mahout-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Grant Ingersoll (Commented) (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAHOUT-964) RowSimilarityJob should exit immediately if an invalid similarity measure specified and it would be nice to have an --overwrite option for the RowSimilarityJob CLI
Date Sat, 28 Jan 2012 15:31:10 GMT

    [ https://issues.apache.org/jira/browse/MAHOUT-964?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13195566#comment-13195566
] 

Grant Ingersoll commented on MAHOUT-964:
----------------------------------------

Hey Suneel,

Content of patch is fine, but can you, going forward, generate your patch so that it can be
applied from the top level directory?  This will make it easier to apply since we don't have
to go chasing down what directory it applies in.

Thanks!
                
> RowSimilarityJob should exit immediately if an invalid similarity measure specified and
it would be nice to have an --overwrite option for the RowSimilarityJob CLI
> -------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: MAHOUT-964
>                 URL: https://issues.apache.org/jira/browse/MAHOUT-964
>             Project: Mahout
>          Issue Type: Improvement
>          Components: Math
>    Affects Versions: 0.6
>         Environment: Mahout 0.6 snapshot from trunk
>            Reporter: Suneel Marthi
>         Attachments: Mahout-964.patch
>
>
> 1. If an invalid Similarity Measure has been specified as input to the RowSimilarityJob,
it presently throws a ClassCastException but still proceeds with executing all of the subsequent
tasks - VectorNormalizer, Cooccurrences Mapper and UnSymmetrify Mapper. We should exit the
process early without having to invoke all of the subsequent tasks (all of them fail anyways).
> 2. It would be nice to have an --overwrite option for the Command line interface which
would delete the temp and output paths at the beginning of RowSimilarityJob execution, similar
to what's being done in seq2sparse, seqdirectory. If I run RowSimilarityJob over and over
again with different similarity measures, I should not be forced to delete my temp and output
paths first prior to invoking the job.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message