For my digitized notes project, I wrote a few scripts that read Markdown cells from Jupyter notebook files. Specifically, I read a notebook’s non-empty Markdown cells and used them for my search index and flashcard database.

Reading Jupyter notebooks as data

Reading Jupyter notebooks as data is pretty easy! Below I’ll read the non-empty markdown cells.

import nbformat

path = '2018-05-02-HMM.ipynb'


with open(path) as f:
    nb =, NB_VERSION)

markdown_cells = [
    for cell in nb['cells']  # go through the cells
    if cell['cell_type'] == 'markdown' and cell['source']  # skip things like 'code' cells, and empty markdown cells

Rendering Markdown and LaTeX

Below shows how to render markdown in a iPython notebook to show what I can do with a dictionary of Jupyter notebook data. This is also how I render flashcards in digitized notes.

from IPython.display import display, Markdown

display(Markdown("**Below is data loaded from [this other file]({})!** \n\n".format(path)))

Bonus application (added 2018/05/08),

I realized that GitHub doesn’t always render Jupyter notebooks nicely, so here’s how I generate Jekyll posts from my Jupyter notebooks. This post is an example! Another example is my alpha recursion post.