Fwd: Desired features for v1 of the DVN Data Deposit API

55 views
Skip to first unread message

Philip Durbin

unread,
Aug 7, 2013, 5:09:17 PM8/7/13
to dataverse...@googlegroups.com
Hello Dataverse community!

Back in May I mentioned* we were planning to add a new "Data Deposit
API" to DVN that implements the SWORDv2 protocol.

Well, now I'm the one working on it. :)

With lots of help from the rest of the team, of course.

So far so good... I'm adding features one by one... obvious features
such as the ability to create a study and deposit data as well as some
others. Please see the mini status report below I sent to the
dataverse team this morning for details.

The first use case we anticipate for this new API is the deposit of
data via a plugin (also under development) for Open Journal Systems
(OJS) so we want version 1 of the Data Deposit API to support
everything they need at the very least. You can read all about this
collaboration at http://projects.iq.harvard.edu/ojs-dvn

But what do YOU want to use the DVN Data Deposit API for?

Phil

p.s. The DVN Data Deposit API is still very much under development so
some implementation details you see below may change! We'll publish
official docs once the API has stabilized.

p.p.s. I did already fix one of the bugs I mentioned in the mini
status report below. :)

* https://groups.google.com/d/msg/dataverse-community/07h0Ca-Ai1I/tPJSRp5Cnf8J

---------- Forwarded message ----------
From: Philip Durbin <philip...@harvard.edu>
Date: Wed, Aug 7, 2013 at 9:32 AM
Subject: Desired features for v1 of the DVN Data Deposit API
To: Dataverse Team

As I mentioned in a previous email, I added "v1" to all the URLs in
upcoming DVN's Data Deposit API.

My thinking is that v1 will be enough for OJS to write their plugin
but I'm not sure v1 will fully implement all of the SWORDv2 spec,
which is pretty extensive.

From a high level, I want to make sure we're all on the same page in
terms of desired features for v1 of DVN's Data Deposit API and I just
started a document at
http://devguide.thedata.org/features/api/data-deposit/v1 that lists
(with `curl` examples) what it's capable of right now:

- Retrieve SWORD service document
- Create a study
- Deposit data into a study
- List studies in a dataverse
- Delete a study
- Replacing cataloging information (title, author, etc.) for a study
- Replace all files on a study with a new zip file
- Release a study

(I should note that the implementation details of the above (i.e. PUT
vs. POST) are still subject to change and I'm still working through
some known bugs in the implementation.)

The list above is more extensive than the "Developers should be able
to use the Data Deposit API to..." list in the description of
https://redmine.hmdc.harvard.edu/issues/3108 but that ticket
description would be a great place for us to edit to make sure we're
all on the same page. Please feel free to go ahead and make edits to
that ticket description to list features you want for v1.

In addition to working on the bugs I mentioned, I'm planning on
implementing additional features such as:

- List files in a study
- Delete an individual file from a study
- Add an additional file to a study (not "replace all files")
- Handle metadata for files (category, description)

Also, I need to think through how to handle study versions. At the
moment, you do all of your edits to a study while it's in draft and
then you release it by sending "In-Progress: false" as and HTTP header
via POST*.

Phil

* Hat tip to Peter for pointing out this header to me:
http://irclog.iq.harvard.edu/dvn/2013-08-01#i_3005

p.s. Unless there are any objections, I was thinking I would forward
this to https://groups.google.com/group/dataverse-community so OJS and
others can see the current status. I'm especially interested in
feedback from people who are familiar with the SWORDv2 spec.

--
Philip Durbin
Software Developer for http://thedata.org
http://www.iq.harvard.edu/people/philip-durbin

Eleni Castro

unread,
Aug 8, 2013, 10:54:11 AM8/8/13
to dataverse...@googlegroups.com, philip...@harvard.edu
We should also include if possible a way to send OJS the Data Citation for the study via the Data Deposit API.

Similarly, how will Dataverse get the bibliographic citation of the article from OJS? 

Alex Garnett

unread,
Aug 8, 2013, 7:00:42 PM8/8/13
to dataverse...@googlegroups.com, philip...@harvard.edu
On Thursday, August 8, 2013 7:54:11 AM UTC-7, Eleni Castro wrote:
We should also include if possible a way to send OJS the Data Citation for the study via the Data Deposit API.

