Holoviews annotator not displaying in Jupyter Lab

HoloViews
#1

I’m trying to follow the ML annotator example from pyviz dot org.

I’m running jupyterlab from the pangeo-notebook docker image which includes:

  • holoviews = 1.14.0
  • jupyterlab = 2.2.9
  • labextension: @pyviz/jupyterlab_pyviz v1.0.4

Here is the code from the example:

import holoviews as hv
import geoviews as gv

hv.extension('bokeh')

tiles = gv.tile_sources.EsriImagery()
tiles

pts = dict(
    Longitude = [-121.932619100, -121.932362392, -121.933530027],
    Latitude  = [  36.631164244,   36.629475356,   36.630623206])
    
opts = dict(size=10, line_color='black', padding=0.1, min_height=400)

points = gv.Points(pts).opts(**opts)

points_annotator = hv.annotate.instance()
hv.annotate.compose(tiles, points_annotator(points, annotations=dict(Size=int, Type=str)))

When I run it in jupyterlab, nothing is displayed.

I can display the points and tiles in an overlay: points * tiles. This works as expected based on the example.

But the annotator creates empty output in the notebook. Rather than displaying the expected result from the example.

image
image714×96 1.93 KB

I hope some one can point me in the right direction. Thanks! :slightly_smiling_face:

#2

If you create an environment from the example, the output will appear properly. Here is the download:

https://examples.pyviz.org/assets/ml_annotators.zip

Output appears, but an annoying byproduct that arises are param warnings. When running each section of code, I’ll be given 20+ lines of param warnings.