This action automatically uploads GitHub repository content to a Dataverse dataset. It can upload the entire repository or its subdirectories into an existing dataset on a target Dataverse installation. The action is customizable, allowing you to fully replace a dataset, add to the dataset, publish it or leave it as a draft version on Dataverse.

The action provides some additional metadata to the dataset, such as the origin GitHub repository, and it preserves the directory tree structure.

To use this action, you will need the following input parameters:

To use the action, create a new YML file (i.e., workflow.yml) and place it in the directory .github/workflows/ in your GitHub repository.

The action workflow can be executed at trigger events such as push and release. If you'd only like to run the workflow manually from the Actions tab, use the workflow_dispatch event option.

Here is an example of a workflow.yml that actives the action on release and manually.

on: 
  release:
  workflow_dispatch:

jobs:
  build:
    runs-on: ubuntu-latest
    steps:
      - name: Send repo to Dataverse 
        uses: IQSS/[email protected]
        with:
          DATAVERSE_TOKEN: ${{secrets.DATAVERSE_TOKEN}}
          DATAVERSE_SERVER: https://demo.dataverse.org
          DATAVERSE_DATASET_DOI: doi:10.70122/FK2/LVUA

If you'd like to upload files from a specific subdirectory only (for instance, a data folder), you should add the GITHUB_DIR argument in your workflow, as follows:

steps:
  - name: Send repo to Dataverse 
    uses: IQSS/[email protected]
    with:
      DATAVERSE_TOKEN: ${{secrets.DATAVERSE_TOKEN}}
      DATAVERSE_SERVER: https://demo.dataverse.org
      DATAVERSE_DATASET_DOI: doi:10.70122/FK2/LVUA
      GITHUB_DIR: data

You can upload multiple subdirectories by listing them as GITHUB_DIR like GITHUB_DIR: data,plots, or in the following format:

steps:
  - name: Send repo to Dataverse 
    uses: IQSS/[email protected]
    with:
      DATAVERSE_TOKEN: ${{secrets.DATAVERSE_TOKEN}}
      DATAVERSE_SERVER: https://demo.dataverse.org
      DATAVERSE_DATASET_DOI: doi:10.70122/FK2/LVUA
      GITHUB_DIR: |
        data
        plots 

By default, the action will sync the GitHub repository and the Dataverse dataset, meaning that it will delete the Dataverse content before uploading the content from GitHub. If you don't want the action to delete your dataset before upload (i.e., if you already have a Dataverse DRAFT dataset), set the DELETE argument to False like:

steps:
  - name: Send repo to Dataverse 
    uses: IQSS/[email protected]
    with:
      DATAVERSE_TOKEN: ${{secrets.DATAVERSE_TOKEN}}
      DATAVERSE_SERVER: https://demo.dataverse.org
      DATAVERSE_DATASET_DOI: doi:10.70122/FK2/LVUA
      GITHUB_DIR: data
      DELETE: False

The action automatically uploads new content to a Dataverse dataset, but it will not publish it as a new version by default. If you'd like the action to publish a new dataset version in Dataverse, set the PUBLISH argument to True.

steps:
  - name: Send repo to Dataverse 
    uses: IQSS/[email protected]
    with:
      DATAVERSE_TOKEN: ${{secrets.DATAVERSE_TOKEN}}
      DATAVERSE_SERVER: https://demo.dataverse.org
      DATAVERSE_DATASET_DOI: doi:10.70122/FK2/LVUA
      GITHUB_DIR: data
      DELETE: False
      PUBLISH: True

1. If you change the content of your GitHub repository, are the changes synchronized in Dataverse? Otherwise, is it possible to synchronize them automatically?

Yes, the action is able to automatically update the Dataverse dataset. In other words, if the action is triggered with every push to the GitHub repository, it will automatically upload its content to Dataverse. You specify the action triggers in the workflow (.yml) file, and in this case, it would contain on: push line to execute the action on every push to the repository.

2. Will the action work with dataset Handles as persistent identifiers (PIDs) instead of DOIs?

Yes, the action uses Dataverse API that supports both DOIs and Handles for retrieving and uploading data.

3. How do I contribute to this project?

Have a look at the instructions in the CONTRIBUTING.md file.

Check out the following related projects:

Visit this page to create a Dataverse DOI Badge for your GitHub repository.

Looking for a stand-alone Dataverse uploader that will work from the command line? Check out DVUploader.

See the official Dataverse documentation here.

Don't hesitate to create an issue, a pull request, or contact us if you notice any problems with the action.