Introductory example doesn't seem to work #415

quasiben · 2019-05-07T15:56:00Z

Setup

Machine: OSX
relevant conda packages:

jupyterlab 0.35.5
panel 0.5.1
python 3.7.2
matplotlib 3.0.3

I naively copied the sine example on the introduction page into a jupyter notebook. The mpl plot appeared however manipulating the widgets produced no outcome:

import numpy as np, pandas as pd, matplotlib.pyplot as plt
%matplotlib inline

def mplplot(df, **kwargs):
    fig = df.plot().get_figure()
    plt.close(fig)
    return fig

def sine(frequency=1.0, amplitude=1.0, n=200, view_fn=mplplot):
    xs = np.arange(n)/n*20.0
    ys = amplitude*np.sin(frequency*xs)
    df = pd.DataFrame(dict(y=ys), index=xs)
    return view_fn(df, frequency=frequency, amplitude=amplitude, n=n)

sine(1.5, 2.5)

import panel as pn
pn.extension()

pn.interact(sine)

philippjfr · 2019-05-07T16:06:51Z

You have the JupyterLab extension installed? You can get it with:

jupyter labextension install @pyviz/jupyterlab_pyviz

quasiben · 2019-05-07T16:12:30Z

Ah! I did not have that installed! Thanks!!!

quasiben · 2019-05-07T16:16:09Z

After installing the extension I am still no seeing any behavior when moving the sliders

philippjfr · 2019-05-07T16:16:10Z

Should maybe mention it again in the Introductory paragraph.

philippjfr · 2019-05-07T16:18:12Z

That's bizarre, does the browser JS console have any errors in it? Could you also try clearing the notebook and reloading the page?

quasiben · 2019-05-07T16:21:10Z

There is a JS console error:

44VM62:7 Uncaught TypeError: Cannot read property '1001' of undefined
    at eval (eval at get (lab:140), <anonymous>:7:38)
    at i.execute (<anonymous>:140:220931)
    at e.<anonymous> (<anonymous>:140:142322)
    at e.t.emit (<anonymous>:140:68660)
    at e.emit (<anonymous>:140:68847)
    at e._setv (<anonymous>:140:25486)
    at e.setv (<anonymous>:140:25916)
    at e.set [as value] (<anonymous>:140:23724)
    at e._slide (<anonymous>:216:6747)
    at Object.<anonymous> (<anonymous>:216:5552)

philippjfr · 2019-05-07T16:25:25Z

That definitely looks to be responsible for the issue. I've encountered that once before but it went away when I tried again in a new session so I was never able to debug it. I'll try to reproduce using your setup:

Install new environment (without JLab extension)
Run Introduction
Install JLab extension
Rerun notebook again.

For better reproducibility could you tell me if you restarted the JLab server, cleared the notebook and/or reloaded the browser window in between any of those steps? Also lastly, could you confirm you have pyviz_comms 0.7.2?

quasiben · 2019-05-07T16:30:01Z

I cleared the notebook, restarted the server, and started an incognito browser session:

❯ conda list pyviz_comms
# packages in environment at /Users/bzaitlen/miniconda3/envs/dask-dev:
#
# Name                    Version                   Build  Channel
pyviz_comms               0.7.2                      py_0

quasiben · 2019-05-07T16:30:10Z

Thanks for looking into this!

philippjfr · 2019-05-07T16:30:35Z

And it's still not working now?

quasiben · 2019-05-07T16:31:06Z

correct

philippjfr · 2019-05-07T16:32:02Z

Thanks for looking into this!

