framewise=False in a function works, framewise=False in a method doesn't work

HoloViews
1

Hello everybody,
My goal was to setup an interactive plot, in which the users themselves can decide to have the axis auto-adjust to the displayed data. And I thought that the option framewise = True (or False) would be the solution to achieve that.
And it looks like framewise = True is working, whereas framewise = False in some cases is not.

In the first code example the interactive plot is behaving like I would expect, when setting framewise = False.

import holoviews as hv
import pandas as pd
import panel as pn
import param

hv.extension("bokeh")

df = pd.DataFrame({'X': [0, 1, 2, 3, 4, 5],
                   'Var_A': [2, 1, 3, 6, 4, 2],
                   'Var_B': [8, 7, 5, 8, 5, 7]})
df_rolling = df.rolling(window=2, center=True).mean()

def update_plot(var):
    plot = hv.Curve(df, ('X', 'Index'), var).opts(color='blue', tools=["hover"], framewise=False)
    plot_rolling = hv.Curve(df_rolling, ('X', 'Index'), var).opts(color='red', tools=["hover"], framewise=False)
    return plot * plot_rolling

def view_plot():
    return pn.panel(hv.DynamicMap(pn.bind(update_plot, var=widget.param.choose_var)))

class Example_Widget(param.Parameterized):
    choose_var = param.Selector(label='Choose Var:', default='Var_B', objects=['Var_A', 'Var_B'])

widget = Example_Widget()
# I read something about framewise not working in pn.Tabs so that's why I used a couple of Tabs here. But all o.k.
pn.Tabs(
    ('Tab A', pn.Tabs(
        ('Tab 1', pn.Row(widget.param, view_plot)),
        ('Tab 2', 'Hello'))),
    ('Tab B', 'Hello'),
)

The second code example is a bit different to the first: I have moved the functions inside the class (but that’s all):

import holoviews as hv
import pandas as pd
import panel as pn
import param

hv.extension("bokeh")

df = pd.DataFrame({'X': [0, 1, 2, 3, 4, 5],
                   'Var_A': [2, 1, 3, 6, 4, 2],
                   'Var_B': [8, 7, 5, 8, 5, 7]})
df_rolling = df.rolling(window=2, center=True).mean()

class Example_Widget(param.Parameterized):
    choose_var = param.Selector(label='Choose Var:', default='Var_B', objects=['Var_A', 'Var_B'])
    
    def update_plot(self, var):
        plot = hv.Curve(df, ('X', 'Index'), var).opts(color='blue', tools=["hover"], framewise=False)
        plot_rolling = hv.Curve(df_rolling, ('X', 'Index'), var).opts(color='red', tools=["hover"], framewise=False)
        return plot * plot_rolling
    
    def view_plot(self):
        return pn.panel(hv.DynamicMap(pn.bind(self.update_plot, var=self.param.choose_var))) 

widget = Example_Widget()
pn.Tabs(
    ('Tab A', pn.Tabs(
        ('Tab 1', pn.Row(widget.param, widget.view_plot)),
        ('Tab 2', 'Hello'))),
    ('Tab B', 'Hello'),
)

In this (the second) code example the plot is behaving like I would expect it to behave if I had set framewise = True. But it’s set to False. So I don’t understand why the plot auto-updates its axis-ranges. If someone could tell me what my mistake is or how I could stop the interactive plot to auto-adjust its axis-ranges, I would really appreciate that. (And if something is unclear, please let me know)
Best regards
(I ran the example code in a Jupyter-Notebook version 6.4.8, holoviews version 1.14.8, panel version 0.12.6, param version 1.12.0)

2

Try calling view_plot function like this:

...
pn.Tabs(
    ('Tab A', pn.Tabs(
        ('Tab 1', pn.Row(widget.param, widget.view_plot())),
        ('Tab 2', 'Hello'))),
    ('Tab B', 'Hello'),
)
3

That’s it. That’s the solution. Thanks a lot!
It always helps to call a method the way a method should be called :blush: