/eftpipe

EFTofLSS analysis pipeline

Primary LanguagePythonMIT LicenseMIT

License: MIT

eftpipe

Warning

The code still needs to be cleaned up, expected to finish before mid-July.

Future developments will be moved to desilike, and all changes will be directly submitted to the upstream PyBird repo.

A Python package for doing the multi-tracer EFT analysis (arXiv: 2308.06206).

If you have any questions, please feel free to contact Ruiyang Zhao and Xiaoyong Mu. Any additional products are also available upon reasonable request.

Dependencies

  • python>=3.10
  • numpy, scipy and pandas
  • cobaya for MCMC sampling
  • typing_extensions for type hints, optional
  • camb or classy boltzamnn code, optional
  • numba for better performance, optional
  • matryoshka for emulator, optional, not well tested yet

Installation

Install all packages using Anaconda

The easiest way to set up the environment and reproduce analysis results in the paper would be using Anaconda

git clone git@github.com:zhaoruiyang98/eftpipe.git
cd eftpipe
conda env create -f environment.yml
# or `conda env create -f environment-dev.yml` for development
# or `conda env create -f environment-emu-dev.yml` for emulator

MPI support could be installed by running

conda install -c conda-forge "mpi4py>=3"
# or `conda install -c intel "mpi4py>=3"` if you are using the intel compiler
# or `pip install "mpi4py>=3" --upgrade --no-binary :all:` if you want to build from source

Install classy

git clone --depth 1 --branch v3.2.2 https://github.com/lesgourg/class_public
cd class_public/
# For mac user: you have to edit class_public/Makefile line21 by hand to set CC to the path to your gcc compiler (gcc points to clang on mac by default)
make

Minimum installation

If numpy, scipy, pandas and cobaya have already been installed, you could run the following code to install eftpipe locally

git clone https://github.com/zhaoruiyang98/eftpipe.git
cd eftpipe
pip install -e .

Acknowledgements

  • Pierre Zhang and Guido D'Amico for developing PyBird and making this wonderful code public. eftpipe heavily relies on PyBird. For the purpose of extension, we include it as a subpackage. Please have a look at README for detailed information and citation.
  • Cheng Zhao for providing EZmock catalogues.
  • Arnaud de Mattia for developing the pypower package, which is used for power spectrum estimation in this work.