Add hover_tooltips, hover_mode, hover_formatters opts to easily modify hover

[ahuang11]

See discussion at #1816 (comment)

Screen.Recording.2024-04-09.at.3.29.35.PM.mov

resolves #1816

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 31.71806% with 155 lines in your changes are missing coverage. Please review.

Project coverage is 88.28%. Comparing base (bf4592e) to head (fb07cf8).
Report is 6 commits behind head on main.

Files	Patch %	Lines
holoviews/tests/ui/bokeh/test_hover.py	20.93%	102 Missing ⚠️
holoviews/plotting/bokeh/element.py	45.91%	53 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #6180      +/-   ##
==========================================
- Coverage   88.70%   88.28%   -0.42%     
==========================================
  Files         316      318       +2     
  Lines       66121    66416     +295     
==========================================
- Hits        58650    58636      -14     
- Misses       7471     7780     +309     

Flag	Coverage Δ
ui-tests	?

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[droumis]

Hey @ahuang11 , no problem for this time, but in the future would you mind pasting in your test code utilized in video demos?

[ahuang11]

Yeah I forgot to do so this time and I've already lost the code for that, but the edited notebook should have everything you need to try out all the functionality

[droumis]

Looks great with subcoordinate_y! Nice work. Not a big deal but the only thing I couldn't figure out yet is how to have the unit included in the tooltip when the unit is set in the hv.Dimension and a formatter is applied in hover_tooltips, like {0.2f} (see 'amplitude' vs 'time'). It's not a big deal to add the unit it manually though.

Code

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

n_channels = 4
n_seconds = 300
fs = 256
init_freq = .01  # Initial sine wave frequency in Hz
freq_inc = 2/n_channels  # Frequency increment
total_samples = n_seconds * fs
time = np.linspace(0, n_seconds, total_samples)
channels = [f'CH {i}' for i in range(n_channels)]

data = np.array([np.sin(2 * np.pi * (init_freq + i * freq_inc) * time)
                 for i in range(n_channels)])
# print(f'shape: {data.shape} (n_channels, samples) ')

time_dim = hv.Dimension('Time', unit='s')
amplitude_dim = hv.Dimension('Amplitude', unit='µV')

curves = []
for channel, channel_data in zip(channels, data):
    ds = Dataset((time, channel_data), [time_dim, amplitude_dim])
    curve = hv.Curve(ds, time_dim, amplitude_dim, group='EEG', label=f'{channel}')
    curve.opts(
        subcoordinate_y=True,
        subcoordinate_scale=.75,
        color="black",
        line_width=1,
        tools=['hover'],
        hover_tooltips=[
            ("Group", "$group"),
            ("Channel", "$label"),
            "Time",
            ("Amplitude", "@Amplitude{0.2f}")],
        # hover_mode = 'vline',
        )
    curves.append(curve)

curves_overlay = hv.Overlay(curves, kdims="Channel")
curves_overlay = curves_overlay.opts(
    ylabel="Channel",
    show_legend=False,
    aspect=3.5,
    responsive=True,
    shared_axes=False,
    xlim=(50,150),
    title='Multi-Channel Timeseries',
)

curves_overlay

and some fun with hover_mode = 'vline':

Screen.Recording.2024-04-11.at.1.19.59.PM.mov

[ahuang11]

Ah, right as you commented, I just finished all the unit tests, and I think I fixed it in the newest push. Let me also try...

[ahuang11]

Oh I guess it's not yet supported. I can add a if unit, then add it as ()

[ahuang11]

Okay it should work with your code now; tests also added!

[ahuang11]

and some fun with hover_mode = 'vline':

Would be a nice Bokeh feature to have to have like hover_align="left", "center", "right"

Interestingly, it does seem to be automatic on bokeh

[ahuang11]

Should be ready for re-review if needed.

The UI tests passes in my local machine, but I noticed it failed here. I added flaky markers.

[hoxbro]

For the flaky tests, try using wait_until.

[ahuang11]

Eh the test just refuses to pass...