I’ve gone ahead and written up a custom WidgetManager
in order to mount and render ipywidgets directly on an empty front-end (i.e. no notebook/lab context, just a webpage). This works great using just vanilla ipywidgets. However, when I send code to the kernel that tries to render a plot (in this case, using plotly), I get a series of error messages in the console:
This is the altered widget code:
from ipywidgets import VBox
from IPython.display import display, HTML
import plotly.graph_objs as go
data = [go.Bar(
x=['giraffes', 'orangutans', 'monkeys'],
y=[20, 14, 23]
)]
layout = {'yaxis': {'range': [0, 25]}}
fig = go.FigureWidget(data, layout)
display(VBox([fig]))
which I’ve confirmed works in both jupyter lab and jupyter notebook.
I’m wondering if the custom widget manager in the front-end is forgoing some setup process that the normal notebook/lab context provides. If anyone has any familiarity with this, I greatly appreciate some advice.