tsv metadata block import: add preview mode

[pdurbin]

While working on #2310 I'm realizing it would be very handy to have a "preview" mode for loading tsv files as metadata blocks but not actually writing to the database. It would be a way to sanity check a tsv file. @sekmiller seems to like this idea as well.

Locally, I'm playing around with a "preview=true" query parameter that looks like this:

curl -s http://localhost:8080/api/admin/datasetfield/load?preview=true -X POST --data-binary @scripts/api/data/metadatablocks/customGSD.tsv -H "Content-type: text/tab-separated-values"

@scolapasta I'm not sure I should change this in 4.2 though. We can probably defer it so I'll push any change to a branch that we can merge in later.

(@bencomp I'm aware from #2357 that you're not a fan of the tsv import.) :)

[pdurbin]

In 948ff41 I have a proposal in a branch for adding the "preview" functionality I'm interested in.

(Mental note that http://localhost:8080/api/metadatablocks is handy way to see which blocks have been loaded.)

[pdurbin]

@sekmiller as we discussed, since you're looking into adding the ability to replace controlled vocabulary values per #2310 (comment) you might want to at least think about my proposal at 948ff41 to add a "preview" mode to the API endpoint for importing tsv files. I'll give this to you and move it to 4.2 but obviously you should feel free to move it out of 4.2 at any point.

This is how I was testing the preview mode I added in that branch (adding ?preview=true as a query parameter):

curl -s http://localhost:8080/api/admin/datasetfield/load?preview=true -X POST --data-binary @scripts/api/data/metadatablocks/customGSD.tsv -H "Content-type: text/tab-separated-values"

[pdurbin]

In a thread called "documentation for metadata block import format" @tdilauro started a Google doc where he's trying to document what he's figured out about how to make new metadata block: https://groups.google.com/d/msg/dataverse-community/VniGWZKlgiA/lKxiYO4iBQAJ

This seems on topic for this issue since before you can import a tsv file you've created (in "preview" mode or otherwise) you have to know enough about about how Dataverse processes the tsv file for a successful import. Basically, you need docs.

On a related note, the experiment by @raprasad to convert tsv files to JSON Schema at https://github.com/IQSS/json-schema-test is very interesting.

[pdurbin]

Does anyone else think this is a good idea? If so, please get in touch. Closing.

[pdurbin]

This new feature request sounds really useful:

• Feature Request/Idea: Delete a metadata block (that hasn't been used at all) #9628