I generate heatmaps with the rect glyph. However it seems I cannot get rid of white lines between the rectangles unless I set an a priori not logical width and height.
Below I plot a 2d gaussian distribution, the x and y axes are just indices, thus evenly spaced by 1 so using rectangles with width and height equal to 1 should fill the whole canvas without white spaces, considering I set line_alpha=0
The dilate=True option helps, but we can still see white lines between the squares.
The only way I found to not have any visible spacing between the squares is to set them larger than what seems a priori required, e.g. 1.1x1.1 instead of 1x1 with dilate=True
import numpy as np
from scipy.signal import gaussian
from bokeh.io import show
from bokeh.plotting import figure
from bokeh.models import LinearColorMapper, ColorBar, ColumnDataSource
from bokeh.palettes import Magma256
some data to plot
x = 10*gaussian(100,std=20)
matrix = np.matmul(x.reshape(100,1),x.reshape(1,100))
IDs = range(100)*100
source = ColumnDataSource( data={‘row’:np.array(sorted(IDs)),‘col’:np.array(IDs),‘mat’:matrix.flatten()} )
mapper = LinearColorMapper(palette=Magma256[::-1], low=10, high=90,low_color=‘grey’,high_color=‘red’)
fig = figure()
fig.rect(x=‘col’,y=‘row’,width=1,height=1,source=source,fill_color={‘field’:‘mat’,‘transform’:mapper},line_alpha=0,line_width=0,dilate=True)
show(fig)
``
Maybe there is a better way to set width and height, or some other parameter, to prevent having any spacing between the rectangles?
Thanks,
Seb