buttons linking directly to downloadable OSF and Zenodo dataset versions in CONP portal

[emmetaobrien]

Our links to OSF in the CONP portal currently point at the overview pages for each dataset: https://osf.io/<example>/overview

OSF are fine with us linking directly to the archived zip files for each dataset, at the address
https://files.us.osf.io/v1/resources/znt9d/providers/osfstorage/?zip=<example>, which has worked fine in my tests and could be automatically generated.

Zenodo are also good with direct links to their data, and links to the relevant files are stored in the files->links->self entry in the response to an API call to https://zenodo.org/api/records/<example>. In this case the structure of the link is more complex and does not look to be automatically derivable from <example> so we would need to retrieve this for every Zenodo dataset.

If we proceed with using these values as the default download links for OSF and Zenodo datasets in the CONP Portal interface, do we still want to store the overview page links in DATS.json as well? It seems to me that the options here are a) replace all the existing links, or b) add a new field storing the zip file download link (or dragoon some other part of the existing DATS format into that purpose).

[emmetaobrien]

Update after CONP dev meeting: we definitely do want to store this information in the DATS file, as the Zenodo model above is more complex than we want to generate in realtime in the CONP portal front-end. We will also want to keep the existing value that is displayed in the portal as Project Landing Page. So I am currently looking through the DATS schema for the best place to store this information, and anticipate using OSF as the first test case here as it is simpler.

[emmetaobrien]

On going through the DATS schema, I think adding another entry under distributions for the direct download link may be a workable way to store this information, but I'm not yet sure it is the best way.

[emmetaobrien]

Looking at this again there is, I think, a better way:

distributions->access->landingPage continues to point at our existing landing page
distributions->access->accessURL points to the direct link.

Note that datasets retrieved by the OSF crawler are stored in the conp-bot repository under conp-bot/conp-dataset-<dataset_name> so there is (intentionally) no simple manual way of editing the DATS.json file in existing OSF-derived datasets.