Create Zenodo JSON file

[epassaro]

Last update: 13/2/2020

• The pipeline consist of two jobs: one creates .zenodo.json file by running a notebook, and the other upload a new GitHub release. The Zenodo webhook automatically makes a release in their site using .zenodo.json from master as author list.

• At the moment the second job is commented and not executed. Once un-commented, the second job will run only if the first job is succesful (see the parameter: dependsOn: zenodo_json).

• The first job is incomplete. To complete we need:

  a) Fetch the gather_data.ipynb notebook from tardis_zenodo private repo. We are using a copy of the notebook stored in Gist and downloaded via wget as a temporary workaround.

  b) Push the generated .zenodo.json to the master branch.

• During the first job the gather_data.ipynb notebook is executed twice:

  1. The first run is with --allow-errors flag and renders the notebook with errors/exceptions. The rendered notebook is saved as an Azure Artifact. This step usually never fails and is useful to debug errors very quickly.

  2. The second run is done without the flag. If there's any error/exception the entire pipeline will fail.