GitHub user viirya reopened a pull request:
https://github.com/apache/spark/pull/19442
[SPARK-8515][ML][WIP] Improve ML Attribute API
## What changes were proposed in this pull request?
The current ML attribute API has issues like inefficiency and not easy to
use. This work tries to improve this API with main changes:
* Support spark vector-typed attributes.
* Simplify vector-typed attribute serialization.
* Keep minimum APIs to support ML attributes.
** THIS WORK is not ready and is working in progress.
## How was this patch tested?
Added tests.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/viirya/spark-1 SPARK-8515
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/spark/pull/19442.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 #19442
----
commit 77d657d8bc8102081e4b0d7b5d42a256e64514d4
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-02T15:03:54Z
Init design of ml attribute.
commit 7837778e7cbbf83851b1a2b5047f4e6a8039f809
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-03T15:03:31Z
revise.
commit 97f6848f0cbb1a76b4434930ce8938da50eaafbe
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-03T15:14:02Z
revise.
commit 2e3a3541fc7a59ac63b2118228de8015c238de40
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-04T05:15:58Z
revise.
commit 0d76eac84f5837aefebc763687fa9c5c7e1aeb4d
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-04T15:07:57Z
revise.
commit 81cca5cccfa2556ff0bba5a73764d3f503040b13
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-05T04:30:48Z
revise.
commit 4813fe8a4bd19a02b7b6bff138f04e7e50f7cdd7
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-05T06:15:53Z
revise.
commit 7951f59027418962ad95465e439bff41876ecfa8
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-05T07:51:50Z
revise.
commit a381af3edf52132086af64360789cb3a7d20d61e
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-05T09:00:02Z
Add builder and test.
commit f25c89dbded0eb9dce25d8da63a1a1aa49ad459f
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-05T15:10:11Z
revise test.
commit 7e237f38088f2375f40f9a4c97aee2e6acd54328
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-06T02:46:07Z
Add new test.
commit 77ced957e7be2169ac0c59c76f60ab9d4fcac3ef
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-06T03:57:12Z
Add more tests.
commit de0aa76199141255258d9d5b12a0d31b1758c6f1
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-06T06:17:29Z
revise.
commit d828cf3d3b13a2b2b1990bdff9593b49e53f6cf9
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-06T13:55:41Z
Add java-friendly APIs for attribute types.
commit 5844fbaef5d5825eafadb7c53196fb2132937e4e
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-09T03:24:26Z
Revise APIs.
commit da0fcef7d3370ebca97d200f01e9f2814a9ed755
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-09T03:26:15Z
revise.
commit 66be26cd7f25614137cfb9722f859f36d9f80c0c
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-09T03:47:43Z
Add default constructors to attribute types.
commit ce80ed5b693745fa4a650e508c6cd9e24350c52e
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-10T12:52:22Z
Use Array instead of Seq in APIs.
commit 2b94dd5c192b1d9302e24c0392fc9a5aaaedb596
Author: Liang-Chi Hsieh <[email protected]>
Date: 2017-10-14T00:21:04Z
Add more compatibility tests.
----
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
