README

TODO: write about the idea, the roadmap, the existing files, make README .rst

This is a Python 2 project meant to be run on Unix platforms. If you run windows, install Linux on a VM. Virtualbox works and is free.

Non-python dependencies:

• NEST (NEST Neural Simulator) + PyNEST OR

• Brian (http://www.briansimulator.org)

• VTK (The Visualization Toolkit) + Python 2 bindings

Python pip-installable dependencies:

• coverage
• PyNN
• SimPy
• mock
• nose
• numpy ((( - ctype, a dep of python-magic, needs to be installed like that (https://github.com/pypa/pip/issues/317): pip install --no-install ctypes cd /build/ctypes/source/libffi) chmod +x configure pip install ctypes )))
• python-magic
• PIL

if using Brian:

• scipy which requires:
• a fortran compiler if you pip install it (On Arch, use gcc-fortran)
• the basic linear algebra subprograms library (blas on Arch)
• the Linear Algebra PACKage (lapack)
• sympy, note that a bug in package numbering semantics may prevent you from pip installing it, in which case you should request for the exact latest version: pip install sympy==0.7.1
• matplotlib

For now, switching between simulators is done manually. For instance, switching from PyNN to Brian: grep -ilr 'pyNN.nest' .py /.py //.py | xargs sed -i 's/pyNN.nest/pyNN.brian/g'

The project should be installed locally as a non-priviledged user except if you know what you are doing.

Install NEST under your home directory, e.g. ~/local-install/

Modify your environment setup script: NEST_INSTALL_DIR=$HOME/local-install NEST_SUBFOLDER_PATH=$HOME/nest NESTRCFILENAME=$NEST_SUBFOLDER_PATH/config/nestrc NEST_DATA_PATH=$NEST_SUBFOLDER_PATH/nest_data NEST_DATA_PREFIX="out_" export NEST_INSTALL_DIR NESTRCFILENAME NEST_DATA_PATH NEST_DATA_PREFIX export LD_LIBRARY_PATH=$NEST_INSTALL_DIR/lib/nest:$LD_LIBRARY_PATH

source it and mkdir -p $NEST_DATA_PATH$NEST_DATA_PATH/../config

edit your /etc/ld.so.conf, add there /home/<>/local-install/lib/nest and run ldconfig

Install VTK using your OS's package manager, e.g. apt-get install vtk, pacman -S vtk, emerge vtk

Do the following modifications to your environment: Add $HOME/local-install/usr/bin:$HOME/local-install/usr/local/bin:$HOME/local-install/bin to your PATH Add .:$HOME/local-install/lib/python2.7/site-packages to your PYTHONPATH, additionally to the place of your system-wide python library (soemthing like /usr/lib/python2.7/site-packages)

Anstall virtualenv and virtualenvwrapper (with pip or your distro)

Modify your environment setup script again: export WORKON_HOME=your_python_virtualenv/directory, for instance $HOME/virtualenvs. Remember to mkdir$HOME/virtualenvs export VIRTUALENVWRAPPER_PYTHON=which python2, or which python2.7 or whatever your python 2 binary is called

Add source /usr/local/bin/virtualenvwrapper.sh to your .bash_profile Add alias ipython='python2 -c "from IPython import embed; embed()" 2>/dev/null || python2 -c "import IPython; IPython.Shell.IPShell().mainloop()"' to your .bashrc

Have bash_profile source bashrc.

mkvirtualenv cogneuro mkdir -p dev/cogneuro/ && cd dev/cogneuro pip install PyNN mock nose git clone git://github.com/agravier/pycogmo.git cd pycogmo/src/attention_net/ python attention_net.py

Detailed instructions to install NEST 2.0.0-rc4

sudo apt-get install build-essential autoconf automake libtool libltdl-dev libreadline5-dev libncurses5-dev libgsl0-dev python-numpy python-scipy python-matplotlib ipython cd nest-2.0.0rc4 ./configure --prefix=$HOME/local-install PYTHON_EXEC='/usr/bin/python2.7' --with-python='/usr/bin/python2.7' && make && make install && make installcheck

to create an environment: mkvirtualenv to activate an environment: . ~/venvs//bin/activate e.g. . ~/virtualenvs/cogneuro/bin/activate to deactivate it: deactivate to install something in the virtual env: pip install

useful: pip install yolk yolk -l