hydra-pspec is a Gibbs sampler-based power spectrum estimation code with foreground filtering and in-painting capabilities. For more details on the underlying mathematics and demonstrations on simulated data please see Kennedy+2023 and/or Burba+2024.
hydra-pspec is written in Python and has the following dependencies:
- astropy
- jsonargparse
- matplotlib
- mpi4py
- multiprocess
- numpy
- pyuvdata
- scipy
- setuptools
- setuptools_scm
If you wish to install all of these dependencies with conda/mamba, you can do so using the included environment.yaml file via
conda env create -f environment.yaml
hydra-pspec can then be installed via
pip install .
hydra-pspec can be run using the provided driver script run-hydra-pspec.py. This code is designed to be run using MPI via e.g.
mpirun -n <number_of_ranks> run-hydra-pspec.py --config <config_file.yaml>
There are several input parameters which are required to run a hydra-pspec analysis. Using jsonargparse, these input parameters can be specified via a configuration yaml file or directly via the command line. For a full list of available input parameters, run
python run-hydra-pspec.py --help
Please see test_data/config.yaml for an example of a configuration yaml file containing the minimum required parameters. Please also see the jsonargparse documentation for more details.
Users of the code are requested to cite the following papers:
hydra-pspec is an open source project which is being actively developed. If you would like to make a contribution or suggest a feature, you are very welcome to do so in the form of an issue and/or pull request.