I was recently trying to embed a clip in a jupyter notebook using https://github.com/Zulko/moviepy, somthing that was working smoothly with python 2.7. After switching to python 3, this was not working anymore and left me scratcthing my head for a solution.

We live in a open-sourced world, so I filled an issue:

https://github.com/Zulko/moviepy/issues/160

which basically describes the problem:

using a video clip generated by:

duration = 20.
fig_mpl, ax = plt.subplots(1, figsize=(6, 5), facecolor='white')

def draw_elementary_pattern(ax, center):
    ax.add_artist(mpl.patches.Wedge(center, 1., 0, 180, width=.1))

def make_frame_mpl(t):
    ax.cla()
    draw_elementary_pattern(ax, t/duration)
    return mplfig_to_npimage(fig_mpl) # RGB image of the figure

animation = mpy.VideoClip(make_frame_mpl, duration=duration)

I am trying to embed it in the notebook using:

mpy.ipython_display("circles.webm", fps=20, loop=1, autoplay=1)

This was working nicely when I was using python 2, but seems to be broken when using python 3:

(The video generated by

animation.write_videofile("circles.webm", fps=20)

would show fine in the browser when opened directly)

$ python3 --version
Python 3.4.3

$ ipython --version
3.1.0

any clue?

screen shot 2015-04-25 at 16 26 43

Let's try to solve this step by step.

In [3]:
!pip install moviepy
Collecting moviepy
  Using cached moviepy-0.2.2.11.tar.gz
Requirement already satisfied (use --upgrade to upgrade): numpy in /Users/lolo/.pyenv/versions/3.4.3/lib/python3.4/site-packages (from moviepy)
Collecting decorator (from moviepy)
  Using cached decorator-3.4.2.tar.gz
Collecting imageio (from moviepy)
  Using cached imageio-1.2.tar.gz
Collecting tqdm (from moviepy)
  Using cached tqdm-1.0.tar.gz
Installing collected packages: decorator, imageio, tqdm, moviepy
  Running setup.py install for decorator
  Running setup.py install for imageio
  Running setup.py install for tqdm
  Running setup.py install for moviepy
Successfully installed decorator-3.4.2 imageio-1.2 moviepy-0.2.2.11 tqdm-1.0
In [4]:
%matplotlib inline
import matplotlib as mpl
import matplotlib.pyplot as plt
from moviepy.video.io.bindings import mplfig_to_npimage
import moviepy.editor as mpy
In [5]:
duration = 3.
fig_mpl, ax = plt.subplots(1, figsize=(6, 5), facecolor='white')

def draw_elementary_pattern(ax, center):
    ax.add_artist(mpl.patches.Wedge(center, 1., 0, 180, width=.1))

def make_frame_mpl(t):
    ax.cla()
    draw_elementary_pattern(ax, t/duration)
    return mplfig_to_npimage(fig_mpl) # RGB image of the figure

animation = mpy.VideoClip(make_frame_mpl, duration=duration)

Great. Now the following code should embed the movie in the notebook:

In [6]:
animation.ipython_display(fps=20, loop=1, autoplay=1)

Out[6]:

Dooh! Something seems wrong - let's check the raw video file:

In [7]:
fname = "/tmp/circle.webm"
animation.write_videofile(fname, fps=20)
mpy.ipython_display(fname, fps=20, loop=1, autoplay=1)
[MoviePy] >>>> Building video /tmp/circle.webm
[MoviePy] Writing video /tmp/circle.webm
[MoviePy] Done.
[MoviePy] >>>> Video ready: /tmp/circle.webm 

Out[7]:

But this video plays fine in my browser (Firefox on a Mac), so something must be wrong in the embedding.

As Zulko noted in the issue's discussion, he is using something like (see https://github.com/Zulko/moviepy/blob/master/moviepy/video/io/html_tools.py):

In [9]:
def video(fname, mimetype):
    """Load the video in the file `fname`, with given mimetype, and display as HTML5 video.
    """
    from IPython.display import HTML
    from base64 import b64encode    
    with open(fname, "rb") as f: 
        video_encoded= b64encode(f.read())
    video_tag= """
<center><video controls style='max-width:100%'>
<source src='data:{mimetype};base64,{b64}' type='video/{mimetype}' loop=1 autoplay>
Your browser does not support the video tag.
</video><center/>""".format(mimetype=mimetype, b64=video_encoded)
    return HTML(data=video_tag)
video(fname, fname.split('.')[-1])
Out[9]:

I remembered during the switch to python3 the different things to change print(), Except and... something around strings that become bytes objects (see https://docs.python.org/3.4/library/base64.html). I also remembered that the excellent https://github.com/ioam/holoviews library as no problem to embed videos in the notebook. Dwelling in their code (see in particular https://github.com/ioam/holoviews/blob/master/holoviews/plotting/__init__.py and https://github.com/ioam/holoviews/blob/master/holoviews/ipython/display_hooks.py) I could find this particular solution working:

In [15]:
def video(fname, mimetype):
    """Load the video in the file `fname`, with given mimetype, and display as HTML5 video.
    """
    from IPython.display import HTML
    from base64 import b64encode
    
    with open(fname, "rb") as f: 
        video_encoded = b64encode(f.read()).decode("utf-8")

    video_tag= """
<center><video controls style='max-width:100%'>
<source src='data:{mimetype};base64,{b64}' type='video/{mimetype}' loop=1 autoplay=1>
Your browser does not support the video tag.
</video><center/>""".format(mimetype=mimetype, b64=video_encoded)
    return HTML(data=video_tag)

video(fname, fname.split('.')[-1])
Out[15]:

Works as expected! Let's try to do a pull request

  • set-up variables

     cd ~/pool/libs/
     github_user='laurentperrinet'
     lib='moviepy'
     project='moviepy3'
     git clone https://github.com/$github_user/$lib
     cd $lib
  • creating a new branch for his project

    git branches
    git remote add $github_user https://github.com/$github_user/$lib
    git checkout -b $project
  • making the one line change to moviepy's code and test it with the minimum working example above:

     vim moviepy/video/io/html_tools.py
     pip install -U .
  • publishing results

    git status
    git commit -am"fix to work with python3"
    git push $github_user $project
  • confirm the pull request

    open https://github.com/$github_user/$lib
  • come back to the master branch

    git checkout master