Add an API endpoint for extended File Level metadata

[raprasad]

5/4/2018 update

Add an API endpoint where File Level metadata can be retrieved using a File ID or File DOI:

• API input: Datafile id or Datafile DOI, optional Dataverse token
• output: Datafile metadata including:
  • Owning dataset DOI
  • File-specific information already included in the native API for datasets
  • JSON-LD Metadata for the file's dataset
  • Any other relevant info

---

Notes:

1. My specific use case involves knowing the Datafile ID and wanting to retrieve the JSON-LD citation for that file's dataset. Currently this is done in 3 steps:
   1. Use the search API to retrieve the file's dataset citation
   2. Parse the text citation for the dataset DOI
   3. Call the existing JSON-LD endpoint using the dataset DOI from above
2. community benefit: If I am understanding Dataverse External Tools correctly, this will help other developers who could use the Datafile ID/DOI to get more display information for users and/or access other existing APIs. * http://guides.dataverse.org/en/4.8.5/installation/external-tools.html
3. google group: https://groups.google.com/forum/?utm_source=digest&utm_medium=email#!topic/dataverse-community/u6VjVQ0h8RU

[scolapasta]

We also want to add an API for editing the metadata of a file, so we'll use this issue to track both edit and view of a file's metadata.

[scolapasta]

It was pointed out you may be able to edit the md5 of a file with the existing dataset APIs. If so, this is a bug, and we should also fix that as a result of adding these new file level APIs (since they should use the same code fragments, ideally.

[matthew-a-dunlap]

@scolapasta By editing file metadata do you mean updating? In other words the ability to upload json that overrides the current metadata?

[scolapasta]

@matthew-a-dunlap yes, that's what I mean.

[matthew-a-dunlap]

Should the new GET endpoint allow access to any arbitrary version of the metadata, or should it just return the latest?

[pdurbin]

@matthew-a-dunlap we don't want users to be able to change the MD5 of a file, for example.

Please see the "Create a dataset with files and directoryLabels using json" test here for the existing API endpoint:

https://github.com/IQSS/dataverse/blob/v4.11/src/test/java/edu/harvard/iq/dataverse/api/FileMetadataIT.java#L115

[matthew-a-dunlap]

The new requirements for this story:

EDIT: we decided instead upon a new story #5674

[pdurbin]

@raprasad hi! Now that pull request #5529 has been merged, the first item on your list is easier:

• Owning dataset DOI

curl -s https://dataverse.harvard.edu/api/search?q=entityId:3371438 | jq -r '.data.items[0].dataset_persistent_id'

doi:10.7910/DVN/TJCLKP

How much more of this issue are you still interested in?

[djbrooke]

I'll close this one out. We can implement other parts of this as specific issues as needed as we build out additional external tools.

[pdurbin]

Related:

• Add info about owning entities to get file api (and get Dataset API) #9598