[ 
https://issues.apache.org/jira/browse/MARMOTTA-423?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13874642#comment-13874642
 ] 

Sergio Fernández commented on MARMOTTA-423:
-------------------------------------------

after some work, I've achieved the initial scope of the task (directly call the 
api and ignore the content type issue), but the provider is not usable at all 
because the API is returning invalid Turtle... mainly:

* literals with non-escaped quotes inside
* several codification issues 

This notes by AndyS may be relevant for the problem:
 
http://people.apache.org/~andy/Freebase20121223/Notes.txt

And there is a Perl script for fixing the dumps:

https://github.com/vigsterkr/freebase-tdb/blob/master/fixit

But I'd prefer to not re-implement it in the provider if possible... but I 
guess there is no other option.

> Freebase returns wrong content type
> -----------------------------------
>
>                 Key: MARMOTTA-423
>                 URL: https://issues.apache.org/jira/browse/MARMOTTA-423
>             Project: Marmotta
>          Issue Type: Improvement
>          Components: LDCache, LDClient
>    Affects Versions: 3.1-incubating
>            Reporter: Sergio Fernández
>            Priority: Minor
>              Labels: freebase
>             Fix For: 3.2
>
>   Original Estimate: 2h
>  Remaining Estimate: 2h
>
> Requests to Freebase return an incorrect content type:
> {quote}
> $ curl -I -H "Accept: text/turtle" http://rdf.freebase.com/ns/m.07zqbwz
> HTTP/1.1 200 OK
> Date: Fri, 17 Jan 2014 07:43:03 GMT
> Content-Type: text/plain; charset=UTF-8
> {quote}
> Even if it's not a problem in our side (Google does not care), we could do 
> something to fix the issue from LDClient. And maybe directly avoid the 301 
> redirection to the API:
> https://developers.google.com/freebase/v1/rdf



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

Reply via email to