This repository contains the code used for making the results and plots in "nestcheck: diagnostic tests for nested sampling calculations" (Higson et. al, 2019). This provides examples of the use of the nestcheck package.
If you have any questions then feel free to email [email protected]. However, note that this is research code and is not actively maintained.
Generating the results in the paper requires PolyChord v1.14, plus the requirements listed in setup.py. Results in the paper were run using Python 3.6, nestcheck v0.1.6 and PolyChord v1.14.
The diagnostic Python module contains high level functions for generating and plotting the results in the paper. Most of this is just convenient wrappers and stored settings for using the nestcheck module, which contains implementations of the tests introduced in the paper. diagnostic can be installed, along with nestcheck and its other dependencies, by running the following command from within this repo:
pip install . --user
generate_data.py can be used to generate the nested sampling runs used in the paper except those using data from the Planck survey - see its documentation for more details. For details of the likelihood used for the Planck survey and how it can be downloaded see the paper and references therein.
All the plots in the paper can be generated using the diagnostic_paper_code.ipynb and planck_results_plotting.ipynb notebooks; see the notes within each for more details.
If you use this code in your academic research, please cite the nestcheck references. These are listed in the attribution section of the documentation at https://nestcheck.readthedocs.io/en/latest/.