[jira] [Commented] (LUCENE-7736) Expose some IndexReader stats via DoubleValuesSources

["Adrien Grand (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/LUCENE-7736?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16210111#comment-16210111 ] 

Adrien Grand commented on LUCENE-7736:
--------------------------------------

Thanks Alan. I think those new functions are going to be useful.

bq. One wrinkle here is that we can't compare closures, so IndexReaderDoubleValuesSource only uses its description field for comparisons. These are all private implementations, and there's a comment to that effect, so I think this is OK.

I think this is incorrect as two instances that have been built on a different reader will be considered equals even though they produce different values? Let's pass things that are actually comparable to {{IndexReaderDoubleValuesSource}} or just live with the fact that two instances that do the same thing are considered unequal (which is ok imo as long as a.equals(b) returns true if a == b)?

Should we use == rather than equals to know whether the source was rewritten? I don't like one more than the other but since queries already use == I'd rather like things to be consistent?

Did you remove the {{if (boost == 1f) return expl;}} from {{FunctionScoreQuery}} on purpose? I know it is not necessary for correctness but I still like the fact that it makes the explanation easier to read in the default case that the boost is 1.

In {{TermFreqDoubleValuesSource}} could we just return an empty instance if the term is not found instead of having to do a null check for every document?

In {{lucene/queries/src/java/org/apache/lucene/queries/function/ValueSource.java}} you seem to be assuming that rewriting only once is enough? Is it true?



> Expose some IndexReader stats via DoubleValuesSources
> -----------------------------------------------------
>
>                 Key: LUCENE-7736
>                 URL: https://issues.apache.org/jira/browse/LUCENE-7736
>             Project: Lucene - Core
>          Issue Type: New Feature
>            Reporter: Alan Woodward
>            Assignee: Alan Woodward
>            Priority: Minor
>         Attachments: LUCENE-7736.patch, LUCENE-7736.patch, LUCENE-7736.patch, LUCENE-7736.patch
>
>
> We have a number of ValueSource implementations that expose IndexReader stats (numDocs, termFreq, etc).  We should re-implement these as DoubleValuesSources, allowing them to be used in FunctionScoreQuery, etc.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

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