No problem, pretty crucial to have first set up going smoothly. Was hoping to release an official announcement blog post next week. Having new users run into this issue immediately is definitely not good.

philippjfr · 2019-05-07T16:33:55Z

Oh one more question, what's the bokeh version? Will assume it's 1.1.0 but you never know.

quasiben · 2019-05-07T16:35:51Z

❯ conda list bokeh
# packages in environment at /Users/bzaitlen/miniconda3/envs/dask-dev:
#
# Name                    Version                   Build  Channel
bokeh                     1.1.0                    py37_0

philippjfr · 2019-05-07T16:44:32Z

Ugh, bad news, I can't reproduce it. One last question did you get the packages conda-forge?

quasiben · 2019-05-07T16:47:11Z

I don't think so, it looks like they all came from defaults. Should I use conda-forge ?

philippjfr · 2019-05-07T16:49:02Z

Ah okay, I'll try a new env from defaults now. Hadn't even realized everything made it onto defaults already.

quasiben · 2019-05-07T16:50:37Z

i can also build a new env with conda-forge.

philippjfr · 2019-05-07T16:53:39Z

That would be appreciated, thanks for being so patient debugging this with me!

philippjfr · 2019-05-07T16:57:55Z

Yes! It's broken with the packages from defaults, now I just need to figure out why.

philippjfr · 2019-05-07T17:48:38Z

Okay, it seems to be a race condition caused by placing the pn.extension() call into the same cell as the first output. I don't know why it occurs in the defaults environment and not in the conda-forge one but I'm guessing it could be the JupyterLab version difference between the two (0.35.5 vs. 0.35.6).

quasiben · 2019-05-07T18:01:00Z

Yup! Moving the call to pn.extension() to another cell fixed the issue. Thanks! Do you want to close this and start another issue ? Also, I'm happy to help with docs if you'd like

philippjfr · 2019-05-07T18:03:12Z

Going to push a minor version of the JupyterLab extension shortly once this is merged: holoviz/pyviz_comms#36

That should fix it when both are in the same cell too.

philippjfr · 2019-05-07T18:08:51Z

Okay, just merged and pushed the updated JupyterLab extension to npm. Would you mind trying to update?

Also, I'm happy to help with docs if you'd like

That would be hugely appreciated, so far I've basically written almost all of them with a little editing here or there by @jbednar. They would definitely benefit from a second pair of eyes, even if you just mean the mention of the JLab extension in the introduction.

sdc50 · 2019-05-07T21:36:04Z

@philippjfr I was having the same issue (but with a different notebook). I just reinstalled the updated JupyterLab extension and that seems to have fixed it.

I now occasionally get a different error:

Uncaught Error: Cannot send
    at CommHandler.send (comm.js:111)
    at Object.sendClosure [as send] (renderer.js:137)
    at process_events (eval at get (lab:140), <anonymous>:43:10)
    at eval (eval at get (lab:140), <anonymous>:91:27)

But it only happens with the following steps:

Render a cell
Restart the kernel (without clearing cell outputs)
Refresh the browser page
Render the cell again
Click on one of the widgets

Even with the error showing up in the console everything still seems to work. I just thought I'd mention it in case it was related.

philippjfr · 2019-05-07T21:55:42Z

Thanks @sdc50 that's probably an unrelated issue. I think the issue there is that the comm which is responsible for affirming that the frontend has received and processed an event from the backend is not being correctly re-established after reloading the page. Those events are important but not crucial which is why it all appears to work.

philippjfr · 2019-05-17T23:04:28Z

Going to close this since the immediate issue has been addressed.

phimit · 2020-04-01T10:34:49Z

Same problem here on Macos Mojave

Installed everything from scratch in a new environnment, recent packages.
Nothing shows in both notebook and jupyter lab.
(and no hv-extension-comm appears in notebook terminal)
I also put pn.extension in separate cell.

jupyter 1.0.0 py38_7
jupyter_client 6.1.2 py_0
jupyter_console 6.1.0 py_0
jupyter_core 4.6.3 py38_0
jupyterlab 2.0.1 py_0 conda-forge
jupyterlab_server 1.1.0 py_0 conda-forge

panel 0.9.3 py_0 pyviz
param 1.9.3 py_0 pyviz
pyct 0.4.6 py_0 pyviz
pyct-core 0.4.6 py_0 pyviz
pyviz_comms 0.7.4 py_0 pyviz

philippjfr · 2020-04-01T10:39:13Z

Can't reproduce, @phimit do you have the jupyterlab extension installed?

