Top > Research > Tools > Samurai-clone

Samurai-clone

Summary

Samurai-clone splits identifiers into words.
This is an implementation (a clone) of the Samurai algorithm.

Download

Perl script: samurai.pl Ver. 1 (6.5KB)
Dictionary file: dict.csv (67MB) (compressed file: dict.csv (20MB))
Dictionary file (You need to download from the Samurai official site): prefixes.txt / suffixes.txt

Install

Download the above perl script and three dictionary files.
Notice that all dictionary files must be in the same directory.
Then, specify the dictionary directory in the heading of the perl script as "$samurai_clone::DICT_DIR = ", where the directory path must be specified by using the fully qualified path name.

Usage

perl samurai.pl source_file (-s identifier | -f identifier_file)

options
- source_file: the source file in which the target identifier appears.
- -s identifier: specify the target identifier to be split.
- -f identifier_file: specify the text file in which the target identifiers are listed; one line, one identifier.

(Example)

$ perl samurai.pl xxxx.java -s ASTVisitor
This is a clone of "Samurai" for Java programs.
now loading dictionaries ... done.
ASTVisitor: AST Visitor

Reference

Eric Enslen, Emily Hill, Lori Pollock and K. Vijay-Shanker,
"Mining Source Code to Automatically Split Identifiers for Software Analysis,"
Proc. 6th IEEE International Working Conference on Mining Software Repositories, pp. 71--80, May 2009.