sword/native api mismatch for data deposit

[raprasad]

(may be a duplicate)

The SWORD data deposit API does not require contact email or subject.

/dvn/api/data-deposit/v1/swordv2/collection/dataverse

However, the GUI does require these fields.

[pdurbin]

All these fields are discussed in "Dataverse 4.0 Data Deposit API Backward Compatibility and Migration Guide for Collaborators" at https://docs.google.com/document/d/11DpdKyp1tagmaJAAzRqQBEZEZ69WOOOYsoqhz8UsfNM/edit?usp=sharing

I'll move this ticket to QA when I'm finished testing.

[pdurbin]

@scolapasta I need help with this. If you look at 19ae153 I'm obviously doing something wrong with controlled vocabulary because "Other" appears twice:

[pdurbin]

We don't have a way to express "Contact E-mail" or "Subject" in dcterms so we now populate the former with the email address of the parent dataverse and the later with "Other". (The longer term plan in #769 is to have datasets created by SWORD inherit the subject of the parent dataverse.)

Centralized validation is being worked on in #605 but for now some quick and dirty validation is in place. If you try to create a dataset without any dcterms:creator elements, for example, you should get the error "Can't create/update dataset. http://purl.org/dc/terms/ equivalent of required field not found: authorName".

Moving to QA (though I haven't looked at the native API at all... maybe that should be a separate ticket).

[esotiri]

when creating dataset via sword, required fields email and subject are auto completed with creator email and subject other.

[esotiri]

issue resolved

[pdurbin]

when creating dataset via sword, required fields email and subject are auto completed with creator email and subject other.

@esotiri can you please also test other required fields? In some cases, the required field is not auto-completed. Rather, you should expect to see an error indicating that the field is required.

In addition, there's some discussion at #457 (comment) with @posixeleni over create vs. edit.

These are the fields that were possible in DVN 3.6: https://github.com/IQSS/dvn/blob/develop/doc/sphinx/source/dataverse-api-main.rst#data-deposit-api

We can't use dcterms:rights yet, per #805.

[esotiri]

create a study-missmatch:
Testing two other required fields, author and description, commenting them out from the atom-entry.

Below is the result of sword error: equivalent of required field not found

<?xml version="1.0" encoding="ISO-8859-1"?>
<sword:error xmlns:sword="http://purl.org/net/sword/terms/" href="http://purl.org/net/sword/error/ErrorBadRequest">
  <atom:title xmlns:atom="http://www.w3.org/2005/Atom">ERROR</atom:title>
  <atom:updated xmlns:atom="http://www.w3.org/2005/Atom">2014-10-29T10:57:32Z</atom:updated>
  <atom:generator xmlns:atom="http://www.w3.org/2005/Atom" uri="http://www.swordapp.org/" version="2.0"/>
  <sword:treatment>Processing failed</sword:treatment>
  <atom:summary xmlns:atom="http://www.w3.org/2005/Atom">Can't create/update dataset. http://purl.org/dc/terms/ equivalent of required field not found: dsDescription</atom:summary>
</sword:error>

 <sword:treatment>Processing failed</sword:treatment>
  <atom:summary xmlns:atom="http://www.w3.org/2005/Atom">Can't create/update dataset. http://purl.org/dc/terms/ equivalent of required field not found: authorName</atom:summary>
</sword:error>

• Note: It would be good that the actual dcterms xml element is not displayed at the failed message. Right now you can look the mapping at foreignmetadatamapping.
  I will open a new ticket for this.