Take a set of fits images (which share a World Coordinate System, but span different timesteps, spectral channels, or polarisations) and write the results into an HDF5 file which is disk-space-efficient, stores all relevant metadata, and is optimised for efficient extraction of timeseries (lightcurve) data.
This software has been written for IPS studies with the MWA. It is currently limited in its ability to handle multiple spectral channels and expects the filenames to be in the format produced by WSClean. Contact me if you wish to use this software for other purposes.
The script make_imstack.py can be used to read the fits files into the hdf5 file. Run the script with the -h flag for instructions. This is by far the least efficient part of the process. Note that this script relies on h5py_cache which is available via pip.
The file image_stack.py contains a class for accessing the image stack you have generated. See the documentation for the various member functions. For example, extracting a time series for a given decimal RA and Dec is as simple as:
from image_stack import ImageStack
imstack = ImageStack("my_hdf5_file.hdf5")
imstack.world2ts(112.334341, +5.32321)
The script make_moment.py gives an example of extracting a summary image (or images) of each pixel. In this case, generating mean, standard deviation and higher moments of the time series with high- and low-pass filters optionally applied. mpi4py is used for parallelisation. Run the script with the -h flag for instructions.
Please cite Morgan et al. (2018) if you make use of this software for research purposes. Further details on imstack are available in the appendix of this paper.
h5pynumpyastropyscipy(moment_image)mpi4py(moment_image-- tested with 3.0.3)
Note that installation of mpi4py is not done automatically via setup.py as it is only required for moment_image.py. Installation of mpi4py via e.g. pip, requires mpicc to be available, not just mpirun.