Hi,

I wonder, do we only need to use the (custom) scorer if the query orders
by the pseudo-property "jcr:score"? I mean, it's probably faster not to
use any scorer, right?

Regards,
Thomas



On 15/01/15 08:17, "Rishabh Kumar Maurya" <[email protected]> wrote:

>Hi,
>
>We have search enhancements requests based on search results relevance
>like -
>
>1. boosting score of recently modified documents.
>2. boosting documents which are created/last updated by current session
>user.(OR boosting on basis specific docfield value).
>3. boosting documents with a field value in certain range.
>
>these are similar to Solr Function Queries[0] use case. I have logged a
>issue[1].
>
>With Chetan's help I came up with a custom scorer implementation[2],
>which exposes lucene CustomScoreQuery[3] via OSGI and one can implement
>custom ScorerProvider and use it by configuring index definition with
>custom ScorerProvider name.
>
>Implementation details  -
>
>1. ScorerProvider - ScorerProvider creates lucene CustomScoreQuery, which
>user can implement and configure its name in index definition.
>
>2. ScorerProviderFactory - ScorerProviderFactoryImpl  tracks all
>ScorerProvider services bound and looks up for  ScorerProvider from its
>name.
>
>3. test - customScoreQuery -  boost document with specific docfield value
>( in this case jcr:createdBy containing 'foo')
>
>Please review and provide feedbacks. Also, this implementation has
>following requirements  -
>
>1. Scorer uses lucene packges(for now exported all lucene packages from
>oak-lucene)
>2. Some way to access field names in scorer implementation on the basis
>of name defined in property definition
>
>
>Thanks
>
>[0] http://wiki.apache.org/solr/FunctionQuery
>[1] https://issues.apache.org/jira/browse/OAK-2399
>[2] https://github.com/rishabhmaurya/jackrabbit-oak/commit/9042200bd
>[3] 
>http://lucene.apache.org/core/4_7_1/queries/org/apache/lucene/queries/Cust
>omScoreQuery.html
>

Reply via email to