libmata 1.2.3

The automata library

Mata: The Automata Library

Mata is an open source automata library that offers interface for different kinds of automata (NFA, etc.). Currently, Mata offers two interfaces:

1. An efficient library implemented in C/C++
2. A flexible wrapper implemented in Python that uses the efficient library

Requirements and dependencies

For a successful installation of Mata, cmake of version 3.15.0 (or higher) and a C++ compiler with a support of C++-20 standard is required. From optional requirements, doxygen is required for a generation of the documentation and catch2 is required for the unit testing (Mata can still be compiled without these optional dependencies).

The Mata library further depends on the following libraries, included in the 3rdparty directory:

• cudd for BDD manipulation,
• re2 for regular expression parsing and the corresponding automata construction, and
• simlib for a simulation computation.

Building and installing from sources

To build the library, run the following:

git clone https://github.com/VeriFIT/mata
cd mata
make release

In order to install the library, you can run

sudo make install

In order to verify the functionality of the library, you can run the test suite:

make test

You might, need to install the dependencies to measure the coverage of the tests. Run the following to install the dependencies for MacOS:

brew install lcov gcovr

Run the following to install the dependencies for Ubuntu:

sudo apt-get install -y build-essential lcov gcovr xdg-utils

Python binding

Mata offers binding of its efficient library to Python. You can install the binding as an Python package on your system as follows.

Install from PyPI

To install a latest version from the PyPI repository, run

pip3 install libmata

Building from sources

To build from sources first, install the necessary requirements for Python and your system. We recommend using the virtual environemnt (venv) to install and use the library.

python -m pip install --upgrade pip
make -C bindings/python init

sudo apt-get -qq update 
sudo apt-get -qq install -y graphviz graphviz-dev

Now, you can install the library.

make -C bindings/python install

Finally, you can verify the binding woks as expected by running the test suite:

make -C bindings/python test

Getting started

To get started, we refer to the examples in our repository. This directory contains examples of various usage in form of:

1. C/C++ example programs. By default, they are built with the library. To run for example the first example:

./build/examples/example01-simple

3. Python example scripts. To run the scripts run the following.

python examples/example01-python-binding.py

4. Python jupyter notebooks. To run the jupyter notebook, one needs to have jupyter installed as a prerequisite. The run the jupyter notebook, that creates an instance on your local server. Navigate to generated link to see the available jupyter notebooks:

pip3 install jupyter
jupyter notebook

Using the library

The library can be used directly in the C/C++ code. The result of compilation is a static or dynamic library, that can be linked to ones project. Note, that the library is dependent on several other 3rd party libraries (e.g., libre2 or libsimlib), which are included in the repository.

First import the library in your code. If the library is properly installed, you can use the standard include.

#include <mata/nfa/nfa.hh>

We recommend to use the mata::nfa namespace for easier usage:

using namespace mata::nfa;

Start by creating an automaton with fixed number of states.

int main() {
    Nfa aut(4);

You can set the initial and final states directly using the initializers.

    aut.initial = {0, 1};
    aut.final = {2, 3};

Further, you can add transitions in form of tripple (state_from, symbol, targets):

    aut.delta.add(0, 0, 2);
    aut.delta.add(1, 1, 3);

You can verify the state of your automaton by generating the automaton in .dot format.

    aut.print_to_DOT(std::cout);

    return 0;
}

We recommend cmake for building projects using Mata. Provided the Mata is installed in the system directories, the CMakeLists.txt file for the example may look like:

cmake_minimum_required (VERSION 3.15.0)
project (mata-example)
set (CMAKE_CXX_STANDARD 20)

find_library(LIBMATA mata REQUIRED)

add_executable(mata-example
    mata-example.cc)
target_link_libraries(mata-example PUBLIC ${LIBMATA})

Using the Python binding

The python binding is installed (by default) to your local python package repository. You can either use the binding in your own scripts or in the python interpreter.

You can start using the binding by importing the libmata package.

import libmata.nfa.nfa as mata_nfa

In your own scripts, we recommend to use the standard guard for running the scripts, as follows.

if __name__ == "__main__":

The usage of the binding copies (to certain levels) the usage of the C++ library.

    aut = mata_nfa.Nfa(4)

    aut.initial_states = {0, 1}
    aut.final_states = {2, 3}
    aut.add_transition(0, 0, 2)
    aut.add_transition(1, 1, 3)

    print(aut.to_dot_str())

You can either run your scripts directly using python or compile it using the cython project.

Publications

• Chocholatý, D., Fiedor, T., Havlena, V., Holík, L., Hruška, M., Lengál, O., & Síč, J. (2023). Mata, a Fast and Simple Finite Automata Library. arXiv preprint arXiv:2310.10136.
  • Chocholatý, D., Fiedor, T., Havlena, V., Holík, L., Hruška, M., Lengál, O., Síč, J.: A replication package for reproducing the results of paper “Mata: A fast and simple finite automata library” (Oct 2023).

Contributing

If you'd like to contribute to the libmata, please fork the repository, create a new feature branch, and finally create a new pull request.

In case you run into some unexpected behaviour, error or anything suspicious either contact us directly through mail or create a new issue. When creating a new issue, please, try to include everything necessary for us to know (such as the version, operation system, etc.) so we can sucessfully replicate the issue.

Note to main contributors

By default, each merge automatically increases the minor version of the library (i.e., 0.0.0 -> 0.1.0 ). This can be overruled using either tag #patch (increasing patch version, i.e., 0.0.0 -> 0.0.1) or #major (increasing major version, i.e., 0.0.0 -> 1.0.0). This tag is specified in the merge message.

Generally, it is recommended to use #major for changes that introduces backward-incompatible changes for people that used previous versions, and #patch for minor changes, such as bug-fixes, performance fixes or refactoring.

Links

• Project (origin) repository: https://github.com/verifit/mata
• Issue tracker: https://github.com/verifit/mata/issues
  • In case of some sensitive bugs (like security vulnerabilities), please contact us directly, instead of using issue tracker. We value your effort to improve the security and privacy of this project!
• Project documentation: https://verifit.github.io/mata
• Jupyter notebooks demonstrating mata usage: https://github.com/VeriFIT/mata/tree/devel/examples/notebooks
  • Example 01: WS1S Logic
  • Example 02: ReDoS Attacks
  • Example 03: Exploring Maze

Also, check out our research group focusing on program analysis, static and dynamic analysis, formal methods, verification and many more: http://www.fit.vutbr.cz/research/groups/verifit/index.php.en.

Licensing

The code of Mata is licensed under MIT licence. See LICENSE.

The folder 3rdparty/ contains 3rd party applications licensed under their own licences, included with the code.

Contacts

See AUTHORS.md

Acknowledgements

We thank for the support received from the Brno University of Technology (BUT FIT).

Development of this tool has been supported by the following projects: ???.

This tool as well as the information provided on this web page reflects only the author's view and no organization is responsible for any use that may be made of the information it contains.