lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roberto Minelli (Commented) (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (LUCENE-2587) Highlighter picks wrong offset for fragment boundaries
Date Wed, 02 Nov 2011 17:37:32 GMT

    [ https://issues.apache.org/jira/browse/LUCENE-2587?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13142329#comment-13142329
] 

Roberto Minelli commented on LUCENE-2587:
-----------------------------------------

Suppose I have a dummy Fragmenter that returns isNewFragment() = true on F and K. I have the
text "A b c d e. F g h i j! K l m n o. ". I search for "g" hence the Highlighter will pick
the middle fragment. The fragment it returns is ". F <b>g</b> h i j". I was wondering
if the correct fragment should be " F <b>g</b> h i j" or even "F <b>g</b>
h i j" without the starting whitespace. The starting "." is a bug and matches the issue LUCENE-2587.
I was wondering if, instead, the starting whitespace at the beginning of the fragment is correct
or not. And, in the end, if I understood correctly the issue LUCENE-2587, but I think so.
Thanks in advance to anyone who wants to answer. RM
                
> Highlighter picks wrong offset for fragment boundaries
> ------------------------------------------------------
>
>                 Key: LUCENE-2587
>                 URL: https://issues.apache.org/jira/browse/LUCENE-2587
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: modules/highlighter
>    Affects Versions: 3.0.2
>         Environment: Java 6 + Lucene 3.0.2
>            Reporter: Terje Eggestad
>            Priority: Trivial
>              Labels: newdev
>
> I have written  a new Fragmenter since we need fragments for hitlines to be on sentence
boundaries and not cross paragraphs. 
> When using it with org.apache.lucene.search.highlight.Highlighter, I get hitlines that
starts with ". ", "? ", "! "...
> Consider the text  "A b c d e. F g h i j! K l m n o. " 
> which become the tokenstream : (A) (b) (c) (d) (e) (F) (g) (h) (i) (j) (K) (l) (m) (n)
(o)  
> If the fragmenter return isNewFragment()  = true on F and K and Highlighter pick the
middle fragment, lets say we search on "g" the hitline becomes:
> ". F <B>g</B> h i j"
> The reason, it seems, is that the offset to the fragment boundaries found by taking the
endOffset of the last token in a fragment , 
> not the startOffset of the first. 
> TJ

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org


Mime
View raw message