Error: Module jupyter-matplotlib, semver range ^0.5.0 is not registered as a widget module

[mfrawley]

I nearly went crazy trying to figure this out, I don't know if it's intentional but the version of the nodejs package (under js) in master is 0.5.0, whereas the overall python/conda package version was 0.4.0.
When installing the extension without specifying any specific version in a fresh conda environment (via docker) the above is happening every time. Using the update GUI in jupyter lab's extension manager seems to have resolved. Hope this helps someone.

[martinRenou]

It is intentional that the front-end version and the back-end versions are different. They have diverged at some point in the development (I guess at some point the front-end had lots of development while the back-end did not change at all).

As said in the README, if you install jupyter-matplotlib for JupyterLab, you need to install the JupyterLab extension from the GUI or running:

jupyter labextension install @jupyter-widgets/jupyterlab-manager jupyter-matplotlib

So, using conda, you need to run the following in a fresh environment:

conda install -c conda-forge ipympl nodejs
jupyter labextension install @jupyter-widgets/jupyterlab-manager jupyter-matplotlib

Would you mind if we close this issue? I feel like it's working as intended.

[mfrawley]

Hi @martinRenou thanks for the prompt reply, but I did install it as you state, and got the error, here is end of my Dockerfile:

RUN conda install -c conda-forge widgetsnbextension
RUN conda install -c conda-forge ipywidgets
RUN conda install -c conda-forge ipympl
RUN jupyter labextension install --minimize=False @jupyter-widgets/jupyterlab-manager
RUN jupyter labextension install --minimize=False jupyter-matplotlib

I set minimize=False to make the build faster but it has nothing to do with the issue.

[martinRenou]

Mmh, what happens if you enforce the version with:

RUN jupyter labextension install --minimize=False jupyter-matplotlib@0.5

[mfrawley]

No, alas, after trying that the extension manager shows the Update link just as it did before and I get the same error. Perhaps this is specific to operation inside Docker, I'm thinking that some json config file gets written but not picked up..I don't know enough about how it works.

[martinRenou]

Perhaps this is specific to operation inside Docker

That was what I wanted to say. It's not the first time I see this kind of comment when using Docker and JupyterLab (see jupyter-widgets/ipywidgets#2623).

• What happens if you simply refresh the page? Does it find the extension?

• Do core widgets work?

from ipywidgets import Button
Button(description='test')

[mfrawley]

• Do core widgets work?

from ipywidgets import Button
Button(description='test')

The above works..the code I was having the error when trying to plot was more or less:

plt.figure(1, figsize=(9,5))
plt.plot(df[['foo', 'bar']).resample('1W').count())
plt.tight_layout()
plt.show()
[Error shown below cell in jupyter notebook]
Error displaying widget: model not found

Where df is a pandas dataframe with a datetime index.

• What happens if you simply refresh the page? Does it find the extension?
  No, it makes no difference.

[mfrawley]

Self-contained example in a separate notebook:

import pandas as pd
import matplotlib.pyplot as plt

df = pd.DataFrame({'a': [1,2,3]})

plt.figure(2)
plt.plot(df['a'])
plt.show()

This gives "Loading widget" but the same JS exception as the title of this issue.

[mfrawley]

I appreciate the help but now even the GUI solution isn't working ..I might be unresponsive for a few days but would be willing to try to debug in the future.

[martinRenou]

I am clueless right now. I guess I will have to try it myself at some point. Maybe it would be good to add that to documentation somewhere (Simple Dockerfile with ipywidgets and some other widgets libraries running in JupyterLab)

[slopedog]

I've got this problem, too (same error in console). It was working before, and it died without my changing anything. I can trigger it with

%matplotlib widget
import matplotlib.pyplot as plt

plt.figure()
plt.plot([1,2])
plt.show()

Since then, I tried:

pip uninstall -y ipywidgets ipympl
pip freeze | grep jupy | xargs pip uninstall -y
pip install ipympl ipywidgets jupyterlab --upgrade
jupyter labextension install @jupyter-widgets/jupyterlab-manager
jupyter labextension install jupyter-matplotlib
jupyter lab

But this doesn't fix anything. Removing my ~/.jupyter doesn't help, either.
I have these versions of things:

jupyter-client==5.3.4
jupyter-core==4.6.1
jupyterlab==1.2.4
jupyterlab-server==1.0.6
ipykernel==5.1.3
ipympl==0.4.0
ipywidgets==7.5.1
matplotlib==3.1.2

JupyterLab v1.2.4
Known labextensions:
   app dir: /home/jah/Hank/RadioInversion2/Python/env/share/jupyter/lab
        @jupyter-widgets/jupyterlab-manager v1.1.0  enabled  OK
        jupyter-matplotlib v0.5.0  enabled  OK

Uninstalled core extensions:
    jupyter-matplotlib

That strange "Uninstalled core extensions" popped up when I tried using the GUI to install and uninstall jupyter-matplotlib, I think... I can't figure out a way to get rid of it.

[slopedog]

Oh - I probably left out the most important thing - I'm not using Docker. This is in a virtualenv on Ubuntu 18.04.

[AllanChain]

In my case (JupyterLab 2.1.0), I installed ipympl v0.5.0 months ago, and recently installed jupyter-matplotlib@0.7.2, causing the verion problem Module jupyter-matplotlib, semver range ^0.6.0 is not registered as a widget module. It's strange that upgrading ipympl to v0.5.6 and reinstall labextension solved the problem, but downgrading jupyter-matplotlib to 0.6.0 did not...

[apoorvmote]

I was in similar situation. It started failing after I updated juypter lab. I had to force reinstall ipympl as follows

pip3 install -I ipympl
jupyter labextension install @jupyter-widgets/jupyterlab-manager jupyter-matplotlib
jupyter lab build

After that it started working

[martinRenou]

@AllanChain It's strange that upgrading ipympl to v0.5.6 and reinstall labextension solved the problem, but downgrading jupyter-matplotlib to 0.6.0 did not...

This is because jupyter-matplotlib 0.6.0 does not work on Jupyterlab 2:
https://github.com/matplotlib/ipympl#install-an-old-jupyterlab-extension