jupyterlite-pdf-exporter

[!TIP] While this extension was originally designed for JupyterLite in mind and was thus named jupyterlite-pdf-exporter, it is agnostic to the Jupyter environment. It works in JupyterLite, JupyterLab 4.x, Jupyter Notebook 7, and JupyterHub deployments, all from a single installation. You do not need a LaTeX distribution set up or a server to export your notebooks to PDF.

Classifiers	Description
Builds
Try it out!
PyPI downloads
conda package

A serverless PDF exporter for JupyterLite, JupyterLab, and Jupyter Notebook, based on WebAssembly distributions of Pandoc and Typst.

• This Jupyter extension registers a PDF exporter with JupyterLite's INbConvertExporters interface.
• In JupyterLab and Jupyter Notebook, it adds an "Export Notebook to PDF" command to the File menu and the command palette.

It runs fully in the browser in a serverless fashion, so it works the same way across all three. It does not require a LaTeX distribution or a server, and it works in environments where you cannot install software, such as on a Chromebook or in a locked-down corporate environment. The PDF is downloaded to your machine at a location of your choice.

Why?

The usual way to convert a notebook into a PDF is nbconvert driving a LaTeX distribution such as TeX Live or MiKTeX, or a headless browser via Playwright (WebPDF). These typically run on a server, and often need administrator rights to install. On a Chromebook, a managed corporate laptop, or any machine where you cannot install software easily, setting them up is frequently not an option. In JupyterLite, there is no server at all, so that route does not exist and notebooks could not be exported to PDF this way.

This extension provides a different mechanism to export notebooks to PDF that runs entirely in your browser:

1. Pandoc, compiled to WebAssembly, converts the notebook to a Typst markup IR (intermediate representation).
2. The Typst compiler, a modern typesetting system that stands in for LaTeX and is also compiled to WebAssembly, renders that IR into a PDF.

Installation

To install the extension into your Jupyter deployment, execute:

pip install jupyterlite-pdf-exporter

• For JupyterLite, rebuild your JupyterLite distribution after installing, so the extension is bundled into your site.
• For JupyterLab or Jupyter Notebook, that is all you need. Start (or restart) the app and you should see "PDF (via jupyterlite-pdf-exporter)" in the "Save and Export Notebook As" section under the "File" menu.

Uninstalling the extension

To remove the extension from your JupyterLite deployment, execute:

pip uninstall jupyterlite-pdf-exporter

and rebuild your JupyterLite distribution.

Usage

In both JupyterLab and JupyterLite environments, the entry point to usage lives in the same place, under "File" > "Save and Export Notebook As". The exported PDF is downloaded to your machine at a location of your choice.

Requirements

• Either:

  • JupyterLite >=0.7.0,<0.9, or

  • JupyterLab >= 4.5.0

  • Jupyter Notebook >=7.5

    This extension does not depend on or enforce a particular version of JupyterLite or JupyterLab, but please note that cutting-edge versions of JupyterLite, JupyterLab, and Jupyter Notebook may introduce breaking changes. If you face any troubles, please file an issue!

• A modern web browser with support for WebAssembly and Web Workers (e.g., Chrome, Firefox, Safari, Edge, and so on). All browsers supported by JupyterLite should work with this extension.

• The extension relies on WebAssembly distributions of Pandoc and Typst. These distributions are quite large (over 50 MiB) and may take some time to download and initialise when the extension is first used. For a better user experience, it is recommended to use this extension in an environment with a stable and reasonably fast internet connection.

JupyterLite

• Install this extension in your JupyterLite deployment via pip install jupyterlite-pdf-exporter and rebuild your JupyterLite distribution.
• Open a notebook in JupyterLite, click on the "File" menu, and select "Save and Export Notebook As" > "PDF".
  • You may also run "Save and Export Notebook: PDF" via the command palette.

In JupyterLab or Jupyter Notebook

• Install this extension in your environment and open JupyterLab/Jupyter Notebook.
• Open a notebook, then pick "File" > "Save and Export Notebook As" > "PDF (via jupyterlite-pdf-exporter)".
  • You may also run "Save and Export Notebook: PDF (via jupyterlite-pdf-exporter)" via the command palette.

Customising the PDFs

It is also possible to change the look and feel of the exported PDF(s) through the settings.

To change the settings for yourself, open the "Settings Editor" from the "Settings" menu in JupyterLite, JupyterLab, or Notebook 7, and pick "JupyterLite PDF Exporter". The form lets you set the page size, font, margins, and more. The new settings apply the next time you export a notebook. Choosing the settings per-notebook is currently not implemented.

