opened 12:57AM - 24 Apr 21 UTC
HoloMaps and DynamicMaps don't respect the `framewise` option if they are operat…ing on composite plot types like `Layout`, `NdOverlay`, or `Overlay`.
Consider this MRE:
```
import numpy as np
import holoviews as hv
hv.extension('bokeh')
def dmap_func(amp, freq, phase):
ts = np.linspace(0,5, num=50)
return hv.Curve(zip(ts, amp * np.sin(freq * ts + phase)))
def dmap_overlay_func(amp, freq, phase):
return dmap_func(amp, freq, phase) * dmap_func(amp, freq+0.5, phase)
dmap = hv.DynamicMap(dmap_func, kdims=[hv.Dimension('amp', range=(1,10)),
hv.Dimension('freq', range=(1,3)),
hv.Dimension('phase', range=(0,7))])
dmap_overlay = hv.DynamicMap(dmap_overlay_func, kdims=[hv.Dimension('amp', range=(1,10)),
hv.Dimension('freq', range=(1,3)),
hv.Dimension('phase', range=(0,7))])
```
Here's `dmap.opts(framewise=True)`
And here's `dmap_overlay.opts(framewise=True)`
As you can see, the frame has not been scaled as I expected, despite `framewise=True`. I've been able to reproduce this with `Overlay`, `NdOverlay`, and `Layout`.
I've also been able to replicate this with `HoloMap`. However, as opposed to `DynamicMap` which will set the scale based on the first frame, `HoloMap` will set the scale based on all the data. I think this shows that both objects ignore `framewise` and revert to their default behavior instead.
#### ALL software version info
`holoviews 1.14.3`
`bokeh 2.3.0`
Chrome & Windows