Custom HoverTool overridden with "by="

hvPlot
1

I’m new to HoloViz and just started using the bokeh HoverTool to customise how the hover tool looks. I noticed that when I use “by=‘category_column’” rather than “color=‘category_column’” the custom hover tool formatting is overridden.

I produced a simple example with scatter plots below to illustrate the issue. Obviously for a scatter plot I could just use “color=‘category_column’” to get around this problem but for line plots you can only use “by=‘category_column’” as far as I know.

import hvplot.pandas
import holoviews as hv
from bokeh.sampledata.autompg import autompg_clean as df
from bokeh.models import HoverTool

hv.extension(“bokeh”)

hover = HoverTool(
tooltips=[(‘miles per gallon’,‘@mpg’),
(‘horse power’,‘@hp’),
(‘region’, ‘@origin’),
(‘manufacturer’, ‘@mfr’)
]
)

df.hvplot(
x=‘mpg’,
y=‘hp’,
kind=‘scatter’,
color=‘origin’, # HoverTool works fine
title=‘color=origin’,
# by=‘origin’, # HoverTool doesn’t work
# title=‘by=origin’,
hover_cols=[‘origin’, ‘mfr’],
).opts(
tools=[hover],
)

image
image

2

Hi @Melanie

Welcome to the community.

My understanding is that when you use by you are really creating an overlay of plots. And you are trying to apply the hover tool to that layout. I would also like/ expect that to work. But since it does not I thought that maybe setting opts of the underlying Curve would work. And it seems to work.

import hvplot.pandas
import holoviews as hv
from bokeh.sampledata.autompg import autompg_clean as df
from bokeh.models import HoverTool

hv.extension("bokeh")

hover = HoverTool(
tooltips=[
    ('miles per gallon','@mpg'),
    ('horse power','@hp'),
    ('region', '@origin'),
    ('manufacturer', '@mfr')
])

print(df)

plot = df.hvplot(
    x='mpg',
    y='hp',
    kind='line',
    by=['origin', 'mfr'],
    title='by=origin',
).opts(hv.opts.Curve(tools=[hover]))

import panel as pn
pn.extension()
pn.panel(plot).servable()

hover-curve
hover-curve712×300 106 KB

3

Hi @Marc thanks so much, that’s sorted my problem for me.