#2877 Bug sequence showing where FFT of a component is mishandled

[Jeremy Faden]

Dave and I were able to recreate a bug he ran into, trying to do the FFT of a component:

1. plot vap+das2server:https://space.physics.uiowa.edu/das/stream?dataset=Tracers/TRACERS-2/QL/MSC/BAC&intrinsic=true&start_time=2025-08-31T18:25:00.000Z&end_time=2025-08-31T18:45:00.000Z
2. delete By and Bz components
3. click on Bx to set focus
4. change operations to "|unbundle('Bx')|fftPower(512,2,'Hanning')"
5. Note the "child" plotElement is deleted and the "parent" is now visible, with no filters, using spectrogram. This is the bug!
6. Enter "|unbundle('Bx')|fftPower(512,2,'Hanning')" in the parent operations. This is what was intended.

[Jeremy Faden]

Here's a script which demos:

plot('vap+inline:ripplesVectorTimeSeries(101024)')
dom.controller.deletePlotElement(dom.plotElements[-1])
dom.controller.deletePlotElement(dom.plotElements[-1])
dom.plotElements[1].component='''|unbundle('X')|fftPower(512,2,'Hanning')'''

[Jeremy Faden]

The gist of the problem is that there is code which says the parent plotElement should get the same render type as its child. So the child is set to spectrogram, then the parent is set, and spectrograms should not have children.

I've changed the code so that only the series/hugeScatter/scatter changes are propagated up.