Short Term Roadmap

(Written 2020-03-27)

Things to Achieve

Use Cases

• Glue solar has a beta feature for creating derived datasets from pixel selections. This is more efficient than the normal subset extraction. This feature needs a few tests / improvements:

  • Ensure that subsets extracted with this tool have correct WCS information and are linked to the other datasets. For example highlighting subsets shows up in the extracted dataset etc.
  • Find a good icon for the pixel extraction tool!
  • Other UX improvements etc

• Extract a 1D profile from a pixel selection of a >3D dataset, then be able to animate that profile along any extra dimensions. For example take a (space, space, wavelength, time) cube, select a pixel in (space, space) leading to a 2D derived dataset with the dimensions of (wavelength, time). Currently, Glue will collapse the time dimension, when it would be good for it to be able to provide a slider so you can see the wavelength profile with time.

• Test dataset autolinking for subsets of ND cubes. Currently, if you load in two ND cubes, like (space, space, time) and (space, space, wavelength) there is special handling in the autolinking code to link the two spatial axes. However, if you load in a (space, space, time) and a (time, wavelength) dataset the time axes will not be autolinked.

• Extract derived datasets from a path, this would allow drawing a line on an image and extracting a derived dataset for the pixels under that line. (Tom R apparently has a prototype of this.)

• Dask support. Especially, not loading all the array into memory for calculating pixel limits etc.

• Image / Movie export, both with axes and without axes via matplotlib.

• Solar physics loaders, specifically, NDCube / NDData, SST, IRIS, EIS, DKIST.

• Auto selection of the "correct" solar physics colourmap.