[ 
https://issues.apache.org/jira/browse/MAHOUT-1570?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15193654#comment-15193654
 ] 

ASF GitHub Bot commented on MAHOUT-1570:
----------------------------------------

GitHub user andrewpalumbo reopened a pull request:

    https://github.com/apache/mahout/pull/187

    MAHOUT-1570: Flink binding b: to review.. not for merge

    still needs to be rebased on top of  Mahout 0.11.2 master possibly.

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/andrewpalumbo/mahout flink-binding-b

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/mahout/pull/187.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #187
    
----
commit b86115c44a5ef998cc073f003ce79472a6b9005e
Author: Till Rohrmann <trohrm...@apache.org>
Date:   2015-11-16T16:58:36Z

    Add type information to flink-bindings

commit 6006e7730723f910bbf3ae7b4cfcdb732411a3ce
Author: Till Rohrmann <trohrm...@apache.org>
Date:   2015-11-16T17:14:47Z

    Remove flinkbindings.blas package object

commit e465b5e68faa3a7549b69e6df566019ea2f01624
Author: Suneel Marthi <smar...@apache.org>
Date:   2015-11-16T17:39:55Z

    Merge pull request #1 from tillrohrmann/flink-binding
    
    Add type information to flink bindings

commit a41d0bda894224b238da246ef222c1ca1ce101a2
Author: smarthi <smar...@apache.org>
Date:   2015-11-12T06:10:54Z

    NOJira: Add missing license header

commit 1c6f73bb747e2072c0af05aa105ca080a998c0f7
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-01-10T19:09:39Z

    fix for FlinkopAtB

commit 1ae78e334b85ff90278d153f925539a8651487c7
Author: smarthi <smar...@apache.org>
Date:   2016-01-10T22:47:36Z

    Modified to use Flink Scala Api, fixed FlinkOpABt

commit 1ec2047cb0f1e400f3211cec6f65cb7354ac3bb6
Author: smarthi <smar...@apache.org>
Date:   2016-01-10T23:18:43Z

    Implemented Sampling methods, other minor fixes

commit f1437460899374648eb3f8fd96c19003c43bcd53
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-01-11T01:10:43Z

    Merge branch 'flink-binding' of https://github.com/smarthi/mahout into 
flink-binding-suneel-2016
    
    checked out --theirs
    
    Conflicts:
        flink/src/main/scala/org/apache/mahout/flinkbindings/FlinkEngine.scala
        
flink/src/main/scala/org/apache/mahout/flinkbindings/blas/FlinkOpAtA.scala
        
flink/src/main/scala/org/apache/mahout/flinkbindings/blas/FlinkOpAtB.scala

commit f457ad44d1cc4670d969ccb771fde30b90525952
Author: smarthi <smar...@apache.org>
Date:   2016-01-16T03:31:20Z

    Code reformatting, added TypeInformation check

commit a351af12f0f0e9381504cae9969f3ffe349d5d6f
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-01-16T03:36:08Z

    Merge branch 'flink-binding' of https://github.com/smarthi/mahout into 
flink-bindings-b

commit dcfea65fe3ddb6b438558ad1d48d3fb1c8ce04b9
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-01-16T23:15:09Z

    Allow 'Any' key type infomation for generateTypeInformation[K: ClassTag]: 
TypeInformation[K]

commit b94c045160b29b55e08ec1f7025d4d60fb8ae82c
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-01-17T20:51:17Z

    wip

commit a0a319643964cdf0fdded31348104014bd161c36
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-02-06T19:56:37Z

    (1) Refactored flinks Hadoop1Utils into an Hadoop2Utils opject to reflect 
their retirement of hadoop 1 API.
    (2) wip: flink's checkpoint action seems to not be working; some trial and 
error. and a note added in the math-scala base tests to note this problem. ie. 
what shoul be deterministic tests are not.
     (3) this could possibly be an error due to flink rebalancing 
pre-checkpoint action.. though unlikely.
    (4) does not compile with today's flink 1.0-SNAPSHOT due to some changes in 
exctracting type information made in the last week or so.

commit cbac921171deef7ea559581adae31ed27ccc65d2
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-02-06T20:54:56Z

