Callisto: read from Jupyter notebooks

Showcase
,
1

Callisto is a package to read Jupyter notebooks. Currently it covers two use cases:

  1. Render notebook cells (or a whole notebook), e.g. to embed a notebook as a chapter in a Typst document,

  2. Extract elements from specific cells, e.g. to insert a cell’s source code or its output in a Typst document.

Here’s a rendering example:

#import "@preview/callisto:0.2.2"

// Render the first 5 cells
#callisto.render(range(5), nb: json("Lorenz.ipynb"))

image
image1191×1309 213 KB

Or using config to set defaults for the reading functions:

#import "@preview/callisto:0.2.2"

#let (render, result) = callisto.config(nb: json("Lorenz.ipynb"))

// Render the first 5 cells
#render(range(5))

Cells can be specified in various ways. For example the cell source can include a header that defines a label (as with Quarto):

#| label: plot1
import matplotlib
plt.plot([1, 2, 3, 4], [1, 4, 9, 16])

We can use this plot1 label to find the cell from Typst:

// Get result of plot1 cell
#result("plot1")

For more examples, see the quick start and tutorial.

Future plans

The current templates (plain and notebook) are very basic. I think there’s potential for designing beautiful templates, so Typst could become one of the preferred ways to render Jupyter notebooks as documents.

For the next major release, I want to experiment with functionality a la prequery, to enable round-trips between Typst and Jupyter: write code blocks in a Typst document, and have them automatically extracted and executed by Jupyter, and the result automatically included in the Typst document.

8 Likes