So with https://issues.apache.org/jira/browse/LUCENE-4749 it's possible to set the ModelFile?
<tokenizer class="solr.UIMAAnnotationsTokenizerFactory" descriptorPath="/uima/AggregateSentenceAE.xml" tokenType="org.apache.uima.SentenceAnnotation" ngramsize="2" modelFile="file:german/TuebaModel.dat" /> ??? Thanks, Kai -----Original Message----- From: Tommaso Teofili [mailto:tommaso.teof...@gmail.com] Sent: Monday, February 04, 2013 2:47 PM To: solr-user@lucene.apache.org Subject: Re: Indexing nouns only with UIMA works - performance issue? see an example at http://svn.apache.org/viewvc/lucene/dev/branches/branch_4x/solr/contrib/uima/src/test-files/uima/uima-tokenizers-schema.xml?view=diff&r1=1442116&r2=1442117&pathrev=1442117where the 'ngramsize' parameter is set, that's defined in AggregateSentenceAE.xml descriptor and is then set with the given actual value. HTH, Tommaso