Harry - A Tool for Measuring String Similarity

This software belongs to the publication

Konrad Rieck and Christian Wressnegger. Harry: A Tool for Measuring String Similarity. Journal of Machine Learning Research (JMLR), 17(9): 1-5, 2016 http://jmlr.org/papers/volume17/rieck16a/rieck16a.pdf

Description

Harry is a small tool for comparing strings. The tool supports several common distance and kernel functions for strings as well as some excotic similarity measures. The focus of Harry lies on implicit similarity measures, that is, comparison functions that do not give rise to an explicit vector space. Examples of such similarity measures are the Levenshtein distance, the Jaro-Winkler distance or the spectrum kernel.

During operation Harry loads a set of strings from input, computes the specified similarity measure and writes a matrix of similarity values to output. The similarity measure can be computed based on the granulartiy of bytes, bits and tokens (words) contained in the strings. The configuration of this process, such as the input format, the similarity measure and the output format, are specified in a configuration file and can be additionally refined using command-line options.

Harry is implemented using OpenMP, such that the computation time for a set of strings scales linear with the number of available CPU cores. Moreover, efficient implementations of several similarity measures, effective caching of similarity values and low-overhead locking further speedup the computation.

Harry complements the tool sally(1) that embeds strings in a vector space and allows computing vectorial similarity measures, such as the cosine distance and the bag-of-words kernel.

Similarity Measures

The following similarity measures are supported so by Harry

dist_bag             Bag distance
dist_compression     Normalized compression distance (NCD)
dist_damerau         Damerau-Levenshtein distance
dist_hamming         Hamming distance
dist_jaro            Jaro distance
dist_jarowinkler     Jaro-Winkler distance
dist_kernel          Kernel-based distance
dist_lee             Lee distance
dist_levenshtein     Levenshtein distance
dist_osa             Optimal string alignment (OSA) distance
kern_distance        Distance substitution kernel (DSK)
kern_spectrum        Spectrum kernel
kern_subsequence     Subsequence kernel (SSK)
kern_wdegree         Weighted-degree kernel (WDK)
sim_braun            Braun-Blanquet coefficient
sim_dice             Soerensen-Dice coefficient
sim_jaccard          Jaccard coefficient
sim_kulczynski       second Kulczynski coefficient
sim_otsuka           Otsuka coefficient
sim_simpson          Simpson coefficient
sim_sokal            Sokal-Sneath coefficient

Dependencies

• OpenMP >= 2.5 (need to be supported by the C compiler)
• zlib >= 1.2.1, http://www.zlib.net/
• libconfig >= 1.3.2, http://www.hyperrealm.com/libconfig/
• libarchive >= 3.1.2, http://libarchive.github.com/

Debian & Ubuntu Linux

The following packages need to be installed for compiling Harry on Debian and Ubuntu Linux

gcc
libz-dev
libconfig8-dev
libarchive-dev

For bootstrapping Harry from the GIT repository or manipulating the automake/autoconf configuration, the following additional packages are necessary.

automake
autoconf
libtool

Mac OS X

For compiling Harry on Mac OS X a working installation of Xcode is needed. Moreover, a C compiler supporting OpenMP is required (clang from Xcode currently does not support OpenMP). The following packages need to be installed from Homebrew.

gcc43 (or download from <http://hpc.sourceforge.net>)
libconfig
libarchive (from homebrew-alt)

OpenBSD

Due to the vague state of OpenBSD multi-threading, neither the default gcc nor the packaged gcc seem to correctly support OpenMP. To get Harry to run you can only try to build gcc from scratch

:(

Compilation & Installation

From GIT repository first run

./bootstrap

From tarball run

./configure [options]
make
make check
make install

Options for configure

--prefix=PATH           Set directory prefix for installation

By default Harry is installed into /usr/local. If you prefer a different location, use this option to select an installation directory.

--enable-prwlock        Enable support for POSIX read-write locks

This feature enables read-write locks (rwlocks) from the POSIX thread library. The locks can accelerate the run-time performance on multi-core systems. However, these POSIX locks are not guaranteed to interplay with OpenMP and thus may not work on all platforms.

--enable-md5hash        Enable MD5 as alternative hash

Harry uses a hash function for mapping tokens to symbols. By default the very efficient Murmur hash is used for this task. In certain critical cases it may be useful to use a cryptographic hash as MD5.