Hi All,
I forgot to mention that the svn diff patch is created on the base of
stanbol-nlp-processing branch
Alessio
On 01/09/2013 10:22 AM, Alessio Bosca wrote:
Hi Rupert,
yesterday I updated the morphological Analysis service (We added
support for the sv language, fixed the umlauts issue for German and
unified all the POS tagsets for the supported languages)
Therefore currently the pos tags returned by the web service are not
coherent with the ones declared in the postag mappings in the
enhancement engine. For this reasons the tests on the morphological
engine should be failing.
Concerning the umlauts (and /ß character)/, they are now recognized by
the system but currently the lemma produced by the morphological
analyzer converts them to sequences of characters (e.g. /ö -> oe, //ß
-> ss)./
In the future version of the system I would like to include both lemma
writings/.
/
The language identifier has been updated as well (adding a few new
languages, now the list of supported languages includes
en,fr,de,hu,pl,it,es,pt,el,et,lv,tr,pt,ru,ar,ro,da) without changes
for the engines
The patch you can find in the attachment contains the fixes for the
morphological service updates (sv language addition and new pos tag
mapping, one for all the languages)
The patch also contains the client and the test classes for the
sentiment analysis engine supporting fr and it
Let me know if you have any problem integrating the patch
Alessio
On 01/07/2013 05:44 PM, Alessio Bosca wrote:
Hi Rupert,
sure tomorrow I'll have a look into that and let you know
bests
Alessio
On 01/04/2013 01:07 PM, Rupert Westenthaler wrote:
Hi Alessio, all
Thanks for looking into that. However with the Jenking build #1200
there is still one remaining issue
tesetEngine(org.apache.stanbol.enhancer.engines.celi.langid.impl.CeliLanguageIdentifierEnhancementEngineTest)
Time elapsed: 0.296 sec <<< FAILURE!
junit.framework.ComparisonFailure: The detected language for text
'Brigitte Bardot, née le 28 septembre 1934 à Paris, est une actrice
de cinéma et chanteuse française.' MUST BE 'fr' expected:<[f]r> but
was:<[a]r>
at junit.framework.Assert.assertEquals(Assert.java:100)
at
org.apache.stanbol.enhancer.engines.celi.langid.impl.CeliLanguageIdentifierEnhancementEngineTest.tesetEngine(CeliLanguageIdentifierEnhancementEngineTest.java:101)
Looks like the Language Identification engine detects the language
Arabic for the French text example. I am also able to reproduce this
issue locally.
Can you have a look into that?
best
Rupert
On Fri, Jan 4, 2013 at 10:03 AM, Alessio Bosca
<[email protected]> wrote:
Hi Rupert,
thanks for the feedback, there was a problem with the the access
control for
anonymous users in the services. Now it has been fixed .
PS: Next week I'll send you a patch for CELI engines with the
sentiment
analysis and bug fixes for the German umlauts.
Sorry for the delay in the release.
Bests,
Alessio
--
*************************************
Alessio Bosca, Ph.D.
CELI s.r.l.
Via San Quintino 31
10121 Torino
Tel. +39 011.562.71.15
Fax +39 011.506.40.86
http://www.celi.it
*************************************
--
*************************************
Alessio Bosca, Ph.D.
CELI s.r.l.
Via San Quintino 31
10121 Torino
Tel. +39 011.562.71.15
Fax +39 011.506.40.86
http://www.celi.it
*************************************
--
*************************************
Alessio Bosca, Ph.D.
CELI s.r.l.
Via San Quintino 31
10121 Torino
Tel. +39 011.562.71.15
Fax +39 011.506.40.86
http://www.celi.it
*************************************
