pip install vptree

Example

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)