[
https://issues.apache.org/jira/browse/LUCENE-8644?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16746128#comment-16746128
]
Alan Woodward commented on LUCENE-8644:
---------------------------------------
Jim and I talked about this offline, and agreed that the best general solution
here is to extend GraphTokenStreamFiniteStrings to record all the token
attributes, and change the QueryBuilder signatures to pass around
AttributeSources rather than restrict things to just the Term and the type.
This means we can still access type information when building queries, but
we're free to use other attributes as well. I've attached a new path
implementing this idea.
> Pass token types to QueryBuilder.newSynonymQuery()
> --------------------------------------------------
>
> Key: LUCENE-8644
> URL: https://issues.apache.org/jira/browse/LUCENE-8644
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Alan Woodward
> Assignee: Alan Woodward
> Priority: Major
> Attachments: LUCENE-8644.patch, LUCENE-8644.patch
>
>
> SynonymGraphFilter distinguishes between input tokens and tokens that it
> injects into the token stream by setting the TypeAttribute to SYNONYM.
> QueryBuilder currently ignores this information, but it would be useful to
> pass it through to the newSynonymQuery() method, as this would allow
> queryparsers to apply differential boosts to synonyms and original terms.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
