Expose time functionality

[Carifio24]

While working on WWT interactives recently that were examining motion over time, I realized this sort of visualization/data exploration is completely impossible in glue-wwt, since the time functionality of the WWT engine isn't exposed at all. This PR looks to remedy that issue by exposing WWT's time functionality in both the Qt and Jupyter viewers.

This PR contains updates to both the viewer and to table layers (I don't know that there's any sort of natural use case for image layers, but if there is one I'd be interested in learning what it is). On the viewer side, we allow setting the following WWT time settings:

• Whether or not to play time
• The time rate (the rate at which WWT plays time relative to real time)
• Setting the current time in WWT (using everyone's favorite widget, a time slider)
• Setting min/max values for the WWT time (which set the slider bounds)

Internally, there are callbacks which maintain the relative order min <= current <= max for the time values. If one wants to run time backwards, the clock rate can be set to be negative. It felt too convoluted to me to allow having both inverted time bounds and a negative clock rate.

On the table layer side, one can toggle whether or not each layer is a time series layer. If this is set to true, we allow users to set:

• A time attribute for the layer that tells glue which component represents time. This component must be a DateTimeComponent.
• The decay rate, as an astropy Quantity. This determines how quickly points fade out of view.

Note that this relies on updates to glue (see glue-viz/glue#2475) and echo (glue-viz/echo#37), to allow serialization of astropy Units (for the decay) and to autoconnect Qt datetime widgets (for the min/max time selection), respectively.

For testing purposes, here's an example session and notebook where I've added a (totally meaningless) time component to the W5 data set. Giving a time rate on the order of 1e6 gives a good rate for the motion.

TimeExamples.zip

[codecov]

Codecov Report

Attention: Patch coverage is 40.53030% with 157 lines in your changes missing coverage. Please review.

Project coverage is 67.36%. Comparing base (f6f0135) to head (8cc2ff6).
Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
glue_wwt/viewer/jupyter_viewer.py	0.00%	84 Missing ⚠️
glue_wwt/viewer/time_dialog.py	0.00%	23 Missing ⚠️
glue_wwt/viewer/jupyter_utils.py	0.00%	20 Missing ⚠️
glue_wwt/viewer/viewer_state.py	61.53%	10 Missing ⚠️
glue_wwt/viewer/options_widget.py	82.35%	9 Missing ⚠️
glue_wwt/viewer/table_layer.py	60.86%	9 Missing ⚠️
glue_wwt/viewer/data_viewer.py	95.23%	1 Missing ⚠️
glue_wwt/viewer/table_style_editor.py	87.50%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #103      +/-   ##
==========================================
- Coverage   71.21%   67.36%   -3.86%     
==========================================
  Files          18       21       +3     
  Lines        1063     1250     +187     
==========================================
+ Hits          757      842      +85     
- Misses        306      408     +102     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[astrofrog]

@Carifio24 thanks for the PR! Can you rebase this to make sure the CI is passing before I review this? Note that setup.cfg doesn't exist anymore (there is a conflict) and anything that used to be in there is now in pyproject.toml.

[Carifio24]

@astrofrog I've rebased (and fixed some codestyle issues), but I expect the CI will fail as this depends on PRs to glue and echo.

[astrofrog]

@Carifio24 - I just tried this out in Qt glue, I did the following:

• Load up W5 catalog
• Create a WWT viewer and use Ra/Dec to show points
• Add a date column to the dataset by doing:

dates = np.arange(17771).astype(np.datetime64(1, 'D'))
dc[0].add_component(dates, label='date')

• Enable time series mode for the layer and ensure 'date' is selected for the attribute, and use decay of 10 days.
• Move time slider

Nothing changes in the view - am I doing something wrong above?

[astrofrog]

Also not sure if related to this PR or not, but in Qt glue, the default tab for the viewer options is 'Other' - but it should probably be the first tab?

[Carifio24]

@astrofrog What you were doing is correct, I think it's probably just a case of the points being too small and thus no change being apparent. I did what you described and changed the color and point size, and here's how it looks for me:

Screen.Recording.2024-09-05.at.9.04.07.AM.online-video-cutter.com.mp4

And the issue with the viewer starting from the wrong tab was from this PR, I just pushed a commit to fix that.

[astrofrog]

Ok thanks, very nice! Feel free to merge once you feel this is ready :)

[Carifio24]

Sounds good! Only thing I still want to look at is whether the CI failures are related to this PR or not.

[Carifio24]

Closed and reopened to re-run CI.