• If you build a JupyterLite site for other people, you can set the defaults for everyone. For JupyterLite, you may add an entry for jupyterlite-pdf-exporter:plugin to an overrides.json file in your build, or to the settingsOverrides field in your jupyter-lite.json file.
• On JupyterLab and Jupyter Notebook, add the same jupyterlite-pdf-exporter:plugin entry to an overrides.json in your app settings directory. Those values become the starting point for every user, who can still change them in their own Settings Editor.

Only the fonts that come bundled with the Typst compiler are available.

Here is a table of what each setting changes. The "Key" column is the name you write in overrides.json or in settingsOverrides. The "Pandoc mapping" column shows where the value ends up. A "variable" is passed as a Pandoc template variable and an "option" is a top-level Pandoc option. You do not need to know these to use the settings, but they are handy if you want to read the Pandoc and Typst docs.

The margin key takes an object with any of top, bottom, left, and right, for example { "top": "2.5cm", "bottom": "2.5cm", "left": "2cm", "right": "2cm" }.

For linkColor, the Settings Editor shows a colour picker for ease of use. In overrides.json or jupyter-lite.json, you can use a HEX value such as #0F4C81 (with or without the #). An invalid value fails the PDF export.

An example deployment configuration

Each key goes under the plugin ID jupyterlite-pdf-exporter:plugin. Here is a full overrides.json that you can drop into your JupyterLite build folder:

{
  "jupyterlite-pdf-exporter:plugin": {
    "pageSize": "us-letter",
    "mainFont": "New Computer Modern",
    "pageNumbers": false,
    "tableOfContents": true,
    "numberSections": true,
    "lineSpacing": 1.15,
    "linkColor": "#0F4C81",
    "margin": {
      "top": "3cm",
      "bottom": "3cm",
      "left": "2.5cm",
      "right": "2.5cm"
    }
  }
}

The same values can instead live under settingsOverrides in your jupyter-lite.json as well:

{
  "jupyter-lite-schema-version": 0,
  "jupyter-config-data": {
    "settingsOverrides": {
      "jupyterlite-pdf-exporter:plugin": {
        "pageSize": "us-letter",
        "pageNumbers": false
      }
    }
  }
}

For further reference, navigate to the following resources:

1. JupyterLite documentation on configuration files
2. JupyterLite documentation on settings overrides

Development

See CONTRIBUTING.md for the full setup. To try your local changes in JupyterLab or Jupyter Notebook, build the extension and link it into your environment:

jlpm install
jlpm build
jupyter-builder develop --overwrite .

Then start JupyterLab with jupyter lab or Jupyter Notebook with jupyter notebook. After a code change, run jlpm build again and refresh the browser.

To try your local changes in JupyterLite, build the extension and link it into your JupyterLite build:

jlpm install
jlpm build
jupyter-builder develop --overwrite .

Then rebuild your JupyterLite distribution and open it in the browser via jupyter lite build and jupyter lite serve (or your usual JupyterLite build and serve commands). After a code change, run jlpm build again, rebuild your JupyterLite distribution, and refresh the browser.

License

The source code is licensed under the terms of the BSD-3-Clause "New" or "Revised" License (BSD-3-Clause; see the LICENSE file for details).

[!IMPORTANT] However, the source distribution and wheel for this extension on PyPI are licensed under the terms of the GNU General Public License version 2.0 (GPL-2.0) or later (GPL-2.0-or-later). Please see the Pandoc license file for details.

The WebAssembly/JavaScript distribution of Typst, @myriaddreamin/typst-all-in-one, is licensed under the terms of the Apache License 2.0 (Apache-2.0). Please see the Typst license file for details.

Why?

The WebAssembly distribution of Pandoc, through its dependency on the pandoc-wasm project on the npm package registry, is licensed under the terms of the GNU General Public License version 2.0 (GPL-2.0-or-later). Binary distributions of this extension bundle the pandoc.wasm file, and as a result, are regarded as derivative works of the WebAssembly distribution of Pandoc.

More details

For an overview of the licenses of all the JavaScript dependencies of this extension at runtime, please navigate to your JupyterLite deployment > "Help" menu > "Licenses" after installing and rebuilding it.

Thanks 💛

This project would not have been possible without the following open source projects:

• JupyterLite: A JupyterLab distribution that runs entirely in the web browser, powered by WebAssembly and Web Workers.
• JupyterLab: The next-generation web-based user interface for Project Jupyter, with a rich ecosystem of extensions.
• Jupyter Notebook: The original web-based interactive computing environment for Jupyter, which continues to be widely used and developed in its own right.
• Pandoc: A universal document converter that supports a wide variety of input and output formats, including Jupyter notebooks and PDF.
• Typst: A modern typesetting system that provides high-quality PDF output and a user-friendly syntax for document design.
• pandoc-wasm: A WebAssembly distribution of Pandoc that allows it to run in web browsers and other JavaScript environments.
• @myriaddreamin/typst-all-in-one: A WebAssembly distribution of Typst that allows it to run in web browsers and other JavaScript environments.