This project is my contribution to Jan Daciuk's FSA package . It provides a set of Java classes and methods to load, traverse and in general utilize the power of dictionaries produced using FSA tools (those tools are written in C++).

I wrote these classes simply because I needed them in my Polish Stemmer project. Thus I kept the implementation to the very minimum: only version 5 of dictionaries encoding produced by Jan Daciuk's tools is supported (tools compiled with FLEXIBLE, STOPBIT and NEXTBIT).

I hope that in the future I will have some more time to implement other storage schemes introduced in FSA tools and possibly implement automaton-building algorithm as well, so that FSA is available not only for Unix environment, but also for any Java-compatible platform.

The package also includes some helpful utilities (may be invoked from command-line).

• com.dawidweiss.fsa.util.FSADump will dump a dictionary to standard output (automatic version recognition)
• com.dawidweiss.fsa.util.FSAFind will attempt to locate words from the standard input in the dictionary.

FSA classes are available as part of the Polish stemmer distribution (Lametyzator)