cadence breaks and clipping data off screen

[jbfaden]

I have a time series which shows two issues with the rendering. First, the cadence is dependendent on what's shown. I thought this was fixed but this data seems to tweak a bug. Second, if an outlier is off screen, connectors aren't drawn, and there is no way to know that the point is there.

[jbfaden]

Here is an animation showing where data is off screen.

[jbfaden]

Here is showing the data breaks, implying that the cadence is changing with axis position.

[jbfaden]

This Autoplot script demos:

plot( 'file:///home/jbf/data/gardenhouse/data2/$Y/$m/gardenhouse.core0.$Y$m$d.csv?timerange=2021-01-01/2021-07-01', renderType='series' )

dom.plots[0].xaxis.range= datumRange( '2021-06-21 18:45 to 21:05' )
writeToPng('/tmp/0001.png')

dom.plots[0].yaxis.range= datumRange( '0 to 2.5' )
writeToPng('/tmp/0002.png')

dom.plots[0].yaxis.range= datumRange( '0.0 to 3.3' )
dom.plots[0].xaxis.range= datumRange( '2021-06-21 18:27 to 21:22' )
writeToPng('/tmp/0003.png')

[jbfaden]

I think the cadence needs to be revisited. It is recalculated each time the data is plotted, causing this erratic behavior and a performance penalty.

Here are tricky datasets (as Autoplot URIs) for cadence:

• https://data.ngdc.noaa.gov/platforms/solar-space-observing-satellites/goes/goes17/l2/data/mpsh-l2-avg5m/2021/06/sci_mpsh-l2-avg5m_g17_d20210602_v1-0-3.nc?AvgIntElectronFlux[:,0] (Actually I think there's fill in there.)

[jbfaden]

Sadie provided a nice data set which reminds that orbits often aren't broken as would one would expect:
https://github.com/autoplot/dev/blob/master/demos/2021/20210729/demoBreakOrbit.jy

[jbfaden]

In Sadie's case above, the timetags are within the data (DEPEND_0 of DEPEND_0). I should experiment with detecting breaks in these when timetags are available.

[jbfaden]

See also https://sourceforge.net/p/autoplot/bugs/1582/

[jbfaden]

See also #38

[jbfaden]

See also #111