Installation

kima's core is written in C++ and needs to be compiled.

You will need to install a fairly recent version of the gcc compiler (at least GCC 4.8.1, see here). You can check the gcc version by typing gcc --version in the terminal.

To analyse the results, kima comes with a helper Python package, pykima.
This package depends on some (fairly standard) Python packages:

• numpy, scipy, and matplotlib
• corner
• loguniform
• kumaraswamy

download

To get the latest version of the code run

git clone --recursive https://github.com/j-faria/kima.git

which will also download the required DNest4 submodule.
Don't forget the --recursive or the installation will fail.

install

Change to the kima directory just created and type
(feel free to add "-j 4" to the make command in order to speed things up)

make
python setup.py install

This will compile the C++ code together with the examples and install the pykima package.

Note: you may choose to analyse the results without pykima. The inputs and outputs of kima itself are text files.

If everything went as planned, move on to the getting started guide.
If not, take a look at troubleshooting.