Hi Monika,
First, i'm sorry for taking so long to reply, i've got to swtich priorities
here and now i'm back to this problem.
I am not clear how much human involvement you think you need in the
> process.
>
> Step 1 ) export to dublin core in a way that DSPACE can import from
> directories
> Step 2) ???
> Step 3) import
>
> do you want a person to review the exported data to make decisions about
> whether to import or not, whether to fix up metadata , and to decide which
> collections items should be assigned to ? If the answer is yes your job
> will be difficult.
>
To answer the last question, kind of yes, i'm gonna try to give a better
explanation of what is my problem:
Let's say somebody is graduating (or getting a masther or doctorate
degree), this person has to do some academic work (tcc, thesis...), this
work is firstly submitted in our academic system with, by "submitted" i
mean that a lot of the metadata is being submitted, like: author names,
title, description, type...
IF this work is approved (example:the person graduates)[1], we then
*have[2]* to pass it to DSpace as a not finished[3] submission because we
have to add some fields that won't come from the academic system, like:
license, author curriculum (like orcid, but we use one called lattes), the
file itself (pdf) and so on, the person that will copy the metadata from
the academic system and add its missing metadata will be someone with a
submitter role for its collection, and after this person finishes it, some
reviewer[4] will have to accept the submission, finally publishing it.
Considering this is an all time ongoing process, you can see every attempt
to automate it would be great.
Now, what i want to do is the following, since we can't ask for an OAI
interface yet:
The academic system to have a button on this record view page that says
something like "export metadata", and it will generate a file containing
all the possible metadata used by Dspace, ready for import;
The person that is used to manually copy the metadata will access the
academic system, download it and import to DSpace, using the WebUI as a
non-admin user (altough as a submitter for this collection), as a non
finished submission;
This same person will now fulfill the missing fields, upload the pdf,
accept the license, simplifying, this person will finish the submission,
passing the ball to a reviewer whom will publish the submission.
We have an opportunity on the next few days of asking for the manager of
this academic system to implement this, at the same time aiming for an OAI
interface on the future (they won't do it now).
So i need to:
1) Define a metadata format that the academic system will follow to export
this data.
2) Find a way in which a non-admin DSpace user can import this metadata as
a non finished submission.
I need help with both 1) and 2)
For 2 i found this
https://wiki.duraspace.org/display/DSDOC5x/Importing+and+Exporting+Items+via+Simple+Archive+Format#ImportingandExportingItemsviaSimpleArchiveFormat-UIBatchImport(JSPUI)
Which says only admin users can import data and only mentions batch import,
i believe i would need the basic "import metadata" option above the batch
one.
For 1 i found some info on the DSpace wiki suggesting that i may use DC
Metadata for importation.
Every comment regarding this problem is highly appreciated, thanks!
If i wasn't clear about anything, please point out and i will happily try
to clarify it.
[1]When it is approved, it's made within the academic system, so there is a
way of identifying only those records we need from the academic system.
[2]We have to, but what is really made is that one person manually copies
the data showed on the academic system, so from this person's POV is just a
normal submission.
[3]By "not finished" i mean that this submission is still pending to the
submitter and is not available for review yet (no license accepted yet, for
example).
[4]This comes from the collection workflow setting, as i understood;
someone submits, another one review, and the review process publishes. The
separation of submitters and reviewers is given by grouping users and
setting their permissions for wanted collections.
Samuel Henrique O. P. [samueloph]
2016-05-13 12:26 GMT-03:00 Monika C. Mevenkamp <moni...@princeton.edu>:
> You are right of cause
>
> I am not clear how much human involvement you think you need in the
> process.
>
> Step 1 ) export to dublin core in a way that DSPACE can import from
> directories
> Step 2) ???
> Step 3) import
>
> do you want a person to review the exported data to make decisions about
> whether to import or not, whether to fix up metadata , and to decide which
> collections items should be assigned to ? If the answer is yes your job
> will be difficult.
>
> If the answer is no , along with we’ll generate Dublin core in batches
> that match
