Installation checklist

Here all obligatory and recommended dependencies are summarized.

This looks first of all like a huge list, but if you use a package manager, most of the dependencies are resolved automatically anyway. The installation scripts provided in the installation instructions show that a fully installation can be done with a few commands only!

Core

python 2.7.x
all development was done with python 2.7.x. Lower version numbers might work, but have not been tested so far.
python-dev
an installation of the python development headers is also required
numpy
python numpy is required for data handling
scipy >0.11
python scipy is used for data analysis
netCDF4 library
The netCDF4 library is the major backend for access to data files.
cdo >1.2
climate data operators are required in the benchmarking environment for efficient data preprocessing and are also used for proper calculation of grid cell areas. The latter is important in many analysis functions to have a proper area weighting.
hdf5 library
To access nc4 and hdf5 files, an installation of the HDF5 library is required. This is mandatory for usage of the CDO’s. See cdo installation instructions. (libhdf5-openmpi-dev)
netcdf library
also the installation of the netcdf library is required (libnetcdf-dev)
openmpi library
the developer version of the openmpi library is required as well for the cdo’s (libopenmpi-dev)
cdo python interface
The python interface is used to make the cdo’s and pyCMBS talk to each other
cython
cython allows for very fast data processing from python code. It is used by some of the analysis functions and in particular for the netCDF4 library
latex
a working installation of latex is required for generating automatic reports in the benchmarking framework. In particular your latex installation needs to support the pdflatex command.
nosetests
nosetests are mandatory to run all the unittests in the development enviroment