RALib

RALib is a library for active learning algorithms for register automata (a form of extended finite state machines). RALib is licensed under the Apache License, Version 2.0.

RALib is developed as an extension to LearnLib. It currently implements the following algorithms for learning register automata:

1. The SLλ algorithm by Simon Dierl, Paul Fiterau-Brostean, Falk Howar, Bengt Jonsson, Konstantinos Sagonas, and Fredrik Tåquist, presented in the paper Scalable Tree-based Register Automata Learning, Tools and Algorithms for the Construction and Analysis of Systems (TACAS 2024), pp 87-108.

2. The SL* algorithm by Sofia Cassel, Falk Howar, Bengt Jonsson, and Bernhard Steffen, presented in the paper Learning Extended Finite State Machines, Software Engineering and Formal Methods (SEFM 2014), pp 250-264.

Installation

RaLib uses maven as a build system. You can simply run

mvn package assembly:single

to build RaLib.

Using RALib

RALib can be used as a library from Java. The test cases that come with RALib demonstrate how this can be done. RALib currently also provides two tools that can be run from the shell. A 'simulator' for inferring RA models from simulated systems (automata) and a 'class analyzer' for inferring RA models of Java classes. Running

java -ea -jar target/ralib-0.1-SNAPSHOT-jar-with-dependencies.jar

will show some help and available options to the tools.

Below we provide two example configurations.

1. For learning a model of java.util.LinkedList with the class-analyzer call:

   java -ea -jar target/ralib-0.1-SNAPSHOT-jar-with-dependencies.jar  class-analyzer -f config
   

   with the following config file:

   target=java.util.LinkedList
   methods=push(java.lang.Object:int)boolean:boolean+\
   pop()java.lang.Object:int
   
   logging.level=WARNING
   max.time.millis=600000
   use.ceopt=true
   use.suffixopt=true
   use.fresh=false
   use.rwalk=true
   export.model=true
   rwalk.prob.fresh=0.8
   rwalk.prob.reset=0.1
   rwalk.max.depth=6
   rwalk.max.runs=10000
   rwalk.reset.count=false
   rwalk.draw.uniform=false
   teachers=int:de.learnlib.ralib.tools.theories.IntegerEqualityTheory
   

2. For learning a model of the SIP protocol with the iosimulator call:

   java -ea -jar target ralib-0.1-SNAPSHOT-jar-with-dependencies.jar iosimulator -f config
   

   with the following config file

   target=src/test/resources/de/learnlib/ralib/automata/xml/sip.xml
   
   logging.level=WARNING
   max.time.millis=600000
   use.eqtest=true
   use.ceopt=true
   use.suffixopt=true
   use.fresh=false
   use.rwalk=true
   export.model=true
   rwalk.prob.fresh=0.8
   rwalk.prob.reset=0.1
   rwalk.max.depth=100
   rwalk.max.runs=10000
   rwalk.reset.count=false
   rwalk.draw.uniform=false
   teachers=int:de.learnlib.ralib.tools.theories.IntegerEqualityTheory