Everthing should be already installed except for pymultinest. This is installed under the home directory (cd ~) the following commands:
git clone https://github.com/JohannesBuchner/PyMultiNest/
cd PyMultiNest
python3 setup.py install --user
Then cd back to the top directory above PyMutliNest
git clone https://github.com/JohannesBuchner/MultiNest
cd MultiNest/build
cmake ..
make
The hard part is getting pymultinest to recognise multinest. One of these should do the job...
echo 'export LD_LIBRARY_PATH=/home/nicholas.farrow/MultiNest/lib/:$LD_LIBRARY_PATH'
echo 'export LD_LIBRARY_PATH=/home/FIRSTNAME.LASTNAME/MultiNest/lib/:$LD_LIBRARY_PATH' >>~/.bashrc
Check if this worked by using python3 then import pymultinest. If nothing shows up then everthing is working.
This should be done first to check that all the required packages are installed (especially pymultinest). Sampling using a specific model (or model pairs) can be done through calling
python3 hypothesisB.py 3
Which runs hypothesis B for the 3rd model pair (single - uniform).
As sampling can take very long periods of time (12 combinations of sub-hypotheses ranging between 2 and 7 parameters), it is required that sampling be done through a distributed computing system. The method of providing a number argument to the code allows for easy batch running of the script.
python3 hypothesisA.py [1-3].
python3 hypothesisB.py [1-9].
These scripts analyse the evidence for all of the 9 or 3 sub-hypotheses, and also aggregate a total evidence for each hypothesis A and B.
python3 hypothesisAnalyse.py A hypothesisA_out/
python3 hypothesisAnalyse.py B hypothesisB_out/
This creates files like evidence lists, relative evidence tables and totals, and also latex tables.