CSharp (C#) version of vantage-point tree (VP tree)

vptree

This is one implementation of vantage-point tree (VP tree) for C#. As usual most of the error handling code has been stripped away, so YMMV.

Vantage-point tree as Wiki says: "A vantage-point tree (or VP tree) is a metric tree that segregates data in a metric space by choosing a position in the space (the "vantage point") and partitioning the data points into two parts: those points that are nearer to the vantage point than a threshold, and those points that are not. By recursively applying this procedure to partition the data into smaller and smaller sets, a tree data structure is created where neighbors in the tree are likely to be neighbors in the space."

So VP tree (once created) can be used to find data that is distance-wise near certain point. e.g. this project provides implementation (in Example.cs) that allows one to find X nearest cities of given latitude and longitude point.

This project is a C# port of C++ code that was given to the world by Steve Hanov in his blog.

VpTree.cs has the VP tree related code (you can copy + paste this file to your project).
LinearSearch.cs provides a basic linear search that you can use for performance comparision.
Example.cs gives sample code and some performance measurements.
tests folder contains test cases.

You can get the full cities.txt file used in Example.cs and Test.cs from here. When extracted the size of the file is 125 806 517 bytes (~119MB). The cities.txt contains duplicate latitude and longitude entries for certain places, so order of results might vary on different runs.

You can find smaller (and zipped) cities example file from tests/testfiles folder

Biggest change is that this C# version uses zero based indexing and not iterators (like the C++ version does).

As Steve Hanov said in his blog: "It is worth repeating that you must use a distance metric that satisfies the triangle inequality. I spent a lot of time wondering why my VP tree was not working. It turns out that I had not bothered to find the square root in the distance calculation. This step is important to satisfy the requirements of a metric space, because if the straight line distance to a <= b+c, it does not necessarily follow that a^2 <= b^2 + c^2."

Which means that square root calculations for distance metrics are MANDATORY!

VpTree<Point> vpTree = new VpTree<Point>();
vpTree.Create(points, CalculatePointDistance);
vpTree.Search(targetPoint, 5, out resultsVpTree, out distancesVpTree);

You can run benchmarks (which compare vptree and linear search) by moving to benchmarks folder and running following command

dotnet run -c Release

This document and source code files are released into the public domain. See LICENSE file