You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Alan Woodward (JIRA)" <ji...@apache.org> on 2019/03/11 12:46:00 UTC
[jira] [Commented] (LUCENE-8719) FixedShingleFilter can miss paths
at the end of a TokenStream
[ https://issues.apache.org/jira/browse/LUCENE-8719?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16789505#comment-16789505 ]
Alan Woodward commented on LUCENE-8719:
---------------------------------------
Here is a patch that fixes the issue. Instead of immediately returning `false` when `incrementGraphToken` hits the end of the token stream, we loop back round and try incrementing the base graph again to see if there are further paths that can produce shingles.
> FixedShingleFilter can miss paths at the end of a TokenStream
> -------------------------------------------------------------
>
> Key: LUCENE-8719
> URL: https://issues.apache.org/jira/browse/LUCENE-8719
> Project: Lucene - Core
> Issue Type: Bug
> Reporter: Alan Woodward
> Assignee: Alan Woodward
> Priority: Major
> Attachments: LUCENE-8719.patch
>
>
> If we have a set of paths of differing lengths at the end of a TokenStream, FixedShingleFilter will currently stop when the first path hits the end of the stream, but this may miss extra longer paths. For example, the stream "a b:3/c d e f" with a shingle size of 3 should produce paths "a b f", "a c d", "c d e" and "d e f", but at the moment FixedShingleFilter will stop when it finds that the path "b f" is too short to produce a shingle.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org