Automatically create Location entry in Thoth when a Dissemination workflow succeeds

[rhigman]

On successful dissemination, add a new Location entry to the relevant Publication in Thoth, recording the URL(s) of the newly-created directory entry and/or copy of the content.

To be added to the existing Internet Archive/Figshare workflows (Crossref is not relevant as DOIs are already present in Thoth at time of submission). For Internet Archive, the Platform type INTERNET_ARCHIVE is now available; for institutional repositories such as Figshare, OTHER will need to be used.

There would then be potential for replacing the current logic for checking whether or not a Work already exists in the target platform, instead looking at whether a Location with the relevant Platform exists.

This can then be extended to new dissemination platforms when they are implemented. Ease of implementation may depend on individual platforms' workflows; Internet Archive returns the relevant URL to the dissemination script immediately on creation, but e.g. FTP-based workflows are unlikely to be as neat.

[rhigman]

As part of this work, add something like PUBLISHER_WEBSITE to the set of location platform types.

[ja573]

As part of this work, add something like PUBLISHER_WEBSITE to the set of location platform types.

Tracked separately now under #561

[rhigman]

• For each platform, determine the appropriate landingPage and fullTextUrl to record in Thoth
  • Internet Archive: landingPage = archive.org/details/[workId], fullTextUrl = archive.org/details/[workId]/[filename].pdf
  • Figshare: note figshare.com (API) vs repository.lboro.ac.uk (UI) versions of same links, Handles, etc; treat as a "Figshare" upload or a "Loughborough repository" upload (repositories may migrate platforms)?
  • CUL: tbd
  • Zenodo (or do under Disseminate to Zenodo #542)
  • OAPEN: works don't acquire these URLs until some hours/days after dissemination. Currently handled manually. Any alternative? Split out as separate task?
  • (Crossref: not relevant here)
• Extend disseminator to (retrieve and) pass back landingPage and fullTextUrl when they have been assigned via successful archiving
  • this will need to be on a per-publication basis as we sometimes disseminate more than one format, so publicationId will also need to be passed back, or at least publicationType
• Add script which takes publicationId, locationPlatform, landingPage and fullTextUrl and writes location to Thoth
  • locationPlatform could be supplied directly or derived from inputs to disseminator
  • publicationId could be passed back directly as above, or obtained from Thoth via e.g. workId + publicationType
• Extend GitHub Actions to take output from each disseminator run and pass it to new script
  • for dissemination of multiple formats, should the script be called multiple times, or should it handle multiple locations itself?
• Determine whether any new locationPlatforms need to be added to Thoth
  • e.g. FIGSHARE - or as above, should it be e.g. LBORO_REPO?
  • Any way of marking/"locking" these locations as created by Thoth Dissemination Service/part of Thoth Archiving Network?
  • Is it still appropriate to permit only one location per locationPlatform for all of these? (e.g. users might independently upload copies to additional Figshare repositories, etc - not necessarily sensible but shouldn't go unrecorded)
• Catchup run: ensure that works disseminated prior to implementation of this feature all have appropriate locations created
  • Could a similar mechanism be used (on a regular, automatic basis) to handle OAPEN, as above?
• Add an appropriate set of Thoth credentials as repository secret (or organisation secret - would require permissions)