This is a development version of a potential data library designed for the Event Horizon Telescope (EHT) and next generation EHT (ngEHT).
The aim of this library is
- Natively handling five dimensional images (two dimensional images over frequency, polaization and time grids)
- Providing more readable and intuitive data access to uv-domain datasets (etc, visibilities, closure quantities, calibration tables)
- Offering the data I/O not only with the conventional data formats for the EHT (FITS for images, UVFITS for visibilities), but also with other public data format for the N-D labeled multi-dimensional data used in the science community (e.g. NETCDF4 (based on HDF5), Zarr) more easily accessible from many other launguages (etc., Julia, C/C++, Fortran, etc).
- Offer lazy loading of data to handle a potentially large data set on the disk that cannot be fully loaded into the memory.
- Sounce independent data editing / flagging of images and visibilities.
To achive the above aims, this library provides data structure based on Xarray's Dataset for both uv- and image-domain data. This allows to build up a readable data structure easily accessible even without this particular library.
You can install this library by
python setup.py install
or
pip install .
Documentation is in preparation.