Prolog bits and pieces

This page exists to collect some little bits of Prolog code I have lying around from various projects.  They're designed to work with SWI-Prolog, but in most cases should be reasonably easy to adapt to other environments.

Unless otherwise specified, these are all under the GNU GPL, version 2.

A vp-tree data structure

This collects points in a general metric space (such as N-dimensional vectors) and supports range queries - that is, you can put in a query point and extract all the points in the structure that are within a specified radius of the query point.  The data structure is reasonably efficient for the case it was designed for, which is when you have a lot of queries and relatively few inserts.  Inserts are not heavily optimized (sometimes large sections of the tree may be re-processed to ensure balance, and the reprocessing includes a quicksort-like median search which could be subject to pathological behaviour) but queries should be about as fast as the limits of the problem allow.  In high dimensions, with genuinely high-dimensional data, even the best possible data structures may take a long time to answer queries.  At some point I plan to enhance it to also support nearest-neighbour queries.  There is undocumented code included in the file but not exported, for testing and performance optimization.

The code, version 0.1 (19 March 2006).

Random selection from, and permutation of, a list

Just a few lines of code which I've found myself using over and over in Prolog projects.

The code, version 0.2 (16 April 2006).