[jira] [Commented] (LUCENE-3215) SloppyPhraseScorer sometimes computes Infinite freq

["Doron Cohen (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/LUCENE-3215?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13100182#comment-13100182 ] 

Doron Cohen commented on LUCENE-3215:
-------------------------------------

An update on this...

This is not related to LUCENE-3142 - the latter was fixed but this one still fails.

The patch fix which 'abs' the distance indeed avoids the infinite score problem, but I was not 100% comfortable with it - how can the distance be none positive?

Digging into it shows a wrong assumption in SloppyPhraseScorer:

{code}
    private int initPhrasePositions() throws IOException {
        int end = 0;
{code}

The initial value of end assumes that all positions will be nonnegative.
But this is wrong, as PP position is computed as 

{code}
  position = postings.nextPosition() - offset
{code}

So, whenever the query term appears in the doc in a position smaller than its offset in the query, the computed position is negative. The correct initialization for end is therefore:

{code}
    private int initPhrasePositions() throws IOException {
        int end = Integer.MIN_VALUE;
{code}

You would expect this bug to surfaced sooner...

Anyhow, for the 3 tests that Robert added, this only resolve testInfiniteFreq1() but the other two tests still fail, investigating...

> SloppyPhraseScorer sometimes computes Infinite freq
> ---------------------------------------------------
>
>                 Key: LUCENE-3215
>                 URL: https://issues.apache.org/jira/browse/LUCENE-3215
>             Project: Lucene - Java
>          Issue Type: Bug
>            Reporter: Robert Muir
>            Assignee: Doron Cohen
>         Attachments: LUCENE-3215.patch, LUCENE-3215_test.patch, LUCENE-3215_test.patch
>
>
> reported on user list:
> http://www.lucidimagination.com/search/document/400cbc528ed63db9/score_of_infinity_on_dismax_query

--
This message is automatically generated by JIRA.
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