Visualization for TMT IRIS

Real-time and offline visualization of IRIS data, both data from the imagers and from the spectrometer.

2D visualization

Test data

• Raw science frame
• Reduced data

Launch 2D visualization

Currently we rely on the JWST project imexam for visualizing 2D datasets, imexam has a programmatic interface to post-process data and interact with a GUI backend which can either be DS9 or Ginga (Browser-based).

See the launch_2d_imexam_ds9.py as an example, it could be automatically launched by the pipeline execution script tmtrun so that once a frame is reduced, it is automatically displayed inside DS9. Even after the image is in DS9, several algorithms can be executed from the command line to further process the image, see the available algoritms.

After the image appears on DS9, in the Python terminal type:

viewer.imexam()

to load a menu of all different algorithms that can be executed.

See an example below of a IRIS reduced frame loaded in DS9 with the imexam menu open and a line plot.

3D visualization

Jdaviz is the new JWST python package which includes Cubeviz for visualizing data cubes from a spectrograph, SpecVis for 1D spectra and quick-look analysis and MOSViz for multi-object spectrographs.

Install jdaviz

Create a dedicated conda environment by prefix or name with the requirements that need to be compiled:

conda create --prefix=~/jdaviz -c glueviz -c conda-forge python==3.8 astropy glueviz voila glue-jupyter asteval
conda create --name=jdaviz -c glueviz -c conda-forge python==3.8 astropy glueviz voila glue-jupyter asteval

Then install jdaviz with pip:

pip install jdaviz

Jdaviz on Galactica

$ conda activate /data/software/repo/anaconda3/envs/jdaviz

After this you should have jdaviz in your path:

$ which jdaviz
/data/software/repo/anaconda3/envs/jdaviz/bin/jdaviz

Open an issue in this repository if the environment is broken. In particular, if colormaps are broken, paste any warning you see in the terminal to #3.

Test data

On Figshare, I posted a simulated observation with the IRIS spectrograph (created by Nils-Erik Bjorn Rundquist).

Launch 3D visualization

We can experiment with Jdaviz launching:

jdaviz iris_cube_model.fits --layout=cubeviz

The Jdaviz allows to:

• go interactively thorugh all the slices
• select a subset of the data in pixel space or frequency bin
• apply a set of algorithms from the builtin library

See for example this screenshot (datasets had different shapes, the rest of the data is zeroed):

Customize jdaviz

Each layout of jdaviz is given by a configuration in:

jdaviz/configs/cubeviz

and a set of plugins.

Plugins should be able to modify the application through events.