Hi HSearch team,

I'm using Hibernate Search as the search engine for my company web application and it is required to work with all languages. Could I achieve that requirement? At the moment, my search feature seems to work fine with English, French, German ... but can not with Arabic ... (something like أنا لا أعرف)

Thanks a lot.
dungleonhart

Hi,
we rely on the Analyzers provided by the Apache Lucene community: any existing Analyzer can be used with Hibernate Search.

So for Arabic for example look at http://lucene.apache.org/core/3_6_0/api/all/org/apache/lucene/analysis/ar/ArabicAnalyzer.html

Hi Sane,

Thanks a lot for your reply.

I can see a lot of analyzers in the HSearch library: EnglishAnalyzer, ArabicAnalyzer, SwedishAnalyzer ...
As I understand, we'll user the appropriate analyzer for the language that we want to support. So in case I want to support both English and Arabic, what should I do? Is there something like a "Unified"Analyzer to use?

Thanks a lot and Best regards,
dungleonhart

Hi,
there is no "Unified" Analyzer: the default one works on most languages, but poorly as you have found out it has problems with many languages.

The best strategy is to index each entity instance using it's correct language, there is an example here:
http://docs.jboss.org/hibernate/search/4.2/reference/en-US/html_single/#d0e3840

You also should parse the user queries using the appropriate Analyzer: so you need to know the language of each entity, and also need to know the language of user input.

I'd suggest to always restrict the results to the matching language, you could use a Filter for that.

Hi Sane,

It is very clear to me now.
Thank you a lot for your help.