Difficulties creating a holomap of composed rasterized images

The code below should generate a holomap of holoviews rasterized images bunched together using + operators.

import numpy
import matplotlib
import bokeh
from matplotlib.cm import viridis, plasma, gist_ncar, gist_stern, PuBu, Blues
import holoviews
from holoviews.operation.datashader import rasterize
holoviews.extension('bokeh')
#AggregationMethods = [ 'mean','max','mode','std']
AggregationMethods = [ 'mean','max']
Hows = [ 'linear', 'bilinear' ]
# uncomment the next line to see a different error message
#Hows = [ 'nearest', 'linear' ]

OutputImageFilePattern = 'TestImage_%s_%i_%s_%s' 
ImageIntensityScale = 100000
PlotHeight = 100
PlotSizes = [500]
for library in ['numpy', 'matplotlib', 'holoviews']:
    print ( library + ' ' + str(eval(library + '.__version__')))

def PlotCluster(MatrixClusters, MatrixDistances, OutputFile, How, AggregationMethod, PlotSize, PlotHeight = PlotHeight):
    "plot the cluster"
    DistBarData = holoviews.Image(MatrixDistances)
    ShadedDistBar = rasterize (DistBarData, aggregator = AggregationMethod, interpolation = How, dynamic  = True).opts(cmap=PuBu, xaxis=None, yaxis=None, height=PlotHeight, width=PlotSize, tools=['hover']) 
    ClusterBarData = holoviews.Image(MatrixClusters)
    ShadedClusterBar = rasterize (ClusterBarData, aggregator = AggregationMethod, interpolation = How, dynamic  = True).opts(cmap=PuBu, xaxis=None, yaxis=None, height=PlotHeight, width=PlotSize, tools=['hover']) 
    Merged = (ShadedDistBar + ShadedClusterBar).cols(1)
    return Merged

HoloMapDict = {}
numpy.random.seed(1)
UnitDistanceVec = numpy.random.rand(25000)
UnitClusterVec = (numpy.random.rand(25000)*100).astype(int)  
print (UnitDistanceVec[:10])    
print (UnitClusterVec[:10])
BarSize = (PlotHeight,1)
UnitClusterVecRearranged = numpy.array(UnitClusterVec)
MatrixClusters = numpy.tile( UnitClusterVecRearranged, BarSize).astype(int)  
UnitDistanceVecsRarranged = numpy.array(UnitDistanceVec)
UnitDistanceVecAsInt = (numpy.array(UnitDistanceVecsRarranged)/max(UnitClusterVec)*ImageIntensityScale).astype(int)
MatrixDistances = numpy.tile( UnitDistanceVecAsInt, BarSize)

for AggregationMethod in AggregationMethods:
    for How in Hows:
        for PlotSize in PlotSizes:
            OutputFile = OutputImageFilePattern%('',PlotSize,How,AggregationMethod)
            Merged = PlotCluster(MatrixClusters,MatrixDistances, OutputFile, How, AggregationMethod, PlotSize)
            HoloMapDict[(AggregationMethod, How )] = Merged
Combined = holoviews.HoloMap(HoloMapDict, kdims=['Aggregation', 'interpolation'])
holoviews.save(Combined, 'combined', 'html')

The code above does not work as expected and generates the following printout:

