A Java implementation of Earley’s efficient parsing algorithm for lambda-free context-free grammars. The name comes from J-Earley.

Originally, I wrote this project in 2004-2005 when I was in France. I decided to put it on GitHub in 2017. I have no intention to maintain this project, so don't send me any pull requests. If you want to add / modify something, feel free to make a fork.

In 2017 I tried this project and it still works fine.

I also wrote a short (unpublished) paper about the Earley algorithm, see the file jerly.pdf. It explains the algorithm, and it also presents this software shortly. It also has an example.

Usage: jerly [switches] <rule_file> <question_file>
Switches:
   --help                          help information (you can see this now)
                                   can be used _alone_
   --version, -V                   version information
                                   can be used _alone_

   rule_file                       rules of the grammar
   question_file                   question (word)

How to try it right away

./start input/paper.txt input/question.txt

Rules:
======
S->S+A
S->A
A->AxB
A->B
B->(s)
B->a


Question:
=========
axa+a

The word 'axa+a' can be generated with the grammar.

HTML output was successfully written to dir. 'output'.

The software also creates an HTML output with a nice table. It can be found in the output folder (see screenshot).

• Laszlo Szathmary
• jabba.laci@gmail.com (this is my current e-mail)