jsROOT visualization of the experiment

[eduard322]

I've been using jsroot for the geometry visualization for a long time, because it is much lighter than running it via root TBrowser and OpenGL, works quite well with VSCode extension and there are plenty of examples on jsroot webpage with quite sophisticated geometry (see for example this). At some point, the geometry visualization started looking like that:

What is possibly happening is that there are some subsystems which make visible too many of their TVolume objects. Previously, only Decay Volume could not be visualized, now it is Muon Shield as well (by parts they can be visualized, but not in the overall layout of SHiP). I had something similar when I set to visibility to fibers in MTC. Would be really useful if we can find problematic volumes so that jsroot will become a main (or at least user friendly) visualizer.

[eduard322]

Sorry, cannot upload pictures for some reason :/

[olantwin]

You should be able to copy and paste pictures

[olantwin]

We should definitely move to jsROOT for the event display (will also allow us to support mac OS users again).

Could you bisect to figure out what commit introduced the issue for the muon shield?

[eduard322]

The issue is caused by the default limitation of the jsroot on the number of objects to visualize. It was caused by the splitcal geometry, that has enormous number of objects that are direct daughters of the splitcal objects, i.e. there is no intermediate container for strips to hide strips and leave only the top volume shape. Muon shield and other top components can be returned by making tiny details manually invisible (object->SetVisibility(False)). The same way decay volume can be visualized, I tried to do that by making invisible all the sbt cells, but it's still to much elements that are visualized by default in Tracking station for example. The idea would be to make the containers for all high-granular structures.

[olantwin]

Thanks for investigating this. Then the issue is how the strawtubes, splitcal and vessel are constructed?

@wcmartylee @kathalbrecht (@matclim)

[wcmartylee]

Thanks for investigating this. Then the issue is how the strawtubes, splitcal and vessel are constructed?

The strawtubes structure is like so: top > station > layer > straw. So it would suffice to set the straws invisible, I assume?

[eduard322]

Thanks for investigating this. Then the issue is how the strawtubes, splitcal and vessel are constructed?

@wcmartylee @kathalbrecht (@matclim)

it's quite tricky since most of the top-level objects of the high-granular systems are TGeoVolumeAssembly: they don't have their own shapes. SST is a good example how it could be done: one creates the top volume of the layer as a box filled with air and put straws there as daughters. Later this air box can be painted as straws and after setting it visible and straws invisible, one get a visible easy and lightweight shape of the complex object.