tqdm_notebook fails with IPython 3

[QuLogic]

Running something simple like:

import tqdm
for i in tqdm.tqdm_notebook(range(1000000)):
    i * 2

produces this traceback:

NameError                                 Traceback (most recent call last)
<ipython-input-6-207cef0f8cd2> in <module>()
      1 import tqdm
----> 2 for i in tqdm.tqdm_notebook(range(1000000)):
      3     i * 2

~/.local/lib/python3.4/site-packages/tqdm/__init__.py in tqdm_notebook(*args, **kwargs)
     17     """See tqdm._tqdm_notebook.tqdm_notebook for full documentation"""
     18     from ._tqdm_notebook import tqdm_notebook as _tqdm_notebook
---> 19     return _tqdm_notebook(*args, **kwargs)
     20 
     21 

~/.local/lib/python3.4/site-packages/tqdm/_tqdm_notebook.py in __init__(self, *args, **kwargs)
    182         # self.sp('', close=True)
    183         # Replace with IPython progress bar display (with correct total)
--> 184         self.sp = self.status_printer(self.fp, self.total, self.desc)
    185         self.desc = None  # trick to place description before the bar
    186 

~/.local/lib/python3.4/site-packages/tqdm/_tqdm_notebook.py in status_printer(file, total, desc)
    103         # Prepare IPython progress bar
    104         if total:
--> 105             pbar = IntProgress(min=0, max=total)
    106         else:  # No total? Show info style bar with no progress tqdm status
    107             pbar = IntProgress(min=0, max=1)

NameError: name 'IntProgress' is not defined

using 4.7.4-a581bbd.

The conditional imports seem a bit confused about versions. The comments in the first block say ipywidgets is IPython 4.x, and IPython.html.widgets is 3.x/2.x, but the second block always imports from ipywidgets.

[lrq3000]

I'll check it out, thank you for reporting the bug. About the import of ipywidgets, it's normal, it's because all imports are aliased to ipywidgets to uniformize the lines of codes importing the widgets after.

[QuLogic]

Aliasing does not appear to apply to imports in either Python version:

$ python
Python 2.7.11 (default, Mar 31 2016, 20:46:51) 
[GCC 5.3.1 20151207 (Red Hat 5.3.1-2)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import lxml as lx
>>> from lx import etree
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: No module named lx
>>> import lxml.etree

$ python3
Python 3.4.3 (default, Mar 31 2016, 20:42:37) 
[GCC 5.3.1 20151207 (Red Hat 5.3.1-2)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import lxml as lx
>>> from lx import etree
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: No module named 'lx'
>>> import lxml.etree

[casperdcl]

that's interesting. Nobody reported issues when I added this. grrr...

[lrq3000]

@casperdcl oops sorry no, mea culpa. I tried with IPython v4 (which works because there's no aliasing), but I didn't try with other versions as I was convinced import aliasing would work in Python.

[casperdcl]

merged in #190 - though python needs to make up its mind about warnings...

[casperdcl]

@philipsd6 I received a notification that you commented here but I can't seem to find your comment. Did you delete it?

[philipsd6]

Yup, I was being dumb, so I erased the evidence.

[casperdcl]

Oh, ok, great! Feel free @philipsd6 and @QuLogic to 👍 this comment!

[lrq3000]

Ah good, because I was already looking on how to fix the issue XD

[philipsd6]

Just for reference, I have a notebook that creates Jupyter kernels by setting up a new virtualenv, installing ipykernel, and creating the kernel file. All I was missing was to also install ipywidgets in the venv. See? Dumb. ;)

[stefanv]

@philipsd6 The error message when ipywidgets is missing is pretty obscure :/ NameError: name 'IntProgress' is not defined <-- Not a great hint where to look.

[b3by]

The error seems to still be there with iPython5!

[balarsen]

I certainly get this error still

Software Version
Python 3.6.1 64bit [GCC 4.2.1 Compatible Apple LLVM 6.0 (clang-600.0.57)]
IPython 5.3.0
OS Darwin 15.6.0 x86_64 i386 64bit
matplotlib 2.0.0
numpy 1.11.3
xarray 0.9.1
tqdm 4.11.2
Fri Mar 31 16:47:45 2017 MDT

[vignesh-m]

The short answer is to install ipywidgets

[stevenmanton]

I'm having the same issue. Installing ipywidgets didn't fix it, unfortunately.

[caidish]

I'm having the same issue.In my windows desktop,it works well.However,when I run the same code on Ubuntu,it breaks down.

[davidsmandrade]

I've installed ipywidgets too and it not working yet. Obs: I'm running on ubuntu. Thanks!

[andrewzwicky]

I tried this and got it to work for me, not sure if this will help anyone else or not.

jupyter-widgets/ipywidgets#541

jupyter nbextension enable --py --sys-prefix widgetsnbextension

Using (irrelevant packages omitted):
ipykernel==4.6.1
ipyparallel==6.0.2
ipython==6.0.0
ipython-genutils==0.2.0
ipywidgets==6.0.0
jupyter-client==5.0.1
jupyter-core==4.3.0
nbconvert==5.1.1
nbformat==4.3.0
notebook==5.0.0
tqdm==4.11.2
widgetsnbextension==2.0.0

[caidish]

@andrewzwicky Thanks a lot!It really helps.jupyter nbextension enable --py --sys-prefix widgetsnbextension

[stolsvik]

Not really sure if this helped or not, but after dozens of attempts, I ran both "install" and "enable". It worked afterwards. One should think the enable performed install, but the output of the two commands was at least different. (Do note that this wasn't exactly a scientific experiment, as I probably had changed dozens of parameters of the system by that time..!)

Anyway: jupyter nbextension install --py --sys-prefix widgetsnbextension

[rodrigoruiz]

I'm still having the same issue when running on Google Colab.

[chengs]

@rodrigoruiz ipywidget is not supported by Google Colab. googlecolab/colabtools#60

[jcorro2]

I did this to make it work:

conda install -c conda-forge ipywidgets
jupyter nbextension enable --py widgetsnbextension
jupyter labextension install @jupyter-widgets/jupyterlab-manager
conda install jupyter

The last command is likely to kill your running kernel. If you are training or just trained a large model, I'd suggest you save it to disk before running the last command so you can recover from where you left.