lucene 7 add token offset problems
eroux opened this issue · comments
when analyzing "dharmottara"@sa-alalc97
, the new Fuseki is unhappy:
startOffset must be non-negative, and endOffset must be >= startOffset, and offsets must not go backwards startOffset=5,endOffset=11,lastStartOffset=7 for field 'rdfsLabel_sa-alalc97'
there might be an ungarded call, and we could update to Lucene 7 API
There was an unguarded call to setOffset
in SkrtSyllableTokenizer.end
. I'll make a release for version 1.0.6 that includes Élie's changes as well and test.
We are now on lucene-sa v1.0.10 and a try/catch is added to TextIndexLucene.addDocument
to capture failures that occur outside of lucene-sa during the indexing process. This allows to collect more specific data that should help isolate the problem in lucene-sa: