Hi Solrists,
thank you for your kind responses.
Grant, François, I'll keep your advice in mind & your links in store;
they may be useful in one of my use cases,
even though I doubt they might in the primary one.
As RDF litterals, my documents are affixed with language tags @en, @fr, @ja &c,
as per ISO 639-1, so that language identification is straightforward.
It's the discovery of the corresponding Analyzer subclasses & constructors
I'm trying to automate.
Solr looks like it's up to the server admins to specify in XML
what Analyzer subclasses they want in a given case,
then it's up to Solr to instantiate those subclasses by Java reflection.
I would like to spare myself the burden to write & maintain this XML.
Rather, I'd use Java code to build the mapping
by inventorying the classpath, with rules like
"on finding jarentry /whats/this/package/analysis/xx/WhatsThisAnalyzer.class,
if class WhatsThisAnalyzer is a subclass of lucene.analysis.Analyzer,
if reflection reveals a public new WhatsThisAnalyzer(lucene.util.Version),
if instantiation succeeds,
then the instance is the presumptive default analyzer for ISO 639-1 code xx".
Might make a Lucene submission, more properly than a Solr one.
Thanks again for your time & your help.
Best regards,
François Jurain.
> Message du 25/03/11 à 23h06
> De : "François Schiettecatte" <fschietteca...@gmail.com>
> A : solr-user@lucene.apache.org
> Copie à :
> Objet : Re: Wanted: a directory of quick-and-(not too)dirty analyzers for
> multi-language RDF.
>
>
> François
>
> I think there is a language identification tool in the Nutch code base,
> otherwise I have written one in Perl which could easily be translated to
> Java. I wont have access to it for 10 days (I am traveling), but I am happy
> to send you a link to it when I get back (and anyone else who wants it).
>
> Cheers
>
> François
>
> On Mar 25, 2011, at 11:59 AM, Grant Ingersoll wrote:
>
> > You are looking for a language identification tool. You could check
> > https://issues.apache.org/jira/browse/SOLR-1979 for the start of this.
> > Otherwise, you have to roll your own or buy a third party one.
> >
> > On Mar 24, 2011, at 12:24 PM, fr.jur...@voila.fr wrote:
> >
> >> Hello Solrists,
> >>
> >> As it says in the subject line, I'm looking for a Java component that,
> >> given an ISO 639-1 code or some equivalent,
> >> would return a Lucene Analyzer ready to gobble documents in the
> >> corresponding language.
> >> Solr looks like it has to contain one,
> >> only I've not been able to locate it so far;
> >> can you point the spot?
> >>
> >> I've found org.apache.solr.analysis,
> >> and thing like org.apache.lucene.analysis.bg &c in lucene/modules,
> >> with many classes which I'm sure are related, however the factory itself
> >> still eludes me;
> >> I mean the Java class.method that'd decide on request, what to do with all
> >> these packages
> >> to bring the requisite object to existence, once the language is specified.
> >> Where should I look? Or was I mistaken & Solr has nothing of the kind, at
> >> least in Java?
> >> Thanks in advance for your help.
> >>
> >> Best regards,
> >> François Jurain.
> >>
> >> ____________________________________________________
> >>
> >> Retrouvez les 10 conseils pour économiser votre carburant sur Voila :
> >> http://actu.voila.fr/evenementiel/LeDossierEcologie/l-eco-conduite/
> >>
> >>
> >>
> >
> > --------------------------
> > Grant Ingersoll
> > http://www.lucidimagination.com/
> >
> > Search the Lucene ecosystem docs using Solr/Lucene:
> > http://www.lucidimagination.com/search
> >
>
>
>
____________________________________________________
Suivez toute l'actualité en photos de l'émission Carré Viiip et retrouvez les
derniers échanges des viiip sur :
http://people.voila.fr/evenementiel/carre-viiip
