Using pandoc with GitHub Actions

You can use pandoc, the universal markup converter, on GitHub Actions to convert documents.

GitHub Actions is an Infrastructure as a Service (IaaS) from GitHub, that allows you to automatically run code on GitHub's servers on every push (or a bunch of other GitHub events). For example, you can use GitHub Actions to convert some file.md to file.pdf (via LaTeX) and upload the results to a web host.

Using docker://pandoc Images Directly

You can now directly reference container actions on GitHub Actions. You do not need a separate GitHub Action.

If you need LaTeX (because you want to convert through to PDF), you should use the docker://pandoc/latex image. Otherwise, the smaller docker://pandoc/core will suffice.

It is a good idea to be explicit about the pandoc version you require, such as docker://pandoc/core:2.9. This way, any future breaking changes in pandoc will not affect your workflow. You can find out whatever the latest released docker image is on docker hub. You should avoid specifying no tag or the latest tag -- these will float to the latest image and will expose your workflow to potentially breaking changes.

Simple Usage

You can use pandoc inside GitHub actions exactly as you would use it on the command line. The string passed to args gets appended to the pandoc command.

The below example is equivalent to running pandoc --help.

You can see it in action here.

name: Simple Usage

on: push

jobs:
  convert_via_pandoc:
    runs-on: ubuntu-22.04
    steps:
      - uses: docker://pandoc/core:2.9
        with:
          args: "--help" # gets appended to pandoc command

Long Pandoc Calls

Remember that as per the GitHub Actions workflow syntax, "an array of strings is not supported by the jobs.<job_id>.steps.with.args parameter. Pandoc commands can sometimes get quite long and unwieldy, but you must pass them as a single string. If you want to break up the string over several lines, you can use YAML's block chomping indicator:

name: Long Usage

on: push

jobs:
  convert_via_pandoc:
    runs-on: ubuntu-22.04
    steps:
      - run: echo "foo" > input.txt  # create an example file
      - uses: docker://pandoc/core:2.9
        with:
          args: >-  # allows you to break string into multiple lines
            --standalone
            --output=index.html
            input.txt

You can see it in action here.

Advanced Usage

You can also:

• create an output directory to compile into; makes it easier to deploy outputs.
• upload the output directory to GitHub's artifact storage; you can quickly download the results from your GitHub Actions tab in your repo.

Remember that wildcard substitution (say, pandoc *.md) or other shell features frequently used with pandoc do not work inside GitHub Actions yaml files args: fields. Only GitHub Actions context and expression syntax can be used here. If you want to make use of such shell features, you have to run that in a separate step in a run field and store the result in the GitHub actions context. The below workflow includes an example of how to do this to concatenate several input files.

You can see it in action (haha!) here.

name: Advanced Usage

on: push

jobs:
  convert_via_pandoc:
    runs-on: ubuntu-22.04
    steps:
      - uses: actions/checkout@v3

      - name: create file list
        id: files_list
        run: |
          echo "Lorem ipsum" > lorem_1.md  # create two example files
          echo "dolor sit amet" > lorem_2.md
          mkdir output  # create output dir
          # this will also include README.md
          echo "files=$(printf '"%s" ' *.md)" > $GITHUB_OUTPUT

      - uses: docker://pandoc/latex:2.9
        with:
          args: --output=output/result.pdf ${{ steps.files_list.outputs.files }}
          
      - uses: actions/upload-artifact@v3
        with:
          name: output
          path: output

Alternatives

If you need to have Pandoc installed and available globally, for example because it is being used in a subprocess by a library or application, you can use one of the two following alternative actions:

• pandoc/actions/setup (Linux and MacOS runners support)
• r-lib/setup-pandoc (Linux, MacOS and Windows runners support)