Is there a way to use hv.archive for Holoviews in Jupyter Lab?

Question

I am trying to use HoloViews hv.archive function in Jupyter Lab, but no archive is generated when I run the code.

When I run:

hv.archive.auto()

df.testplot() # create holoviews plot

hv.archive.contents()

hv.archive.export()

I get this: Javascript Error: IPython is not defined

When running in Jupyter Notebook I don't get the error and an archive folder is created, but I still can't generate the archived plots.

Any way forward would be much appreciated. Thanks!

Answer 1

Currently hv.archive() doesn't work with Jupyter Lab, but it should work with Jupyter Notebook though.

See this github issue:
https://github.com/holoviz/holoviews/issues/3570

If I run the code below in Jupyter Notebook, it works for me.
Note: run the code below not all at once, but just part by part.
Otherwise I think the filesystem can't keep up and you get an error.

import holoviews as hv

# you can also choose the bokeh backend, but
# no .svg image will be saved
hv.extension('matplotlib')

# start archiving automatically
hv.archive.auto()

# create a simple plot
hv.Curve(range(0,3))

# check contents of current archive
hv.archive.contents()

# export archive to disk
hv.archive.export()

# check status of export
hv.archive.last_export_status()

# recreate simple plot from archive    
import os
from holoviews.core.io import Unpickler
path = os.path.join(hv.archive.notebook_name, 'Curve.hvz')

if os.path.isfile(path):
    obj = Unpickler.load(open(path,"rb"))
    print(obj)

obj

This is what my archive looks like when exported to disk:

More info on archiving can be found here:
http://holoviews.org/user_guide/Exporting_and_Archiving.html