Radial HeatMap option doesn't work with Dataset.to

[toddrjen]

Radial HeatMaps work when declaring a HeatMap directly, but don't seem to work when using Dataset.to. Consider the following Dataset:

import numpy as np
import pandas as pd
import holoviews as hv
hv.extension('bokeh')

x = np.arange(0, 1000)
y = np.sin(x/1000*2*np.pi)
index = x//10
rep = x%10
thirds = x%3==0

df = pd.DataFrame({"Value": y, "Index": index, "Repetition": rep, "Thirds": thirds})
ds = hv.Dataset(df, ['Repetition', 'Index', 'Thirds'], ['Value'])

If we select one value of Thirds and make a radial HeatMap with that, it works properly:

%opts HeatMap [radial=True]
hv.HeatMap(ds.select(Thirds=True), ['Index', 'Repetition'])

However, if we use Dataset.to on the same data to get a dropdown, the HeatMap isn't radial:

%opts HeatMap [radial=True]
ds.to(hv.HeatMap, ['Index', 'Repetition'])

I would be willing to use an ugly workaround to get this working if it is available.

[toddrjen]

It looks like overlays don't work, either.

Using the same Dataset as above, ds, this works:

%opts HeatMap [radial=True]
hm1 = hv.HeatMap(ds.select(Thirds=True), ['Index', 'Repetition'])
hm1

And this works

%opts HeatMap [radial=True]
hm2 = hv.HeatMap(ds.select(Thirds=False), ['Index', 'Repetition'])
hm2

And this works:

%opts HeatMap [radial=True]
hm1 + hm2

But this doesn't:

%opts HeatMap [radial=True]
hm1*hm2

[toddrjen]

There is also no colorbar with radial heatmaps:

%%opts HeatMap [radial=True colorbar=True]
hv.HeatMap(ds.select(Thirds=True), ['Index', 'Repetition'])

Doesn't show any legend.

[philippjfr]

Thanks, hopefully we can get a fix for this into 1.10.2.