Text-based Notebooks

Text-based Notebooks#

MyST Markdown notebooks allow you to write your Jupyter Notebook entirely in markdown, utilising the MyST Markdown notebook format. This allows you to store notebook metadata, Markdown, and cell inputs in a text-based format that is easy to read and use with text-based tools.

MyST notebooks have a 1-to-1 mapping with Jupyter notebook, so can be converted to .ipynb files and opened as notebooks in Jupyter interfaces (with jupytext installed). When used with myst_nb, MyST notebooks are also integrated directly into the Execution and Caching machinery.[1]

The MyST notebook Structure#

MyST Markdown Notebooks (or MyST notebooks for short) have four main types of content:

cell/notebook level metadata that are written as YAML wrapped in ---
markdown cells that can be written as CommonMark or MyST Markdown
code cells that are written with the MyST Markdown code-cell directive syntax
raw cells that are written with the MyST Markdown raw-cell directive syntax

Notebook-level metadata#

Begin a MyST notebook file with YAML top-matter metadata, containing at least the file_format: mystnb signifier. This will be used as notebook-level metadata for the resulting Jupyter Notebook. This metadata takes the following form:

---
file_format: mystnb
kernelspec:
  name: python3
otherkey1: val1
otherkey2: val2
---
# Notebook title
...

The kernel that your code cells use is determined by the kernelspec.name field, and should relate to a Jupyter kernel installed in your environment and registered with Jupyter. If no kernel is given, then the default kernel will be used.

Syntax for Markdown#

Anything in-between code cells will be treated as Markdown. You can use any Markdown that is valid MyST. If you are using MyST notebooks with the myst_nb Sphinx extension, you can write Sphinx directives and roles. However, note that most Jupyter notebook environments may not be able to render MyST Markdown syntax.

To denote a break between two markdown cells, use the following syntax:

Some markdown
+++ {"optionalkey": "val"}
More markdown

This will result in two markdown cells in the resulting notebook. The key:val pairs specified in the {} brackets will be cell-level metadata in the second markdown cell.

Syntax for code cells#

When writing MyST notebooks, use the following syntax to define a code cell:

```{code-cell} ipython3
a = "This is some"
b = "Python code!"
print(f"{a} {b}")
```

The argument after {code-cell} (above, ipython3) is optional, and is used for readability purposes. The content inside {code-cell} makes up the content of the cell, and will be executed at build time.

This will result in the following output after building your site:

a = "This is some"
b = "Python code!"
print(f"{a} {b}")

This is some Python code!

Cell-level metadata#

You can begin code-cell blocks with top-matter metadata. These will be used as cell-level metadata in the resulting notebook cell. The same metadata tags can be used as you would in a normal notebook, for example those discussed in Hide code cells:

```{code-cell} ipython3
---
tags: [hide-output]
---
for i in range(20):
    print("Millhouse did not test cootie positive")
```

Yields the following:

for i in range(20):
    print("Millhouse did not test cootie positive")

There is also an alternative short-hand syntax for cell-level metadata. This takes the following form:

```{code-cell}
:key: val
print("hi")
```

For example, the following syntax adds a raises-exception tag to the cell, which means our code will execute without halting the kernel:

```{code-cell} ipython3
:tags: [raises-exception]

raise ValueError("oopsie!")
```

raise ValueError("oopsie!")

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[3], line 1
----> 1 raise ValueError("oopsie!")

ValueError: oopsie!

Convert between MyST notebooks and `.ipynb`#

MyST notebooks can be converted to Jupyter notebooks using the mystnb-to-jupyter CLI command.

$ mystnb-to-jupyter path/to/text-notebook.md
Wrote notebook to: path/to/text-notebook.ipynb

MyST notebooks can also be converted back-and-forth from ipynb files using jupytext, a Python library for two-way conversion of ipynb files with many text-based formats.

To let jupytext know the format of the notebook, add the notebook top-matter similar to:

---
kernelspec:
  name: python3
  display_name: python3
jupytext:
  text_representation:
    extension: .md
    format_name: myst
    format_version: '0.13'
    jupytext_version: 1.13.8
---

Then you can run:

To convert .ipynb to a MyST notebook, run: jupytext notebook.ipynb --to myst
To convert a MyST notebook to .ipynb, run: jupytext mystfile.md --to ipynb

MyST notebooks in Jupyter interfaces#

You can use MyST notebooks in Jupyter interfaces by using Jupytext extensions. This allows you to open a MyST Markdown Notebook as a “regular” Jupyter Notebook in Jupyter Lab and the Classic Notebook interface. For more information, see the Jupytext documentation.

MyST notebooks in Jupyter Book#

In addition to using MyST notebooks with Sphinx, you may also use them with the Jupyter Book project. See Notebooks written entirely in Markdown.

Code from Files#

Warning

This is an experimental feature that is not part of the core MyST markup specification, and may be removed in the future. Using :load: will also overwrite any code written into the directive.

myst_nb provides a convenience feature for importing executable code into a {code-cell} from a file. This can be useful when you want to share code between documents. To do this you specify a load metadata attribute such as:

```{code-cell} ipython3
:load: <path>
```