Running Open refine without opening a browser window (new option in refine.ini)

[Mateja Verlic]

While integrating LOD-enabled version of OpenRefine (LODGrefine) into LOD2 stack we encountered an issue when we tried to run OpenRefine as a service. The script tries to open up a browser window, which is not allowed in our case. 

For this purpose I added a new option (refine.start_browser=false) to the refine.ini and changed code in Refine and RefineClient class (server/src/Refine.java), so that it checks whether it should open up a browser window or not. 

Because this is not really a general issue for OpenRefine, I didn't create a new issue on Github. But if you think this could come handy for other users, I can provide a patch.

BR,

Mateja (sparkica)

[David Huynh]

Hi Mateja,

I admit to not having much experience with git, but I presume you could just submit to your own branch and then we can pull it in, right?

David

--
 
 

[Mateja Verlic]

Hi,

yes, I think it should work this way. Thanks.

Mateja

[Sergio Fernández]

Hi Mateja,

a proper application server performs much better that the embedded
jetty. That's why we work on a proper WAR file for Refine some months
ago. See the thread at:

http://markmail.org/message/jb4w5ugfpq5jkz6x#query:+page:1+mid:zkodesocwzkaeui4+state:results

Right now such fork doesn't include the latest improvements since
OpenRefine, but we already talked about discuss it after the 2.6
release, just in case you'd be interested.

Best,

> --
>
>



--

Sergio Fernández <ser...@wikier.org>

[Mateja Verlic]

Hi Sergio,

thanks for good news. I am interested, I think this could make integration of OpenRefine easier in our case. 

AFAIK we use Tomcat for other components of LOD2 stack and currently there are some issues with redirecting refine. Furthermore, we have to create a Debian package for OpenRefine and I don't know (yet) how this affects installation of WAR files.

Unfortunately, I don't have enough experience to help you test it in other containers, but if I can contribute in any other way, let me know.

Best,

Mateja

[Mateja Verlic]

It seems I've somehow missed the option "refine.headless=true" :)

Best, 

Mateja

On Sunday, November 18, 2012 4:52:18 AM UTC+1, David Huynh wrote:

[Tom Morris]

On Mon, Nov 19, 2012 at 3:12 AM, Sergio Fernández <ser...@wikier.org> wrote:

a proper application server performs much better that the embedded
jetty. That's why we work on a proper WAR file for Refine some months
ago. See the thread at:

http://markmail.org/message/jb4w5ugfpq5jkz6x#query:+page:1+mid:zkodesocwzkaeui4+state:results

Right now such fork doesn't include the latest improvements since
OpenRefine, but we already talked about discuss it after the 2.6
release, just in case you'd be interested.

If you're work is forked from the git repository, it should be fairly easy to track the changes being made for 2.6 and reintegrate when desired.

Also, I'm presuming this is different than the relative path changes that you requested and provided the patch for.  Those have been integrated, although not extensively tested (unless you guys have tested them).

Tom 

[Sergio Fernández]

Hi,

On 20 November 2012 16:05, Tom Morris <tfmo...@gmail.com> wrote:
> If you're work is forked from the git repository, it should be fairly easy
> to track the changes being made for 2.6 and reintegrate when desired.

Not, the fork was made from the svn in the old times, but I can track
the changes from our mercurial. The

> Also, I'm presuming this is different than the relative path changes that
> you requested and provided the patch for.

I think all the functional code (java or javascript) had been added
with the provided patches. The single missing thing right now affects
on how the war is build, details at:
http://code.google.com/p/lmf/wiki/GoogleRefineDevelopersDocumentation#Maven

But, as we already discuss, let's keep this discussion for after the
2.6 release.

> Those have been integrated,
> although not extensively tested (unless you guys have tested them).

I'd say that more or less tested, yes, since we include Refine in our
LMF installer (see http://goo.gl/WrHKA for instance) and use it in in
some industrial scenarios.

Cheers,

--

Sergio Fernández <ser...@wikier.org>

[Tom Morris]

On Tue, Nov 20, 2012 at 10:35 AM, Sergio Fernández <ser...@wikier.org> wrote:

But, as we already discuss, let's keep this discussion for after the
2.6 release.

That's fine.  I'm a little worried about this statement from that page though:

"Or, if you'd preferred, we have a forked version of the code with some patches that are not still applied in the official repository and fix some important issues:"

Can you provide a pointer these important issues which are unfixed?  We'd like to make sure the 2.6 release includes as many cleanups as possible.

Tom

[Sergio Fernández]

On 20 November 2012 19:24, Tom Morris <tfmo...@gmail.com> wrote:
> I'm a little worried about this statement from that page though:
>
> "Or, if you'd preferred, we have a forked version of the code with some
> patches that are not still applied in the official repository and fix some
> important issues:"
>
> Can you provide a pointer these important issues which are unfixed? We'd
> like to make sure the 2.6 release includes as many cleanups as possible.

Apologize, since you already applied the patches, that text is old and
not updated with the current status. I'll fix it right now. Sorry.

--

Sergio Fernández <ser...@wikier.org>