if(mode == MONO_LINGUAL){
new BuildIntDocVectors(conf).run();
new BuildWeightedIntDocVectors(conf).run();
LOG.info("Job BuildWeightedIntDocVectors finished in "+(System.currentTimeMillis()-startTime)/1000.0+" seconds");
}else{
BuildTargetLangWeightedIntDocVectors weightedIntVectorsTool = new BuildTargetLangWeightedIntDocVectors(conf);
LOG.info("Job BuildTargetLangWeightedIntDocVectors finished in "+(System.currentTimeMillis()-startTime)/1000.0+" seconds");
int finalNumDocs = weightedIntVectorsTool.run();
if(finalNumDocs > 0){
LOG.info("Changed doc count from "+env.readCollectionDocumentCount() + " to = "+finalNumDocs);
env.writeCollectionDocumentCount(finalNumDocs);
}
// set Property.CollectionTermCount to the size of the target vocab. since all docs are translated into that vocab. This property is read by WriteRandomVectors via RunComputeSignatures.