[jira] [Commented] (SOLR-11386) Extracting learning to rank features fails when word ordering of EFI argument changed.

["Michael A. Alcorn (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SOLR-11386?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16203861#comment-16203861 ] 

Michael A. Alcorn commented on SOLR-11386:
------------------------------------------

-I just set up a local install of Solr 6.6.0 with a toy data set and tested multi-term EFI arguments using single quotes and it worked as expected. The issue seems to be isolated to older Solr versions. We'll upgrade our development version and see if that fixes it.-

I was incorrect. The issue persists in Solr 6.6.0, however, I believe I've discovered a workaround. If you use:

{code}
    {
        "store": "redhat_efi_feature_store",
        "name": "case_description_issue_tfidf",
        "class": "org.apache.solr.ltr.feature.SolrFeature",
        "params": {
            "q":"{!dismax qf=text_tfidf}${text}"
        }
    }
{code}

instead of:

{code}
    {
        "store": "redhat_efi_feature_store",
        "name": "case_description_issue_tfidf",
        "class": "org.apache.solr.ltr.feature.SolrFeature",
        "params": {
            "q": "{!field f=issue_tfidf}${case_description}"
        }
    }
{code}

you can then use single quotes to incorporate multi-term arguments as [~alessandro.benedetti] suggested.

> Extracting learning to rank features fails when word ordering of EFI argument changed.
> --------------------------------------------------------------------------------------
>
>                 Key: SOLR-11386
>                 URL: https://issues.apache.org/jira/browse/SOLR-11386
>             Project: Solr
>          Issue Type: Bug
>      Security Level: Public(Default Security Level. Issues are Public) 
>          Components: contrib - LTR
>    Affects Versions: 6.5.1
>            Reporter: Michael A. Alcorn
>
> I'm getting some extremely strange behavior when trying to extract features for a learning to rank model. The following query incorrectly says all features have zero values:
> {code}
> http://gss-test-fusion.usersys.redhat.com:8983/solr/access/query?q=added couple of fiber channel&rq={!ltr model=redhat_efi_model reRankDocs=1 efi.case_summary=the efi.case_description=added couple of fiber channel efi.case_issue=the efi.case_environment=the}&fl=id,score,[features]&rows=10
> {code}
> But this query, which simply moves the word "added" from the front of the provided text to the back, properly fills in the feature values:
> {code}
> http://gss-test-fusion.usersys.redhat.com:8983/solr/access/query?q=couple of fiber channel added&rq={!ltr model=redhat_efi_model reRankDocs=1 efi.case_summary=the efi.case_description=couple of fiber channel added efi.case_issue=the efi.case_environment=the}&fl=id,score,[features]&rows=10
> {code}
> The explain output for the failing query can be found here:
> https://gist.github.com/manisnesan/18a8f1804f29b1b62ebfae1211f38cc4
> and the explain output for the properly functioning query can be found here:
> https://gist.github.com/manisnesan/47685a561605e2229434b38aed11cc65



--
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