Rasterizing curves and area charts using DynamicMap

[minhphan]

I have quite a complicated DynamicMap in Holoviews with the call function returning a Layout Object of Multiple Overlays. The structure is basically like this:

Dynamic Map → Layout (2 Columns of 9 Overlays) → Overlay from 8-9 Curves.

I want to rasterize this DynamicMap but don’t know how to start. I tried rasterize(DynamicMap) but it returned

(Value Error: the following streams are set to be authamically linked to a plot, but no stream_mapping specifying which item in the layout to link it to was found: PlotSize(height=400, scale=1.0, width=400), RangeXY(…). This is my first attempt trying to do something like this. I used to have a HoloMap with Panel, but for some additional changes I had to resort to DynamicMap. Because of the structure, the file turn out to be quite big to save as html using Bokeh (taking hours to just render 300 frames).

My code is roughly currently like this:

 def plot_curve(dropdown1_value, dropdown2_value, dropdown3_value):
            
            if df_res.empty:
                return hv.Layout([self.null_graph(year) for year in years].append(self.null_graph('2015-2022'))).cols(2)
            else:
                year_plots = []
                for year in years:
                    df_year = df_res[df_res['year']==year]
                    year_plots.append(self.build_year_plot(data=df_year, title=f"{self.site_name}, {year}", legend_args= self.legend, min_freq=min_freq, max_freq=max_freq))
                
                year_plots.append(self.build_year_plot(data=df_res, title=f"{self.site_name}, 2015-2022",legend_args=self.legend, min_freq=min_freq, max_freq=max_freq))
                return hv.Layout(year_plots).cols(2)

        dmap = hv.DynamicMap(plot_curve, kdims=["dropdown1", "dropdown2", "dropdown3"])
        dmap = dmap.redim.values(dropdown1=dropdown1_items.keys(), dropdown2=dropdown2_items.keys(), dropdown3=dropdown3_items).redim(dropdown1 = dropdown1_name, dropdown2 = dropdown2_name, dropdown3=dropdown3_name)
        return dmap

and build_year_plot is roughly like this

list_of_curves = list()
  

# build curves and shaded areas from subdata
for label, group, legend_color in zip(legend_args[0].keys(), groupedDict.values(), legend_args[1]):

    curve = hv.Curve(
        data,
        kdims='frequency',
        vdims='mean',
    ).opts(...)

    std_area = hv.Area(
        data,
        kdims='frequency',
        vdims=['lb', 'ub']
    ).opts(...)
    
    list_of_curves.append(curve*std_area)

  
# build overlay
return hv.Overlay(list_of_curves).opts(
    height=400, 
    width=600,
    ...)