Hi All!

Has anyone ever had the following problem...

If use FastVectorHighlighter and boundaryScanner with hl.bs.type set to SENTENCE there are situations when not all matched terms are highlighted in fragments.

I.e.

1. FastVectorHighlighter generates FieldFragList with info on what terms to highlight

(org.apache.lucene.search.vectorhighlight.FastVectorHighlighter.getFieldFragList(FragListBuilder, FieldQuery, IndexReader, int, String, int))

2. but then boundaryScanner extends fragment further to keep entire SENTENCE. I guess it just adds text till the nearest dot. If there are any matching terms in that text added they not get highlighted (the reason is that in 1. there is no terms highlight info for the text added).



Has anyone was able to get the entire fragment terms highlighting and preserve sentence boundaries at the same time?

I guess that solution can be just not generate fragments but highlight the entire text and fragment it myself.

What do you think guys?

Thanx,

Reply via email to