numpy 1.16.3
matplotlib 3.0.3
holoviews 1.12.1
[4.17022005e-01 7.20324493e-01 1.14374817e-04 3.02332573e-01
 1.46755891e-01 9.23385948e-02 1.86260211e-01 3.45560727e-01
 3.96767474e-01 5.38816734e-01]
[91 86 65 77 59 47 48 45  8 87]
WARNING:param.Warning: Nesting Layouts within a HoloMap makes it difficult to access your data or control how it appears; we recommend calling .collate() on the HoloMap in order to follow the recommended nesting structure shown in the Composing Data user guide (https://goo.gl/2YS8LJ)
Traceback (most recent call last):
  File "Test2.py", line 50, in <module>
    holoviews.save(Combined, 'combined', 'html')
  File "C:\Users\Work\Anaconda2\envs\py3\lib\site-packages\holoviews\util\__init__.py", line 749, in save
    return renderer_obj.save(obj, filename, fmt=fmt)
  File "C:\Users\Work\Anaconda2\envs\py3\lib\site-packages\holoviews\plotting\renderer.py", line 549, in save
    plot = self_or_cls.get_plot(obj)
  File "C:\Users\Work\Anaconda2\envs\py3\lib\site-packages\holoviews\plotting\bokeh\renderer.py", line 134, in get_plot
    plot = super(BokehRenderer, self_or_cls).get_plot(obj, renderer, **kwargs)
  File "C:\Users\Work\Anaconda2\envs\py3\lib\site-packages\holoviews\plotting\renderer.py", line 190, in get_plot
    obj = collate(obj)
  File "C:\Users\Work\Anaconda2\envs\py3\lib\site-packages\holoviews\plotting\util.py", line 75, in collate
    return obj.collate()
  File "C:\Users\Work\Anaconda2\envs\py3\lib\site-packages\holoviews\core\spaces.py", line 349, in collate
    drop_constant=drop_constant)()
  File "C:\Users\Work\Anaconda2\envs\py3\lib\site-packages\holoviews\core\element.py", line 468, in __call__
    dict(constant_keys))
  File "C:\Users\Work\Anaconda2\envs\py3\lib\site-packages\holoviews\core\element.py", line 519, in _add_dimensions
    new_item[k] = self._add_dimensions(v, dims[::-1], constant_keys)
  File "C:\Users\Work\Anaconda2\envs\py3\lib\site-packages\holoviews\core\element.py", line 508, in _add_dimensions
    new_item = new_item.add_dimension(dim, 0, val)
  File "C:\Users\Work\Anaconda2\envs\py3\lib\site-packages\holoviews\core\spaces.py", line 1804, in add_dimension
    raise NotImplementedError('Cannot add dimensions to a DynamicMap, '
NotImplementedError: Cannot add dimensions to a DynamicMap, cast to a HoloMap first.

I tired using collate on the holomap object created after creating the holomap object and adding data directly to the object, yet I am getting the same error.

I need a solution to have combo boxes that will alow me to switch between different aggregation and interpolation method - if holoviews does not implement it, can I do it with panel? My alternative is using tab layout in panel - however, it would be nice having it with the holoviews combo boxes if possible. In any case, I want it to be a static html page without the need to run a server. Hopefully iy is possible.

[jbednar]

Apart from adding collate, I think all you need is dynamic=False, which isn’t a problem if you’re exporting to HTML anyway. The underlying problem is that rasterize creates a DynamicMap if dynamic=True, and you can’t simply collate and reorganize a dynamic process like that. Instead, this code should be rewritten to create one big DynamicMap with non-dynamic components, but that’s a complete reorganization, and this should work for now since you don’t need the result to be dynamic anyway…

import numpy
import matplotlib
import bokeh
from matplotlib.cm import viridis, plasma, gist_ncar, gist_stern, PuBu, Blues
import holoviews
from holoviews.operation.datashader import rasterize
holoviews.extension('bokeh')
#AggregationMethods = [ 'mean','max','mode','std']
AggregationMethods = [ 'mean','max']
Hows = [ 'linear', 'bilinear' ]
# uncomment the next line to see a different error message
#Hows = [ 'nearest', 'linear' ]

OutputImageFilePattern = 'TestImage_%s_%i_%s_%s' 
ImageIntensityScale = 100000
PlotHeight = 100
PlotSizes = [500]
for library in ['numpy', 'matplotlib', 'holoviews']:
    print ( library + ' ' + str(eval(library + '.__version__')))

def PlotCluster(MatrixClusters, MatrixDistances, OutputFile, How, AggregationMethod, PlotSize, PlotHeight = PlotHeight):
    "plot the cluster"
    DistBarData = holoviews.Image(MatrixDistances)
    ShadedDistBar = rasterize (DistBarData, aggregator = AggregationMethod, interpolation = How, dynamic  = False).opts(cmap=PuBu, xaxis=None, yaxis=None, height=PlotHeight, width=PlotSize, tools=['hover'], axiswise=True, framewise=True) 
    ClusterBarData = holoviews.Image(MatrixClusters)
    ShadedClusterBar = rasterize (ClusterBarData, aggregator = AggregationMethod, interpolation = How, dynamic  = False).opts(cmap=PuBu, xaxis=None, yaxis=None, height=PlotHeight, width=PlotSize, tools=['hover'], axiswise=True, framewise=True) 
    Merged = (ShadedDistBar + ShadedClusterBar).cols(1)
    return Merged