  jupyter labextension install @pyviz/jupyterlab_pyviz

phimit · 2020-04-01T10:42:30Z

Yes I have.
I should have added my conda version: current version: 4.6.9

If it helps, here's the output from the classic notebook:
[W 12:30:12.681 NotebookApp] 404 GET /nbextensions/hv-comm.js?v=20200401122952 (::1) 34.32ms referer=http://localhost:8889/notebooks/Test.ipynb
[W 12:30:12.688 NotebookApp] 404 GET /nbextensions/hv-extension-comm.js?v=20200401122952 (::1) 5.61ms referer=http://localhost:8889/notebooks/Test.ipynb
[W 12:30:12.705 NotebookApp] 404 GET /nbextensions/pyviz_comms.js?v=20200401122952 (::1) 5.03ms referer=http://localhost:8889/notebooks/Test.ipynb
[W 12:30:13.072 NotebookApp] 404 GET /nbextensions/widgets/notebook/js/extension.js?v=20200401122952 (::1) 4.01ms referer=http://localhost:8889/notebooks/Test.ipynb

phimit · 2020-04-01T12:42:31Z

Tested on a linux machine, same conditions (new conda env), same results (both with classic notebook and lab). no hv-extension-comm warning though.

The only weird common thing is I had to install matplotlib... I thought it was already in base conda.

evanfollis · 2020-04-07T14:11:18Z

I am having the same issue. pn.interact(fn) does not do anything. I am running panel version 0.9.3

philippjfr · 2020-04-07T14:23:32Z

Could both of you try to upgrade to 0.9.5? All the packaging issues there should have been resolved now.

evanfollis · 2020-04-07T14:28:21Z

Yes! 0.9.5 fixed the issue. Thank you!

phimit · 2020-04-09T08:59:53Z

working now, thanks ! Le mar. 7 avr. 2020 à 16:28, Evan Follis <[email protected]> a écrit :

…

Yes! 0.9.5 fixed the issue. Thank you! — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#415 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACCNMCC6KNZVVYIDVWT35CDRLM2BJANCNFSM4HLKHIOA> .

-- I needed a clue and a drink; one of them I knew where to find.

philippjfr · 2020-04-09T09:43:50Z

Great! Thanks for reporting back.

quasiben closed this as completed May 7, 2019

quasiben reopened this May 7, 2019

philippjfr closed this as completed May 17, 2019

yanovs mentioned this issue Dec 6, 2022

Panel embed not working with more than one pane in the same panel/layout #1222

Open

Introductory example doesn't seem to work #415

Introductory example doesn't seem to work #415

Comments

quasiben commented May 7, 2019

Setup

philippjfr commented May 7, 2019

quasiben commented May 7, 2019

quasiben commented May 7, 2019

philippjfr commented May 7, 2019

philippjfr commented May 7, 2019

quasiben commented May 7, 2019

philippjfr commented May 7, 2019

quasiben commented May 7, 2019

quasiben commented May 7, 2019

philippjfr commented May 7, 2019

quasiben commented May 7, 2019

philippjfr commented May 7, 2019

philippjfr commented May 7, 2019 • edited Loading

quasiben commented May 7, 2019

philippjfr commented May 7, 2019

quasiben commented May 7, 2019

philippjfr commented May 7, 2019

quasiben commented May 7, 2019

philippjfr commented May 7, 2019

philippjfr commented May 7, 2019 • edited Loading

philippjfr commented May 7, 2019 • edited Loading

quasiben commented May 7, 2019

philippjfr commented May 7, 2019

philippjfr commented May 7, 2019

sdc50 commented May 7, 2019

philippjfr commented May 7, 2019

philippjfr commented May 17, 2019

phimit commented Apr 1, 2020

philippjfr commented Apr 1, 2020

phimit commented Apr 1, 2020

phimit commented Apr 1, 2020

evanfollis commented Apr 7, 2020

philippjfr commented Apr 7, 2020

evanfollis commented Apr 7, 2020

phimit commented Apr 9, 2020 via email

philippjfr commented Apr 9, 2020

philippjfr commented May 7, 2019 •

edited

Loading

philippjfr commented May 7, 2019 •

edited

Loading

philippjfr commented May 7, 2019 •

edited

Loading