I really love blogging with Hugo+Blogdown, but unfortunately Blogdown is still mostly restricted to R (although Python is now also possible using the reticulate package).

Jupyter offers a great literate programming environment for multiple languages and so being able to publish Jupyter notebooks as Hugo blogposts would be a huge plus.

I have been looking for a way to incorporate Jupyter notebooks into my blogging workflow for a while now and I narrowed my options down to the following:

• hugo_jupyter Python package (see here for a short tutorial)
• Writing plain vanilla markdown without code output
• Converting Jupyter notebooks to markdown using nbconvert

Today I decided to give nbconvert a go, because I did not want dive into setting up additional libraries.

As it turns out, using nbconvert to convert Jupyter notebooks to .md files is actually really straightforward.

My current setup is:

    blog-folder/
        jupyter-notebooks/
        other-folders/

To create a blogpost all you have to do is cd into the directory that contains your notebooks and run:

jupyter nbconvert --to markdown <notebook-name>.ipynb

## This gets rendered as code:
print('test')

test

Now let’s try some graphs:

from plotnine import *
from plotnine.data import *
import warnings
warnings.filterwarnings('ignore')

mtcars.head()

(ggplot(mtcars, aes(x='disp',y='mpg'))
    + geom_point()
    + geom_smooth()
    + ggtitle('A graph')
)

<ggplot: (-9223371921876142392)>

Running nbconvert will put all images in a folder next to your jupyter notebook. So we need to copy the pic folder to /static/ in our hugo directory and fix the links.

That’s it more or less:) When I find the time, I will look into a more automated setup. Hope you found this post interesting and happy blogging!