[jira] Updated: (MAHOUT-167) Convert code to Hadoop 0.20 API

["Sean Owen (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/MAHOUT-167?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Sean Owen updated MAHOUT-167:
-----------------------------

    Fix Version/s: 0.5
                       (was: 0.4)

This is good progress. The remaining code that has not been converted are:

Bayes
CBayes
Lanczos/Eigen/SVD bits
Transpose/MatrixMult/etc Job

Still some way to go. What's the right way to think about this? Can this code be converted, superseded? 
I tried myself but it's a lot of change and I'm not confident.

Ideally these get rewritten in terms of AbstractJob which just takes care of the integration points with Hadoop's new APIs. Is that feasible?

> Convert code to Hadoop 0.20 API
> -------------------------------
>
>                 Key: MAHOUT-167
>                 URL: https://issues.apache.org/jira/browse/MAHOUT-167
>             Project: Mahout
>          Issue Type: Improvement
>          Components: Classification
>    Affects Versions: 0.3
>            Reporter: Jeff Eastman
>             Fix For: 0.5
>
>         Attachments: jeastman.vcf, MAHOUT-167.patch, MAHOUT-167.patch, MAHOUT-167e.patch, MAHOUT-167e.patch, MAHOUT-167u.patch
>
>
> We need to update the various implementations to remove the deprecated Hadoop API calls.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.