Adding labels to jittered data

Hello!
As the title suggests I’m interested in adding labels to jittered data. There doesn’t seem to be an attribute to label individual jittered datapoints. Currently I’m just plotting the data and adding the labels after:

source = ColumnDataSource(both_df)
p.circle(x=jitter('Gene', .2), y='Value', color = 'Color', alpha=1, 
            size='Size', source=source)
labels = LabelSet(x=jitter('Gene', .2), y='Value', text='Label', level='glyph',
              x_offset=.6, y_offset=.6, source=source, render_mode='canvas')

Which results in something like this:

forgithub600×600 16.4 KB

After #7915 was merged, the Jitter transform began to cache jitter offsets based on the length of the input series. Generally speaking, for a glyph and a corresponding LabelSet, these should be the same. But to take advantage of the caching, you would need to use the same Jitter object in both places. Currently, you are creating two separate Jitter objects, one for the glyph, and a different one for the LabelSet.

Instead, you should re-use:

x_jit = jitter('Gene', .2)

p.circle(x=x_jit, ...)

labels = LabelSet(x=x_jit, ...)

So that the same cached jitter offsets are use in both places.

Wow! I need to tweak it a bit, but it looks much better. Thanks!

forgithub600×600 16.4 KB