[jira] [Commented] (STANBOL-1271) Allow passing content by reference to the Stanbol Enhancer

["Rupert Westenthaler (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/STANBOL-1271?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13884328#comment-13884328 ] 

Rupert Westenthaler commented on STANBOL-1271:
----------------------------------------------

This would basically mean to

(1) update the ContentItemReader (enhancer.jersey module) to specially treat requests with the Content-Type "message/external-body". 

Instead of reading the POST data it will need to read the URL paraemter and create a UrlReference (enhancer.servicesapi module). With this UrlReference it can create the ContentItem using the ContentItemFactory.


(2) Change how ContentItemFactory handles ContentReferences:

Currently the Content referenced by a ContentReference (e.g. a UrlReference) is dereferenced completely downloaded when creating the Blob. There is also an option to delay this until the first access to the Blobs methods. But in any case it will be downloaded before the user can acquire a stream from the Blob. 

This behavior should be changed in a way that the content is not locally cached, but is directly streamed from the reference. In case of a UrlReference this would mean to open an UrlConnection on every call to Blob.getStream() 


> Allow passing content by reference to the Stanbol Enhancer
> ----------------------------------------------------------
>
>                 Key: STANBOL-1271
>                 URL: https://issues.apache.org/jira/browse/STANBOL-1271
>             Project: Stanbol
>          Issue Type: Improvement
>          Components: Enhancer
>    Affects Versions: 0.12.0
>            Reporter: Sergio Fernández
>              Labels: RFC2017
>
> For some use cases we have in mind in the MICO EU project, we'd need to send really large content to Stanbol. Therefore the current REST interface (posting the content directly) is not suitable in such scenario. So we'd need to have a way to send content by reference to the Stanbol Enhancer.
> I've found RFC2017 (http://tools.ietf.org/html/rfc2017), which defines a way to parsed external content by using
>     Content-type: message/external-body; access-type=URL; URL="http://example.org/file"
> If the community fins this feature interesting, we can start to play a bit with a draft implementation, and contribute it as soon as it's working.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)