PyFHD

Python Fast Holographic Deconvolution

TODO: Add Testing Results and Testing coverage

FHD

FHD is an open-source imaging algorithm for radio interferometers, specifically tested on MWA Phase I, MWA Phase II, PAPER, and HERA. There are three main use-cases for FHD: efficient image deconvolution for general radio astronomy, fast-mode Epoch of Reionization analysis, and simulation.

PyFHD is the translated library of FHD from IDL to Python, it aims to get close to the same results as the original FHD project. Do expect some minor differences compared to the original FHD project due to the many differences between IDL and Python. These differences are often due to the difference in precision between IDL and Python with IDL being single-precision (accurate upto 1e-8) and Python being double-precision (1e-16). Some of the IDL functions are double-precision but most default to single-precision.

Quick Start

Clone (and move into) the PyFHD repo:

$ git clone https://github.com/ADACS-Australia/PyFHD && cd PyFHD

Then just pip install the repo:

$ pip install -r requirements.txt .

For full installation notes, including dependencies on FHD, check out the ReadTheDocs installation page.

Note: Once PyFHD is feature-complete, we aim to make this both a conda install and a pip install.

Useful Documentation Resources

PyFHD documentation
MWA ASVO - service to obtain MWA data
FHD repository - the original IDL code
FHD examples - examples on how to use the original IDL code
FHD pipeline scripts - pipeline scripts using the original IDL code

Community Guidelines

We are an open-source community that interacts and discusses issues via GitHub. We encourage collaborative development. New users are encouraged to submit issues and pull requests and to create branches for new development and exploration. Comments and suggestions are welcome.

Please cite Sullivan et al 2012 and Barry et al 2019a when publishing data reduction from FHD.

Maintainers

FHD was built by Ian Sullivan and the University of Washington radio astronomy team. Maintainance is a group effort split across University of Washington and Brown University, with contributions from University of Melbourne and Arizona State University.

PyFHD is currently being created by Nichole Barry and Astronomy Data and Computing Services (ADACS) members Joel Dunstan, Paul Hancock, and Jack Line. ADACS is a collaboration between the University of Swinburne and Curtin Institute for Data Science (CIDS) located in Curtin University.