HvPlot colormesh dynamic colorbar

hvPlot
,
1

I simply want to have a range slider that dynamically updates a 3 color colormap for a xarray quadmesh where the range slider takes the min and max values of the values of the variable plotted and on interaction with the range slider the filtering and colorbar of the hvplot is rerendered.

dss = fstd2nc.Buffer(file_path_list).to_xarray()
max_TT = dss.TT.max().data.compute()
min_TT = dss.TT.min().data.compute()
masked_dss = dss.TT.where((dss.TT < 10) & (dss.TT > min_TT))
masked_dss_2 = dss.TT.where((dss.TT > 10) & (dss.TT < 25))
masked_dss_3 = dss.TT.where((dss.TT > 25) & (dss.TT < max_TT))
hv1 = masked_dss.hvplot(
        cmap='greens', # Echelle de coulleurs
        colorbar=False,
        kind="quadmesh", # Type de graphique, dans notre cas important que ce soit quadmesh car on a une grille a des intervalle non reguliers
        rasterize=True, # Rasterize l'image pour une meilleur qualite et consistence
        data_aspect=1, # On assure de garder l'echelle des donnees eux-memes
        crs=ccrs.PlateCarree(), # On determine la projection des donnees
        projection=ccrs.PlateCarree(), # On projete les donnees sur n'importe laquelle des projections de Cartopy ex LambertConformal https://scitools.org.uk/cartopy/docs/v0.15/crs/projections.html
        project=True,
        geo=True,
        coastline=True,
        widget_location='bottom')
hv2 = masked_dss_2.hvplot(
        cmap='oranges', # Echelle de coulleurs
        colorbar=False,
        kind="quadmesh", # Type de graphique, dans notre cas important que ce soit quadmesh car on a une grille a des intervalle non reguliers
        rasterize=True, # Rasterize l'image pour une meilleur qualite et consistence
        data_aspect=1, # On assure de garder l'echelle des donnees eux-memes
        crs=ccrs.PlateCarree(), # On determine la projection des donnees
        projection=ccrs.PlateCarree(), # On projete les donnees sur n'importe laquelle des projections de Cartopy ex LambertConformal https://scitools.org.uk/cartopy/docs/v0.15/crs/projections.html
        project=True,
        geo=True,
        coastline=True,
        widget_location='bottom')
hv3 = masked_dss_3.hvplot(
        cmap='reds', # Echelle de coulleurs
        colorbar=False,
        kind="quadmesh", # Type de graphique, dans notre cas important que ce soit quadmesh car on a une grille a des intervalle non reguliers
        rasterize=True, # Rasterize l'image pour une meilleur qualite et consistence
        data_aspect=1, # On assure de garder l'echelle des donnees eux-memes
        crs=ccrs.PlateCarree(), # On determine la projection des donnees
        projection=ccrs.PlateCarree(), # On projete les donnees sur n'importe laquelle des projections de Cartopy ex LambertConformal https://scitools.org.uk/cartopy/docs/v0.15/crs/projections.html
        project=True,
        geo=True,
        coastline=True,
        widget_location='bottom')

hv1[0].object * hv2[0].object * hv3[0].object

which results in

image
image1365×463 136 KB

but all of this is very hacky .object and having to specify the limits statically. Below I have made a little demo of what i would like to have but in MatPlotlib and a simple countour plot which bizarly does not update on range slider input that looks like

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.colors as mc

range_slider = pn.widgets.RangeSlider(name='Range Slider', start=0.5, end=1.5, value=(0.3, 1.45), step=0.05)

def rs_print(rangeSlider):
    # print(rangeSlider)
    x,y = np.meshgrid(np.linspace(-3,3), np.linspace(-3,3))
    z = np.exp(-x**2-y**2)*1.5+0.2
    colors = ["green", "yellow", "red"]
    cmap= mc.ListedColormap(colors)
    boundaries=[z.min(),rangeSlider[0],rangeSlider[1], z.max()]
    # print(boundaries)
    m = plt.contourf(x, y, z, levels = boundaries, cmap=cmap)
    plt.colorbar(m, spacing="proportional")
    return m
pn.Column(range_slider, pn.bind(rs_print, rangeSlider=range_slider))

image
image1273×489 12 KB

I would love to have the range slider for either a single color or colormaps (ideally) for one hvplot on the same dataset.

2

Hi @StuckDuckF

According to Matplotlib — Panel you need to use a matplotlib Figure object instead of matplotlib.pyplot.

The code below updates on range slider input. But I don’t understand the rest of your question.

import matplotlib.colors as mc
from matplotlib.figure import Figure
import numpy as np

import panel as pn

range_slider = pn.widgets.RangeSlider(name='Range Slider', start=0.5, end=1.5, value=(0.3, 1.45), step=0.05)

def rs_print(rangeSlider):
    # print(rangeSlider)
    x,y = np.meshgrid(np.linspace(-3,3), np.linspace(-3,3))
    z = np.exp(-x**2-y**2)*1.5+0.2
    colors = ["green", "yellow", "red"]
    cmap= mc.ListedColormap(colors)
    boundaries=[z.min(),rangeSlider[0],rangeSlider[1], z.max()]
    fig0 = Figure(figsize=(8, 6))
    ax0 = fig0.subplots()
    
    m = ax0.contourf(x, y, z, levels = boundaries, cmap=cmap)
    fig0.colorbar(m, spacing="proportional")
    return fig0

pn.Column(range_slider, pn.bind(rs_print, rangeSlider=range_slider)).servable()

panel serve name_of_script.py

2023-01-14_08-49-00

1 Like
3

Thank you @Marc , I am trying to apply that dynamic Color map to an xarray dataset. So when the user inputs range slider it affects the dataset plotted with hvplot with three .where(ds.TT > min_value && ds.TT < range_slider_1) like statements. In my question I do this statically but I want to do so dynamically and with green, yellow and red. At the moment it is static and it is greens, oranges and reds and on an unrelated note I can not get the ESRI or OSM tiles to work.

4

I sorry. But I don’t think I’m able to help here. @maximlt do you know how to solve? Thanks.

1 Like