Docusaurus Prince PDF Generator

Extract rendered data from Docusaurus and generate PDF, the hard way

Demo/Examples

You can download it in GitHub Actions artifacts section to see the result.

This project is using the method 1 (see below) for generating PDF. You must have Prince installed on your local machine.

Usage

Install Prince first.

Run the following commands to generate PDF:

# Genrate PDF from specific site under `docs` scope
npx docusaurus-prince-pdf -u https://docusaurus.io/docs

# Change generating scope to `/docs/cli/`
npx docusaurus-prince-pdf -u https://docusaurus.io/docs/cli

# Custom working (output) directory
npx docusaurus-prince-pdf -u https://openbayes.com/docs --dest ./pdf-output

# Custom output file name
npx docusaurus-prince-pdf -u https://openbayes.com/docs --output docs.pdf

To generate PDF from a local Docusaurus instance. You need to first build the site locally:

# Build the site
(npm|yarn|pnpm) build

# Serve built site locally
(npm|yarn|pnpm) serve

# Generate PDF from local Docusaurus instance
npx docusaurus-prince-pdf -u http://localhost:4000/docs # Change port to your serving port

See help screen for all options:

npx docusaurus-prince-pdf -h

Docker (Recommended)

• Docker Hub
• ghcr.io

You can run this program with Docker image:

docker run --rm -it --init \
  -v $(pwd)/pdf:/app/pdf \
  openbayes/docusaurus-prince-pdf \
  -u https://docusaurus.io/docs/

If you need Asiatic languages support like Chinese and Japanese. You can mount your custom fonts directory to Docker image:

docker run --rm -it --init \
  -v $(pwd)/pdf:/app/pdf \
  -v $(pwd)/fonts:/root/.fonts \
  openbayes/docusaurus-prince-pdf \
  -u https://docusaurus.io/docs/

GitHub Actions

You can also run this program inside GitHub Actions:

jobs:
  build:
    # prerequisites...

    - name: Install Prince
      run: |
        curl https://www.princexml.com/download/prince-14.2-linux-generic-x86_64.tar.gz -O
        tar zxf prince-14.2-linux-generic-x86_64.tar.gz
        cd prince-14.2-linux-generic-x86_64
        yes "" | sudo ./install.sh

    - name: Build PDF
      run: npx docusaurus-prince-pdf -u https://docusaurus.io/docs/

    - name: Upload results
      uses: actions/upload-artifact@v3
      with:
        name: result
        # The output filename can be specified with --output option
        path: pdf/docusaurus.io-docs.pdf
        if-no-files-found: error

    # ...other steps

You can also run prince with prebuilt Prince Docker image:

jobs:
  build:
    # prerequisites...

    - name: Build PDF
      run: docker run --rm -it -v $(pwd)/pdf:/app/pdf openbayes/docusaurus-prince-pdf -u https://docusaurus.io/docs/

    # ...other steps

How it works

Like mr-pdf, this package looks for the next pagination links on generated Docusaurus site. Collect them in a list and then pass the list to Prince to generate the PDF.

You can specify the CSS selector if you're using custom Docusaurus theme:

npx docusaurus-prince-pdf -u https://openbayes.com/ --selector 'nav.custom-pagination-item--next > a'

Why this package?

I made a comparison list for the two methods of generating PDF from Docusaurus.

Method 1: Prince

The good:

• Best font subsetting support
• Text can be selected and copy/paste correctly
• Fancy Table of Contents

The bad:

• Watermark on first page of generated PDF make it hard to handle in CI/CD environments
• Doesn't work with some CSS syntax (e.g. mask-image)
• Doesn't work with some HTML features (e.g. srcset)
• Commercial license is expensive ($3,800)

The ugly:

• None

Method 2: mr-pdf (not used in this project)

The good:

• Free and open-source
• Works with Docusaurus sites
• CI/CD friendly
• Based on Puppeteer make it works for most modern CSS syntax (e.g. mask-image)

The bad:

• Doesn't work well with system Dark Mode. You will get a dark background in generated PDF when you have respectPrefersColorScheme enabled in your Docusaurus instance. But it's not an issue in Ci/CD environments
• No Table of Contents

The ugly:

• Based on Puppeteer make the text cannot be copied or searched correctly
• Link anchors (links start with #) not well handled

Usage:

npx mr-pdf --initialDocURLs="https://openbayes.com/docs/" --paginationSelector=".pagination-nav__item--next > a" --contentSelector="article"

License

MIT