Building Stimfit¶

Author:	Jose Guzman, Christoph Schmidt-Hieber
Date:	20 February, 2019

This document describes how to install Stimfit 0.15.8 under GNU/Linux. The installation was tested on a GNU/Debian testing/unstable with support for Python 2.*. It should work on other Debian-based systems (e.g. Ubuntu) as with newer version of Stimfit as well. I assume that you have the GNU C compiler (gcc) and the GNU C++ compiler (g++) and that both versions match. For this installation, we tested 4.2.4 versions.

What we need before we start¶

For the impatient, here are all Stimfit build dependencies:

$ sudo apt-get install libboost-dev \
                       python-dev \
                       python-numpy \
                       python-matplotlib \
                       libhdf5-serial-dev \
                       swig \
                       python-wxgtk3.0-dev \
                       libwxgtk3.0-dev \
                       wx-common \
                       fftw3-dev \
                       liblapack-dev \
                       chrpath \
                       git \
                       automake \
                       autoconf \
                       libtool \
                       libgtest-dev

This will get you, amongst others:

[wxWidgets]: C++ graphical user interface toolkit (version >= 2.8; tested with 2.8.12 and 3.0.4)
[wxPython]: GUI toolkit for the Python language.
[boost]: C++ library that is mainly used for its shared pointers.
[Lapack]: A linear algebra library.
[fftw]: Library for computing Fourier transformations.
[NumPy]: To handle numerical computations with Python (use version >=1.7.1).
[HDF5]: Hierarchical Data Format 5 (HDF5) to manage large amount of data.
[Matplotlib]: Plotting library for Python (use version >= 1.5.1)

In addition, you can install doxygen, python-sphinx (with graphviz and Latex) if you want to build the documentation.

Optional: PyEMF¶

[PyEMF] is needed to export figures to the windows meta file format (WMF/EMF). EMF is a vector graphics format and can be imported in different Office software including LibreOffice. To install it, do:

$ wget http://sourceforge.net/projects/pyemf/files/pyemf/2.0.0/pyemf-2.0.0.tar.gz/download -O pyemf-2.0.0.tar.gz
$ tar xvf pyemf-2.0.0.tar.gz && cd pyemf-2.0.0
$ sudo python setup.py install

Download the Stimfit source code¶

You can download the latest development code for Stimfit from the Github code repository. For that, type from your current $HOME directory:

$ git clone https://github.com/neurodroid/stimfit.git

This will grab all the required files into $HOME/stimfit. If you’d like to update at a later point, do:

$ cd $HOME/stimfit
$ git pull

Build Stimfit¶

Go to the stimfit directory (in our example $HOME/stimfit) and type:

$ cd $HOME/stimfit
$ ./autogen.sh

to generate the configure script. Remember that we need Autoconf, Automake and LibTool to use autogen. After that, you can call it with

$ ./configure

The –enable-python option is activated as a default.

Finally, after running configure, you can type

$ make -j[N]

where [N] is the number of parallel builds you want to start. And finally:

$ sudo make install
$ sudo /sbin/ldconfig

Note

If you want to install Stimfit as local user (e.g in ~/.local) with a local version of Python (e.g ~/.local/lib/python2.6) you have to add the following argument to configure script:

$ PYTHON=$HOME/.local/lib/python2.6 ./configure --prefix=$HOME/.local

and after that call make and make install as normal user. The Stimfit executable will be now in $HOME/.local

Building Stimfit with BioSig import filter¶

We recommend to build Stimfit with the BioSig library to import files in from different biomedical disciplines. It is necessary to read files acquired with HEKA amplifiers or with Signal from CED. To do it, follow this instructions:

Install libsuitesparse and libz libraries:

sudo apt-get install libsuitesparse-dev libz-dev gawk

To avoid ABI incompatibilities, we recommend building Stimfit with the version of biosig that ships with the Stimfit source:

$ ./configure --with-biosiglite

Alternatively, get the full version of biosig: you can obtain the latest BioSig version in BioSig downloads . Choose BioSig for C/C++, libbiosig (v1.5.6 or higher is recommended). Alternatively, you can obtain the latest developmental version from the git repository:

git clone https://git.code.sf.net/p/biosig/code biosig-code

Enter the directory biosig4c++ and compile and install the sources:

cd biosig-code/biosig4c++
autoconf # needed first time after getting repository
./configure
make
sudo make install

After that you can enter the option –with-biosig in the configure script of Stimfit and compile as usual.

Building documentation¶

The manual of Stimfit including the documentation is accessible on-line in http://www.stimfit.org/doc/sphix/. To have your local copy, you will need to install sphinx version 1.7 or older:

sudo apt-get install python-sphinx

To build a local copy call:

sphinx-build $HOME/Stimfit/doc/sphinx/ <destination> # destination folder could be $HOME/tmp/stf/doc/

The html documentation will be located in <destination>/index.html

Additionally, the source code is documented with [Doxygen] and is also accessible on-line in http://www.stimfit.org/doc/doxygen/html/. If you want to have a local copy of the documentation, you will need to install the doxygen and gravphvix:

sudo apt-get install doxygen gravphvix

Enter a directory called doc inside Stimfit (e.g $HOME/stimfit/doc) and type:

cd $HOME/stimfit/doc
doxygen DoxyFile

The local documentation of the source code will be in $HOME/stimfit/doc/doxygen/html

[wxWidgets]

http://www.wxwidgets.org

[wxPython]

http://www.wxpython.org

[boost]

http://www.boost.org

[Lapack]

http://www.netlib.org/lapack/

[HDF5]

http://www.hdfgroup.org/HDF5/

[NumPy]

http://www.numpy.org

[PyEMF]

http://pyemf.sourceforge.net

[fftw]

http://www.fftw.org

[Doxygen]

http://www.doxygen.org

[Matplotlib]

https://matplotlib.org