Changing the format of nested xaxis ticks labels for the boxplot chart

[dinya]

I want to change the format of the nested xaxis ticks labels for boxplot chart (major, group and subgroup in terms of bokeh.models.CategoricalAxis attributes).

The following code is based on Boxplot Chart — HoloViews v1.18.1

import holoviews as hv
from holoviews import dim
hv.extension('bokeh')

from bokeh.sampledata.autompg import autompg as df
title = "MPG by Cylinders and Data Source, Colored by Cylinders"
boxwhisker = hv.BoxWhisker(df, ['cyl', 'origin'], 'mpg', label=title)
boxwhisker.opts(show_legend=False, width=600, box_fill_color=dim('origin').str(), cmap='Set1')

I want to change the format of the nested group-axis ticks labels (cyl-level).

If I change figure.xaxis.formatter the major axis (origin-level) format is changed:

from bokeh.plotting import show
from bokeh.models import NumeralTickFormatter
figure = hv.render(boxwhisker)
figure.xaxis.formatter = NumeralTickFormatter(format="000")
show(figure)

But I can’t find the attributes (or any other way) to change group-axis ticks labels format (cyl-level). Smth like figure.x axis.major_label_orientation for major-axis ticks label orientation and figure.x axis.group_label_orientation for group-axis ticks label orientation.

Maybe this problem requires a very specific (custom) solution, for example pass x_range parameter while initializing bokeh.plotting.figure(x_range=bokeh.models.FactorRange(...)) (seen on Bar charts — Bokeh 3.3.2 Documentation ).
Also I read that categorical ranges is sequences of strings (here). So I have to convert my dataframe columns for xaxis as strings in the format I need.

Thanks in advance for any advice.

[ahuang11]

I believe you may need to use hooks: FAQ — HoloViews v1.18.1

import holoviews as hv
from bokeh.models.tickers import FixedTicker
hv.extension('bokeh')

def update_axis(plot, element):
    b = plot.state
    b.axis[0].ticker = FixedTicker(ticks=list(range(0, 10)))

h = hv.Curve([1,2,7], 'x_col', 'y_col')
h = h.opts(hooks=[update_axis])
h

[dinya]

According to Customizing Plots — HoloViews v1.18.1

Categorical axes

<…> If the data is a string type or other object type it is formatted as a string and each unique category is assigned a tick along the axis. <…>

Is there any way to hook string formatting of the object during data preparation for the axis representation?

[dinya]

I can “patch” the source dataframe or a copy of it, but this solution is a bit rough:

import holoviews as hv
from holoviews import dim
hv.extension('bokeh')

from bokeh.sampledata.autompg import autompg as df

# "Patch"
df["cyl"] = df["cyl"].apply(lambda x: f"Patched {x}")

title = "MPG by Cylinders and Data Source, Colored by Cylinders"
boxwhisker = hv.BoxWhisker(df, ['cyl', 'origin'], 'mpg', label=title)
boxwhisker.opts(show_legend=False, width=600, box_fill_color=dim('origin').str(), cmap='Set1')