HoloMapDict = {}
numpy.random.seed(1)
UnitDistanceVec = numpy.random.rand(25000)
UnitClusterVec = (numpy.random.rand(25000)*100).astype(int)  
print (UnitDistanceVec[:10])    
print (UnitClusterVec[:10])
BarSize = (PlotHeight,1)
UnitClusterVecRearranged = numpy.array(UnitClusterVec)
MatrixClusters = numpy.tile( UnitClusterVecRearranged, BarSize).astype(int)  
UnitDistanceVecsRarranged = numpy.array(UnitDistanceVec)
UnitDistanceVecAsInt = (numpy.array(UnitDistanceVecsRarranged)/max(UnitClusterVec)*ImageIntensityScale).astype(int)
MatrixDistances = numpy.tile( UnitDistanceVecAsInt, BarSize)

for AggregationMethod in AggregationMethods:
    for How in Hows:
        for PlotSize in PlotSizes:
            OutputFile = OutputImageFilePattern%('',PlotSize,How,AggregationMethod)
            Merged = PlotCluster(MatrixClusters,MatrixDistances, OutputFile, How, AggregationMethod, PlotSize)
            HoloMapDict[(AggregationMethod, How )] = Merged

Combined = holoviews.HoloMap(HoloMapDict, kdims=['Aggregation', 'interpolation']).collate()
holoviews.save(Combined, 'combined', 'html')

[Jacob-Barhak]

This specific problem was solved by James a long time ago - I should have been closed. So I am closing it so it wont bother the developers. There was another issue there with saving the html with the widgets which is related to another issue I started (https://github.com/pyviz/holoviews/issues/3849) that deals with saving.

To close this problem I am reiterating the code that will work.

import numpy
import matplotlib
import bokeh
from matplotlib.cm import viridis, plasma, gist_ncar, gist_stern, PuBu, Blues
import holoviews
from holoviews.operation.datashader import rasterize
holoviews.extension('bokeh')
#AggregationMethods = [ 'mean','max','mode','std']
AggregationMethods = [ 'mean','max']
Hows = [ 'linear', 'bilinear' ]
# uncomment the next line to see a different error message
#Hows = [ 'nearest', 'linear' ]

OutputImageFilePattern = 'TestImage_%s_%i_%s_%s' 
ImageIntensityScale = 100000
PlotHeight = 100
PlotSizes = [500]
for library in ['numpy', 'matplotlib', 'holoviews']:
    print ( library + ' ' + str(eval(library + '.__version__')))

def PlotCluster(MatrixClusters, MatrixDistances, OutputFile, How, AggregationMethod, PlotSize, PlotHeight = PlotHeight):
    "plot the cluster"
    DistBarData = holoviews.Image(MatrixDistances)
    ShadedDistBar = rasterize (DistBarData, aggregator = AggregationMethod, interpolation = How, dynamic  = False).opts(cmap=PuBu, xaxis=None, yaxis=None, height=PlotHeight, width=PlotSize, tools=['hover'], axiswise=True, framewise=True) 
    ClusterBarData = holoviews.Image(MatrixClusters)
    ShadedClusterBar = rasterize (ClusterBarData, aggregator = AggregationMethod, interpolation = How, dynamic  = False).opts(cmap=PuBu, xaxis=None, yaxis=None, height=PlotHeight, width=PlotSize, tools=['hover'], axiswise=True, framewise=True) 
    Merged = (ShadedDistBar + ShadedClusterBar).cols(1)
    return Merged

HoloMapDict = {}
numpy.random.seed(1)
UnitDistanceVec = numpy.random.rand(25000)
UnitClusterVec = (numpy.random.rand(25000)*100).astype(int)  
print (UnitDistanceVec[:10])    
print (UnitClusterVec[:10])
BarSize = (PlotHeight,1)
UnitClusterVecRearranged = numpy.array(UnitClusterVec)
MatrixClusters = numpy.tile( UnitClusterVecRearranged, BarSize).astype(int)  
UnitDistanceVecsRarranged = numpy.array(UnitDistanceVec)
UnitDistanceVecAsInt = (numpy.array(UnitDistanceVecsRarranged)/max(UnitClusterVec)*ImageIntensityScale).astype(int)
MatrixDistances = numpy.tile( UnitDistanceVecAsInt, BarSize)

for AggregationMethod in AggregationMethods:
    for How in Hows:
        for PlotSize in PlotSizes:
            OutputFile = OutputImageFilePattern%('',PlotSize,How,AggregationMethod)
            Merged = PlotCluster(MatrixClusters,MatrixDistances, OutputFile, How, AggregationMethod, PlotSize)
            HoloMapDict[(AggregationMethod, How )] = Merged

Combined = holoviews.HoloMap(HoloMapDict, kdims=['Aggregation', 'interpolation']).collate()
holoviews.save(Combined, 'combined.html')

Thanks again.