Implement histogram-style (steps, not real histogram) plots for specviz

[eteq]

This was discussed in several places in #133 but I'm now realizing we never broke out an issue to specifically do this.

The goal here is to implement "histogram-style" plots instead of "line plots" - i.e., have the points on the spectrum be plotted as the center of a "box"-like plot instead of joining the segments directly as a line. For optical spectra this is typically more physical because they are closer to the concept of "flux in boxes" vs "samples from a continuous signal".

Not everyone agrees on what I just said there, so for this issue that should be an option. After this is implemented we can discuss whether it should be the default (which is what I advocated for in #133, but others were less convinced.)

🐱

[nancymac519]

The "histogram-style" plot mode is essential; please implement. It's the most realistic way to view any data that are binned into pixels. i agree with eteq that it should be the default; I used to get really tired of nagging students to use it with IRAF.

R calls this plot style "step plot" or "stair step" plot, a much better term than "histogram," which properly refers to a completely different type of graph.

I'm not yet a SpecViz user, and will not be a consistent user until this is implemented, but i would like to be. Thanks!

[karlglazebrook]

Found this when I was seeing if anyone else had requested this.

Yes this is the superior way to plot spectra, as they contain pixels.

In matplotlib the command would be plt.step(wav, flux, where='mid') or plt.plot(..., drawstyle='steps-mid')

[nancymac519]

Thanks! Yes, I also requested this. R has a plot type (in the 'plot' package) "s" for stair steps.

…

________________________________________ From: Karl Glazebrook ***@***.***> Sent: Tuesday, November 9, 2021 2:37 AM To: spacetelescope/jdaviz Cc: Morrison, Nancy D.; Comment Subject: [EXTERNAL] Re: [spacetelescope/jdaviz] Implement histogram-style plots for specviz (#385) Found this when I was seeing if anyone else had requested this. Yes this is the superior way to plot spectra, as they contain pixels. In matplotlib the command would be plt.step(wav, flux, where='mid') or plt.plot(..., linestyle='steps-mid') — You are receiving this because you commented. Reply to this email directly, view it on GitHub<https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.heygears.com%2Fspacetelescope%2Fjdaviz%2Fissues%2F385%23issuecomment-963886870&data=04%7C01%7Cnancy.morrison%40utoledo.edu%7C6aec760f27bf46a78c0408d9a353b7cb%7C1d6b1707baa94a3da8f8deabfb3d467b%7C0%7C0%7C637720402249164606%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=ePYTpCgbl9pEv4QVflmHq2GyWJ6BbjXFwqlNqr5sGiQ%3D&reserved=0>, or unsubscribe<https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.heygears.com%2Fnotifications%2Funsubscribe-auth%2FASF32JDTTCHCNBMCMIAYYILULDFRZANCNFSM4WBBSYMA&data=04%7C01%7Cnancy.morrison%40utoledo.edu%7C6aec760f27bf46a78c0408d9a353b7cb%7C1d6b1707baa94a3da8f8deabfb3d467b%7C0%7C0%7C637720402249174601%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=z3f%2FMQ4g1DTKRCYFwMgrC%2FQ5DYr5qBrwQPpjyQDj%2F2E%3D&reserved=0>. Triage notifications on the go with GitHub Mobile for iOS<https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fapps.apple.com%2Fapp%2Fapple-store%2Fid1477376905%3Fct%3Dnotification-email%26mt%3D8%26pt%3D524675&data=04%7C01%7Cnancy.morrison%40utoledo.edu%7C6aec760f27bf46a78c0408d9a353b7cb%7C1d6b1707baa94a3da8f8deabfb3d467b%7C0%7C0%7C637720402249174601%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=G2uz1gVYZEhBmfpLTHGcTEcPSwfNStwt61lkwSGyL4c%3D&reserved=0> or Android<https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fplay.google.com%2Fstore%2Fapps%2Fdetails%3Fid%3Dcom.github.android%26referrer%3Dutm_campaign%253Dnotification-email%2526utm_medium%253Demail%2526utm_source%253Dgithub&data=04%7C01%7Cnancy.morrison%40utoledo.edu%7C6aec760f27bf46a78c0408d9a353b7cb%7C1d6b1707baa94a3da8f8deabfb3d467b%7C0%7C0%7C637720402249184596%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=JSztzaHOATsihTmEb8%2FycAaZg0Z%2Ft1SF1Vu5YnUNy8k%3D&reserved=0>.

[pllim]

Solution from @astrofrog : Change profile viewer over at glue-jupyter to pass in marks for edges of the steps instead of middle.

https://github.com/glue-viz/glue-jupyter/blob/d6c9dd72714d5f3d872289bdcac5a9acc2161232/glue_jupyter/bqplot/profile/layer_artist.py#L71

[PatrickOgle]

I disagree that this should be default. For high SNR spectra with thousands of wavelength bins, it adds clutter.
It should definitely be an option.

[kecnry]

It will be an exposed option in the layer options with glue-viz/glue-jupyter#309 (glue work already merged in glue-viz/glue#2292), but will default to using steps (as this is how the profile viewer already worked in matplotlib within glue). If we want to change the default in jdaviz to not using steps, we will need extra logic on our end whenever a new layer is added to override glue's default or we can suggest the change upstream.