    Use Object instead of Any

commit 957e9f194f9fbd728c36f2f715f62561b2471a2a
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-02-06T23:10:59Z

    upgrade kryo dep in math-scala to match that used by flink.  still getting 
OOM errors in tests

commit c82ed88812dfe324a6829ebdf3c25a8c58b67b13
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-02-07T02:08:11Z

    clean up,
    make sure flink module is using correct kryo version,
    use 'ProblemTestSuite' as example of odd problems

commit 22f456613aa18f16ab35f6861ed279dc8bb1b020
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-02-07T06:30:47Z

    replace accidentaly deleted line

commit e0da8bba715d38b87f1d08353c857161dee75c47
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-02-09T02:27:55Z

    wip: drmDfsWrite: explicitly define all Writable Subclasses for ds.map() 
keys

commit 759d024f2f45b232e333b9fd742902d8401ff16b
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-03-02T21:10:11Z

    Include java.io.serilazable interface in Martix and Vector Kryo serilizers. 
 Update flink pom for flink 1.0 release module renaming.  wip: failing 
stackOverFlow on FlinkOpTimesRightMatrix

commit 44a36d7b1c0350b6cda731f1b35f5593e5a31cad
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-03-02T21:44:11Z

    Hack to fix kryo stackOverFlow when using native flink broadcasting of a 
Mahout matrix in FlinkOpTimesRight

commit dc9104ca064c1f35f4d31dfa1a9c338308566f2a
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-03-03T02:14:32Z

    set the environment's parallelism to 10 in the DistributedFlinkTestSuite so 
that tests like dsqDist(X,Y) do not fail witl not enough slots

commit b14f38880be9407395b3d3009b831066ef605c65
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-03-03T02:19:34Z

    remove all instances of @runwith[JunitRunner] from tests

commit 1079d32f78d8cdaabd8e10d366347e576a962838
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-03-03T02:22:21Z

    Bump flink to 1.1-SNAPSHOT

commit 935e9eb0f743c1170913fa7090a203af6c7c993e
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-03-03T03:36:31Z

    wip, move all 6 failing tests into FailingTestSuite

commit dfb1fae5dea753e9bfc62f3f84fc20cf3c12387d
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-03-03T04:37:54Z

    wip: intwritable expected to be longwritable

commit f9c39ea355aaf5b81d3c14f3e9c1d81bc7f3c48e
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-03-03T22:26:35Z

    isolate a skingle key type to show error with drm.dfsWrite()

commit 73f8a70a6ba356427f85bc31c74d1a8499fa947a
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-03-11T00:47:02Z

    some clean up

commit f05418cef87ff8f2d9efa59ce3a71f9f84181fda
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-03-11T01:14:07Z

    set flink version to 1.0.0

commit a06f02d426d96a36800d72206692d03c68f8805f
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-03-11T03:09:52Z

    write a simple (Int,Int) unit test for flink hadoop output

commit 29fe274eeaf6f2cb8d01bd6b75627ed70b624291
Author: Andrew Palumbo <apalu...@apache.org>
Date:   2016-03-14T01:30:41Z

    fix for failing Writables

----


> Adding support for Apache Flink as a backend for the Mahout DSL
> ---------------------------------------------------------------
>
>                 Key: MAHOUT-1570
>                 URL: https://issues.apache.org/jira/browse/MAHOUT-1570
>             Project: Mahout
>          Issue Type: Improvement
>            Reporter: Till Rohrmann
>            Assignee: Suneel Marthi
>              Labels: DSL, flink, scala
>             Fix For: 0.12.0
>
>
> With the finalized abstraction of the Mahout DSL plans from the backend 
> operations (MAHOUT-1529), it should be possible to integrate further backends 
> for the Mahout DSL. Apache Flink would be a suitable candidate to act as a 
> good execution backend. 
> With respect to the implementation, the biggest difference between Spark and 
> Flink at the moment is probably the incremental rollout of plans, which is 
> triggered by Spark's actions and which is not supported by Flink yet. 
> However, the Flink community is working on this issue. For the moment, it 
> should be possible to circumvent this problem by writing intermediate results 
> required by an action to HDFS and reading from there.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to