I think this was just addressed by Phil w/r/t exposing the Data Citation string, but correct me if I've misunderstood.
 
Similarly, how will Dataverse get the bibliographic citation of the article from OJS? 

Well, we definitely don't want to send a string here -- while there aren't many competing plaintext citation formats for datasets, there certainly are for articles, and OJS doesn't currently set a given journal's preferred citation style anywhere (this is also handled by the new XML functionality we're working on, but we can't develop against it just yet). It's reasonably possible to construct a BibTeX-style citation object from the Dublin Core that's currently exposed by OJS for a given article (small amount of regex hacking around dc:source but not too bad). That'd be our most direct recourse without changing any additional functionality, until the XML stuff is in place -- unless the Dataverse user community is such that we can just assume APA across the board (but I don't think we can do that).

Merce Crosas

unread,
Aug 8, 2013, 8:19:23 PM8/8/13
to dataverse...@googlegroups.com
Dataverse follows the data citation format proposed by Altman, King 2007:


which is compatible with the DataCite recommended format - also a good format to follow:



--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.



--
Mercè Crosas, Ph.D.
Director of Data Science, IQSS
Harvard University



Eleni Castro

unread,
Aug 9, 2013, 9:21:54 AM8/9/13
to dataverse...@googlegroups.com, philip...@harvard.edu
You're right Alex. Phil did address this in the latest email "Exposing Dataverse Fields for OJS to Properly Display Dataverse Data Citation (vice versa for Article Publication)"

Also w/r/t Publication (Journal Article) Citation field in the Cataloging Section of the Dataverse Network study: it is just a plain text (HTML supported) field where researchers can input whatever publication citation they wish. I'm assuming that since it was originally meant for social sciences that a lot of the publication citations may be in APA format but going forward it would imo be wrong to continue assuming that everyone (Medical, Sciences, etc) would be citing in that format.

Philip Durbin

unread,
Aug 9, 2013, 12:54:19 PM8/9/13
to Eleni Castro, dataverse...@googlegroups.com
On Fri, Aug 9, 2013 at 9:21 AM, Eleni Castro <posix...@gmail.com> wrote:
> You're right Alex. Phil did address this in the latest email "Exposing
> Dataverse Fields for OJS to Properly Display Dataverse Data Citation (vice
> versa for Article Publication)"

Yes, exactly, but after playing* with a SWORDv2 server implementation
written by the spec lead (DSpace), I think I'm going switch to
bibliographicCitation from http://purl.org/dc/terms/ to expose the
data citation of a DVN study.

Jen from OJS and I are discussing this at
http://irclog.iq.harvard.edu/dvn/2013-08-09

Everyone's welcome. There's a link at the top of that page. :)

Phil

* SWORDv2 GET examples from dash.harvard.edu (DSpace):
https://docs.google.com/document/d/1vgLGhsuF-5W_KQ_h0tCEvIBP2ah0L6NRckk698ZeGCY/edit?usp=sharing

Merce Crosas

unread,
Aug 12, 2013, 11:09:07 AM8/12/13
to dataverse...@googlegroups.com, Eleni Castro
Phil,

I thought of another feature to be supported by the API that will be important, at least in the second phase:

- the API should allow to link a study (either during in Review or after is Released) to another dataverse ; that is, when a study is created in a journal dataverse, the author might already have his/her own dataverse, so we want to offer the option to link that study to his/her dataverse.

Merce


--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Eleni Castro

unread,
Aug 12, 2013, 11:18:39 AM8/12/13
to Merce Crosas, dataverse...@googlegroups.com
I added this v2 of API suggestion to our Use Cases Workflow documentation https://docs.google.com/document/d/1COP7Qg9XjPnaxTqG-dSgB8tq3X504QOCm9A7lk4Iqaw/edit

Alex Garnett

unread,
Aug 12, 2013, 12:46:24 PM8/12/13
to dataverse...@googlegroups.com, Eleni Castro
Thanks for that, Merce -- I know we discussed it at one point in the context of whether we wanted to allow OJS to hook into two different DVs or provide the "mirroring" functionality on the DV side but it might've fallen out of our notes in the meantime.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-community+unsub...@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.
Reply all
Reply to author
Forward
0 new messages