Repo2docker: Add support for DRYAD data repository

[nuest]

[Sorry if my issues sound like a broken record! I hope more contentproviders will open up BinderHub for more communities.]

Proposed change

Support DRYAD (or stash) as a content provider (DOI + URL-based): https://datadryad.org/

Their guide to data sharing explicitly mentions code + data, see https://datadryad.org/docs/QuickstartGuideToDataSharing.pdf, so _they should be a good fit_.

_The biggest problem right now_ is that I did not yet find a suitable repository (chicken and egg problem?!), and DRYAD is not for free unless you are affiliated with specific institutions.

Anyone here who can upload a test binder to DRYAD?

Relevant API documentation:

• https://datadryad.org/api/v2/docs/#/default/get_datasets__doi_
• https://datadryad.org/api/v2/docs/#/default/get_datasets__doi__download (zipped and direct file download possible, didn't see any limitation on ZIP archive size)

Alternative options

None? :fist_raised:

Who would use this feature?

DRYAD is widely known in some research communities, e.g. ecology, so any researcher publishing a dataset or workflow there would be familiar with it.

How much effort will adding it take?

1 day, with the current examples of contentproviders as a boilerplate.

Who can do this work?

A developer familiar with the content provider implementation.

[nuest]

Ping @sfisher - would this be something you're interested in?

[dlowenberg]

Hi all,

I am the Dryad product manager and wanted to update you with new documentation. Our API is at: https://datadryad.org/api/v2/docs. It should not look different from the stage. We are very interested in building this integration with you, we just need to schedule this into our priorities. If you could email me we could either set up a quick moment to chat and plan this.

Best,
Daniella Lowenberg ([email protected])

[nuest]

_Thanks for reaching out!_ I'll contact you via email to set a time for chat, we can report results here afterwards.

[API links fixed - thanks!]

[choldgraf]

hey @nuest and @dlowenberg - let me know if I can help move this partnership forward as well

[nuest]

Thanks @dlowenberg for the conversation and some clarifications about DRYAD. Publishing code on DRYAD is not that common, but maybe (IM0) an integration with Binder can push that a little bit.

There are no special requirements - the DRYAD API offers everything we need. A very helpful pointer by @dlowenberg is to the staging server, where one can easily create a repo (for free, but ephemeral) for testing:

• https://dryad-stg.cdlib.org/stash/sessions/choose_login
• https://dryad-stg.cdlib.org/api/v2

NOTE: For Login the _sandbox_ ORCID server is used!

The only thing holding this back is developer time. So if anyone is interested @choldgraf and @nuest would be happy to get someone started or review a PR.

[nuest]

@betatim @minrk What do you think, does this deserve a "good first issue" label, since the existing content providers are quite a good guideline?

[betatim]

No idea if this would make a good first issue as I had never heard of Dryad before.

Do you need to be logged in to retrieve a repository as well or only to create one?

[sfisher]

Here is a little more of a walkthrough of submission (I think it's linked from the documentation, also). https://github.com/CDL-Dryad/dryad/blob/master/documentation/api_submission.md

You will need an Application ID and Secret to enable submission by API. I can certainly create one if you want to test on stage (please login to the UI first, which uses the ORCID sandbox for our stage server as mentioned above).

[nuest]

We don't need API-based submission. We can create a test repo by hand. repo2docker only needs read access.

@sfisher you do not need to login for downloading a record, right?

@betatim: I was thinking "good first" because the other content providers give good guidance what to do.

[dlowenberg]

No login necessary for downloading