Manual Provenance Integration for Files

[djbrooke]

File and dataset level, we'll need a few things in order to populate the Provenance system with relevant information from Dataverse so that provenance graphs can be displayed (#3245).

At the file level we need to:

• Allow a User to upload a provenance file and/or allow a user to add text explaining the transformation
• Call “prov.setversiontransformation” when free text is added in a draft state
• Call “prov.(TBD)versiontransformation” when a file is added in a draft state
• Call ? to import Provenance when a file is published

[djbrooke]

We'll get a demo and code walkthrough from @jacksonokuhn next Friday and learn more about what we should expect.

[djbrooke]

Mockups are set, but we need information from the Provenance team about what will be displayed under the tab and what error conditions to expect.

[sekmiller]

On the provenance tab we allow the user to upload a provenance file. We have to add a process for ingesting these files. It's my understanding that they will be saved to the provenance file directory. Phil suggested that this might need to be configurable.

[pdurbin]

Yeah, I was suggested that for now the files could be put in 10.5072/FK2/ZPWUDQ/provenance or whatever. I'm sure people like @pameyer and @bmckinney will be very glad to hear that the eventual plan is to put the provenance files on a separate file system because they definitely don't want these files intermixed with data they're trying to do computation on! But yes, a sane default for the location and a configuration option would be welcome, I can only imagine.