Separating hvplot plot and groupby widgets

[rsignell]

Following the interactive time series example from @ahuang11, I have this code:

import panel as pn
import xarray as xr
import holoviews as hv
import hvplot.xarray

pn.extension()
ds = xr.tutorial.open_dataset('air_temperature')
ds = ds.assign_coords(lon=(((ds.lon + 180) % 360) - 180), lat=ds.lat*1.0)
image, select = ds.hvplot('lon', 'lat', geo=True, cmap='viridis', widgets={'time':pn.widgets.Select})
stream = hv.streams.Tap(source=image.object, x=-88, y=40)

def timeseries(x, y):
    return ds.sel(lon=x, lat=y, method='nearest').hvplot('time', grid=True)

timeseries_plot = pn.bind(timeseries, x=stream.param.x, y=stream.param.y)
pn.Column(pn.Row(select), pn.Row(image, timeseries_plot))

which produces:

I thought I would enhance this by putting a marker on the map indicating where the user clicked to extract the time series.

I was able to create a points object with hvplot:

import pandas as pd
import hvplot.pandas
df = pd.DataFrame({'lon':[-88.0], 'lat':[40], 'station':[0]})
station_plot = df.hvplot.points(x='lon', y='lat', geo=True, by='station', size=15)

which displays fine on it’s own, but when I try to overlay it on the image object above:

image * station_plot

I get back:

image * station_plot
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Cell In[5], line 1
----> 1 image * station_plot

File /home/conda/global/b214724c3d4b6220c3e488e703fac35666e3afcee1879723c6ab99b60541a259-20240520-181650-902933-130-pangeo/lib/python3.11/site-packages/holoviews/core/ndmapping.py:1023, in UniformNdMapping.__rmul__(self, other)
   1022 def __rmul__(self, other):
-> 1023     return self.__mul__(other, reverse=True)

TypeError: Overlayable.__mul__() got an unexpected keyword argument 'reverse'

And when I tried to figure out what was wrong by simply plotting the image object, I see that it still has the widget attached:

image

produces:

If this is the problem, how do I remote the widget from this object?
Or if not, how can I overlay my points object on the map?

In case it’s useful, the whole rendered notebook is here: Jupyter Notebook Viewer (nbviewer.org)

[ahuang11]

Perhaps
image.object * station_plot

[rsignell]

Yep! That works!

New notebook at: Jupyter Notebook Viewer (nbviewer.org)

[Marc]

If I drop into Panel for the most part I can get the below working.

import panel as pn
import xarray as xr
import holoviews as hv
import hvplot.xarray
import pandas as pd
import hvplot.pandas
import numpy as np
import datetime as dt
pn.extension()

ds = pn.cache(xr.tutorial.open_dataset)('air_temperature')
ds = ds.assign_coords(lon=(((ds.lon + 180) % 360) - 180), lat=ds.lat*1.0)
options = list(np.unique(ds.time))
select = pn.widgets.DiscreteSlider(options=options, name='Time')

def plot(time):
    return ds.sel(time=time, method="nearest").hvplot('lon', 'lat', geo=True, cmap='viridis')

image_dynamic_map = hv.DynamicMap(pn.bind(plot, select))
tap_stream = hv.streams.Tap(source=image_dynamic_map, x=-88, y=40)

df = pd.DataFrame({'lon':[-88.0], 'lat':[40], 'station':[0]})

station_plot = df.hvplot.points(x='lon', y='lat', geo=True, by='station', size=15, frame_height=200)
overlay = image_dynamic_map * station_plot

pn.Column(
    select,
    overlay,
    tap_stream,
).servable()

Getting the select widget having nice timestamps is left as an exercise to the reader (I could not figure it out).

[ahuang11]

options = list(np.unique(ds.time)) can be

options = ds["time"].dt.strftime("%Y-%m-%d %H:%M").values.tolist()

[rsignell]

@Marc I don’t see the station marker plotted on your solution either?
(it seems the same result as cell [17] in my solution?)

Am I missing something?
P.S. thanks for taking a look!

[Marc]

Is the station marker not the little blue dot?

[rsignell]

Was there a blue dot on the map? I only saw it in the legend

[ahuang11]

There is; I guess should be red.

[Marc]

It was already there

[rsignell]

I’m so sorry guys. Super embarrassing! I’ve selected Andrew’s original reply as the solution and edited my response. Kind of wish we could delete everything after that!