v1.5.0

Parallelise over baselines using MPI

Run hydra-pspec/run-hydra-pspec.py with MPI.
Test strong scaling on DiRAC (using commit 53474e0).
Further details given in README.md and scaling_tests_README.md

v1.3

Fix spurious correlations between model components at different times

A bug was found where child processes spawned by multiprocess.Pool shared the random state of the parent. This resulted in identical draws for the omegas (Gaussian, zero-mean, unit-variance random vectors) at different times which produced correlated EoR and FG model samples. This version update fixes #7.

v1.2

Time-dependent flags and noise covariance matrices

Allow for time-dependent flags with shape (Ntimes, Nfreqs) and noise covariance matrices (Ntimes, Nfreqs, Nfreqs). Also adds support for loading in a UVData compatible file with time- and frequency-dependent noise variance estimates per baseline and polarization (via the --noise_uvdata command line argument).

v1.1

Complex sampled signal covariance

The sampled signal covariance is no longer restricted to be purely real. This may have been a bug in the original version of the code. Tests with real and complex sampled signal covariance matrices display similar results.