# CoPaR, the `Co`algebraic `Pa`rtition `R`efiner

Copar implements a generic partition refinement algorithm with a fast run-time.
The genericity of copar covers the minimization of

  - Transition Systems (under strong bisimilarity)
  - (Tree) Automata (under tree isomorphism)
  - Markov Chains, and other weighed systems (under weighed bisimilarity)
  - and *combinations* thereof (under derived bismilarity notions, e.g. backwards bisimilarity for weighted tree automata).

The algorithm runs in *O(m log n)* on a system with *n* states and *m*
transitions and has been developed in a series of publications, see the [literature file](doc/literature.md).

## Usage
Installation instructions can be found in the [install file](INSTALL.md).

The core partition-refinement algorithm can be called by

    ./copar refine FILENAME

Examples for input files are in the directory `examples/`.

Command line syntax is described in `copar --help` and the input
file syntax in `copar help`.

## Benchmarks

Extensive benchmarks of CoPaR on different input files can be found at
[git8.cs.fau.de/software/copar-benchmarks](https://git8.cs.fau.de/software/copar-benchmarks).

Micro-Benchmarks for internal data structures and routines are available in the
`bench/` directory and can be executed with `stack bench`.