How do I get the HoloViews box selection tool to synchronize multiple connected plots?

Question

For example, with

import numpy as np
import pandas as pd
import holoviews as hv

from holoviews import opts 

hv.extension('bokeh')

def scatter_hist(src, x, y):
    p = hv.Scatter(src, x, y).hist(num_bins=20, dimension=[x, y]).opts(
            opts.Scatter(show_title=False, tools=['hover','box_select']), 
            opts.Histogram(tools=['hover','box_select']),
            opts.Layout(shared_axes=True, shared_datasource=True, merge_tools=True)
        )
    return p

scatter_hist(ds, 'x', 'y')

I can get box_zoom to work as expected (zooming in any one of the three figures imposes the corresponding zoom on the other two) but box_select does not behave in this way. I can only box-select in the top histogram (though the toolbar doesn't highlight the tool), and this has no effect on the other plots.

I've experimented with different values for tools (including omission) for both opts.Scatter and opts.Histogram but can't discern any pattern with what's going in when I do (in fact I'm a bit vague on what I need to specify for tools since some appear without being listed).

How do I get the HoloViews box selection tool to synchronize multiple connected plots, so that selection in one, selects the corresponding values in the others?

I've read through the documentation on "linked brushing" but can't get that to apply to the provided MWE even though that documentation states that linked brushing should (as expected) be "automatic", and "enabled for Bokeh plots sharing a common data source by simply adding a selection tool"

How do I get the HoloViews box selection tool to synchronize multiple connected plots?

0 Answers