sns.heatmap annotation

Hi there!

I just discovered that in the dev version you can annotate a heat map with the contents of another matrix (wohoo!). I’ve google around (with no success) to see if you can do two things:

• Conditional annotation. In my case, I’ve got a big matrix with a lot of 0.0 values, and those are boring. I’d like to annotate only those cells whose value is != 0. Is that possible?
• Annotate every cell even if xticklabels and yticklabels are set to something other than True. I’ve noticed that if I change them (to avoid text cluttering in the x and y labels), the annotations on the map are also displayed with an offset. Is it possible to turn that off?

Thanks for any comments and keep up prettyfing matplotlib! 😄

[MartinHjelm]

Question: with the mask the annotation and values are not shown if cell is true. Is it possible to just keep the values an remove the annotation? I have a matrix where I want to annotate the n largest values for each row.

[diallobakary4]

Thank you @jeiros . I want something similar. I have seen another solution: to access the text object and modify it.

for text in sns_clustermap.texts:
    text.set_size(14)
    if text.get_text() == '2': # conditional for cells one would like to highlight. This can then be changed to the highest value...
        text.set_size(25)
        text.set_weight('bold')
        text.set_style('italic')