benjaminhr / eye-oh

A set of tools for automata learning with Tomte

eye-oh

A set of tools for converting π-calculus processes into Register Automata in XML. Each component can be run in isolation or the whole pipeline can be run at once. Additionally, there exists a tool for converting Register Automata from Tomte's XML format to DEQ's XML format. Pifra is used to convert π-calculus processes to a labelled transition representing FRAs.

Below we can see a high level overview of all the components and their relevant data formats.

Notes about π-calculus definition files

• file extension must be .pi
• syntax must conform to Pifra's syntax

Notes about input model

• initial state must have initial="true"
• states with no outgoing transitions will be considered final
• relies on input transitions starting with I, output with O (as does Tomte)
• make sure not to have states prefixed vk_N and kk_ or named sink and sink_two in your input model already

Notes about output model

• the output symbols ODummy, OOK, and OFinal will be created
• two sink states will be added, which will consume the added input enabling transitions
• to make alternating input/output transitions, the model will add OOK transitions inbetween

Usage

• npm run package && npm run link to get binary installed in /usr/local/bin
• eye-oh --help to list all available commands
• deq-helper.sh is used to in the output directory with a learnedConcreteModel.xml and a test_processes/temp.pi to produce deq models (handles all pruning, converting, etc).