GNU/Linux

Installation was tested on limited number of GNU/Linux systems. On the most of modern installations you can simply follow generic instructions, for example under Ubuntu 16.04 you can type:

sudo pip install aquaduct

NetCDF4 & MDAnalysis installation Ubuntu 14.04

Other systems may require additional work, in particular installation of NetCDF4 is sometimes cumbersome. Following is an example how to install all required packages under Ubuntu 14.04:

# install required python packages
sudo apt-get install python-dev python-pip python-numpy python-scipy python-matplotlib python-scikits-learn

# install necessary libraries and git -  all required to compile netCDF4
sudo apt-get -y install libnetcdf-dev libhdf5-dev git

# clone netcdf4 python repository
git clone https://github.com/Unidata/netcdf4-python.git
# cd to cloned repository
cd netcdf4-python
# modify setup.cfg to add paths of hdf5 and netcdf4 libraries
sed -i '/\[directories\]/a \
HDF5_dir = /usr/lib \
HDF5_libdir = /usr/lib \
HDF5_incdir = /usr/include \
netCDF4_dir = /usr/lib \
netCDF4_libdir = /usr/lib \
netCDF4_incdir = /usr/include' setup.cfg
# run setup.py
sudo python setup.py install

# install MDAnalysis
sudo pip install "MDAnalysis[amber]>=0.15"

SciPy update and Ubuntu/Debian

Debian (and Ubuntu) uses strange approach to Python installation. To install newer version of SciPy (if required) try following procedure:

# install libraries required for SciPy compilation
apt-get build-dep python-scipy

# install SciPy
easy_install-2.7 --upgrade scipy

Warning: The above procedure will remove current SciPy from easy-install.pth file.

If everything went fine you can follow generic instructions, type:

sudo pip install aquaduct