A general sampling tool for the exploration of inflationary observables.
This code is designed to derive general predictions for inflationary cosmology with random sampling techniques. The goal is not to optimize a model's compatibility with observation, but to make predictions from a robust exploration of initial conditions and parameter spaces. A recent application of this software can be found in this paper.
The code serves as an extension to PyTransport (see accompanying paper) which was developed by David Mulryne and John Ronayne. Some parts of the current technology were developed with David Seery.
Contents
- Updated python 3 version of PyTransport, with internal routines for tensor algebra.
- A priori and Latin hypercube sampling methods.
- MPI ready: run sampling jobs locally or on computing clusters.
To install the code as a package run the following command from the root directory:
python -m pip install -e .The setup should have installed an example model file dquad_2sphere.py. To test whether the submodule was installed,
you can run some unittests via
pytest tests/test_install.pyDemo files for single field & multiple field models can be found in the /notebooks directory. These files explain how to use the code: Installation, Implementing compiled modules, and sampling over models.