Python wrapper for selected SLICOT routines, notably including solvers for Riccati, Lyapunov, and Sylvester equations.

Slycot supports Python versions 3.10 or later.

To run the compiled Slycot package, the following must be installed as dependencies:

• Python 3.10+
• NumPy

If you are compiling and installing Slycot from source, you will need the following dependencies:

• Python 3.10+
• NumPy
• scikit-build
• CMake
• C compiler (e.g. gcc, MS Visual C++, clang)
• FORTRAN compiler (e.g. gfortran, ifort, flang)
• BLAS/LAPACK (e.g. OpenBLAS, ATLAS, MKL)

To run the Slycot unit tests and examples, you'll also need SciPy and pytest.

There are a variety of ways to install these dependencies on different operating systems. See the individual packages' documentation for options.

The easiest way to get started with Slycot is to install pre-compiled binaries from conda-forge (see below); these are available for Linux, OSX, and Windows.

Compiling the Slycot source is unfortunately a bit tricky, especially on Windows, but we give some pointers further below for doing this.

First install Miniconda or Anaconda. Slycot can then be installed from the conda-forge channel with the following command:

conda install -c conda-forge slycot

The hardest part about installing from source is getting a working version of FORTRAN and LAPACK (provided by OpenBLAS, MKL, etc.) installed on your system. Depending on where you get your NumPy and SciPy from, you will need to use a compatible LAPACK implementation. Make sure that the correct header files are installed, and specify the CMake variable BLA_VENDOR, if necessary. We recommend to use BLA_VENDOR=Generic in order to produce a Slycot module, which is binary compatible with all implementations.

We publish Slycot to the Python package index, but only as a source package, so to install using pip you'll first need to install the build prerequisites (compilers, libraries, etc.)

If you have these build prerequisites, the command:

pip install slycot

will download the latest release of the source code from PyPI, compile, and install Slycot into the currently configured location (virtual environment or user site-packages).

Get the source code of the latest release is available from PyPI. It contains both the Python to Fortran wrappers as well as the SLICOT-Reference Fortran sources.

If you decide to download a source code archive from GitHub (tagged release or a specific branch), you also have to get the correct version of our SLICOT-Reference fork and place it into slycot/src/SLICOT-Reference:

1. Download and unpack https://github.com/python-control/Slycot/archive/master.zip
2. Go to https://github.com/python-control/Slycot/master/slycot/src
3. Follow the link of SLICOT-Reference @ <commit-id>
4. Download the archive of SLICOT-Reference from the Code download button (https://github.com/python-control/SLICOT-Reference/archive/<commit-id>.zip)
5. Unpack the contents of the SLICOT-Reference archive into slycot/src/SLICOT-Reference

Replace master with the release tag or branch name, which you want to build.

Directly checkout the submodule, when cloning the git repository:

git clone --recurse-submodules https://github.com/python-control/Slycot.git

or if you forked the repository:

git clone --recurse-submodules https://github.com/<your-username>/Slycot.git

If you already have a local checkout, but still need to init the submodule:

git submodule init
git submodule update

If you need to specify a specific compiler, set the environment variable FC before running the install:

# Linux/OSX:
export FC=/path/to/my/fortran

# Windows:
set FC=D:\path\to\my\fortran.exe

To build and install, execute:

cd /path/to/slycot_src/
pip install -v .

You can use conda to compile and install Slycot from source. The recipe is located in the folder conda-recipe and is intended to work for all platforms.

The conda-forge channel provides almost all requirements to compile Slycot with conda-build, except:

• On macOS, you need the macOS SDK. See the conda-build documentation for macOS how to get it.
• On Windows, you need to install Microsoft Visual C++ 14.x provided e.g. by Microsoft Visual Studio. To build, you'll need a command shell setup for both conda and the Visual Studio build tools. See conda activation and Microsoft Visual Studio setup for information on this.

To build and install:

conda build -c conda-forge conda-recipe
conda install -c conda-forge --use-local slycot

A similar method can be used for Linux and macOS, but is detailed here for Windows. This method uses conda and conda-forge to get most build dependencies, except for the C compiler.

1. Install Microsoft Visual Studio.
2. Unpack the source code to a directory of your choice,
3. Create a command shell setup that can run the conda commands and the Visual Studio build tools (see above)

4. In such a command shell, within the Slycot source code directory, run the following commands to build and install Slycot (this example creates a Python 3.8 environment):

   conda create --channel conda-forge --name build-slycot python=3.10 numpy scipy libblas=*=*netlib liblapack=*=*netlib scikit-build flang pytest
   conda activate build-slycot
   
   pip install -v .

Additional hints for how to install Slycot from source can be found in the .github directory , (commands used to build and test in the GitHub Actions CI), the logs from the GitHub Actions, and the conda-recipe directory (conda pre-requisites, install and test commands) which is included in the source code repository.

To test if the installation was successful, you can run the slycot unit tests:

pytest --pyargs slycot

You may also run the tests by calling slycot.test() from within the python interpreter:

import slycot
slycot.test()

Importing slycot or running pytest without --pyargs slycot from inside the source directory will fail, unless the compiled wrapper library has been installed into that directory. Note that the [tool:pytest] section in setup.cfg enforces the --pyargs slycot argument by default.

Up until version 0.4, Slycot used a version of SLICOT that was released under the GPLv2 license. This requires Slycot to be released under the same license. In December 2020, SLICOT 5.7 was released under BSD-3-Clause. However, as the existing Slycot wrappers have been submitted by many contributors, we cannot move away from GPLv2 unless we get the permission to do so by all authors. Thus, Slycot remains licensed under GPLv2 until further notice.