[SPARK-7380][MLLIB] pipeline stages should be copyable in Python by mengxr · Pull Request #6088 · apache/spark

mengxr · 2015-05-12T18:20:33Z

This PR makes pipeline stages in Python copyable and hence simplifies some implementations. It also includes the following changes:

Rename paramMap and defaultParamMap to _paramMap and _defaultParamMap, respectively.
Accept a list of param maps in fit.
Use parent uid and name to identify param.

@jkbradley

AmplabJenkins · 2015-05-12T18:22:11Z

Merged build triggered.

AmplabJenkins · 2015-05-12T18:22:18Z

Merged build started.

SparkQA · 2015-05-12T18:23:03Z

Test build #32521 has started for PR 6088 at commit 9630eae.

SparkQA · 2015-05-12T18:24:47Z

Test build #32521 has finished for PR 6088 at commit 9630eae.

This patch fails Python style tests.
This patch merges cleanly.
This patch adds no public classes.

AmplabJenkins · 2015-05-12T18:24:48Z

Merged build finished. Test FAILed.

AmplabJenkins · 2015-05-12T18:24:49Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/32521/
Test FAILed.

AmplabJenkins · 2015-05-12T19:27:11Z

Merged build triggered.

AmplabJenkins · 2015-05-12T19:27:19Z

Merged build started.

SparkQA · 2015-05-12T19:28:01Z

Test build #32527 has started for PR 6088 at commit a163413.

SparkQA · 2015-05-12T21:11:24Z

Test build #32527 has finished for PR 6088 at commit a163413.

This patch passes all tests.
This patch merges cleanly.
This patch adds the following public classes (experimental):
- decisionTreeCode = '''class DecisionTreeParams(Params):
- class DecisionTreeParams(Params):
- class LinearRegression(JavaEstimator, HasFeaturesCol, HasLabelCol, HasPredictionCol, HasMaxIter,
- class LinearRegressionModel(JavaModel):
- class TreeRegressorParams(object):
- class RandomForestParams(object):
- class GBTParams(object):
- class DecisionTreeRegressor(JavaEstimator, HasFeaturesCol, HasLabelCol, HasPredictionCol,
- class DecisionTreeRegressionModel(JavaModel):
- class RandomForestRegressor(JavaEstimator, HasFeaturesCol, HasLabelCol, HasPredictionCol, HasSeed,
- class RandomForestRegressionModel(JavaModel):
- class GBTRegressor(JavaEstimator, HasFeaturesCol, HasLabelCol, HasPredictionCol, HasMaxIter,
- class GBTRegressionModel(JavaModel):
- s"FileOutputCommitter or its subclass is expected, but got a $
- trait FSBasedRelationProvider
- abstract class OutputWriter

AmplabJenkins · 2015-05-12T21:11:29Z

Merged build finished. Test PASSed.

AmplabJenkins · 2015-05-12T21:11:29Z

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/32527/
Test PASSed.

jkbradley · 2015-05-13T00:18:44Z

I realized that, while I believe this works, it will make a deep copy of the Estimator and Evaluator, which is a little inefficient. Not that important

jkbradley · 2015-05-13T00:19:51Z

It looks good, although I agree it may need more testing to make sure UIDs behave as expected.

One other thought: Do you think it would be worthwhile to try to use the same uid for a Java*Wrapper instance in Python and the Scala instance it is wrapping?

mengxr · 2015-05-13T01:18:26Z

Yes, that certainly helps. Let's get #6019 in first, which should be ready for review.

AmplabJenkins · 2015-05-14T09:07:10Z

Merged build triggered.

AmplabJenkins · 2015-05-14T09:07:19Z

Merged build started.

SparkQA · 2015-05-14T09:09:18Z

Test build #32698 has started for PR 6088 at commit 46cb6ed.

SparkQA · 2015-05-14T10:58:11Z

Test build #32698 has finished for PR 6088 at commit 46cb6ed.

This patch passes all tests.
This patch merges cleanly.
This patch adds the following public classes (experimental):
- class Pipeline(override val uid: String) extends Estimator[PipelineModel]
- final class DecisionTreeClassifier(override val uid: String)
- final class GBTClassifier(override val uid: String)
- class LogisticRegression(override val uid: String)
- final class OneVsRest(override val uid: String)
- final class RandomForestClassifier(override val uid: String)
- class BinaryClassificationEvaluator(override val uid: String)
- final class Binarizer(override val uid: String)
- final class Bucketizer(override val uid: String)
- class ElementwiseProduct(override val uid: String)
- class HashingTF(override val uid: String) extends UnaryTransformer[Iterable[_], Vector, HashingTF]
- final class IDF(override val uid: String) extends Estimator[IDFModel] with IDFBase
- class Normalizer(override val uid: String) extends UnaryTransformer[Vector, Vector, Normalizer]
- class OneHotEncoder(override val uid: String)
- class PolynomialExpansion(override val uid: String)
- class StandardScaler(override val uid: String) extends Estimator[StandardScalerModel]
- class StringIndexer(override val uid: String) extends Estimator[StringIndexerModel]
- class Tokenizer(override val uid: String) extends UnaryTransformer[String, Seq[String], Tokenizer]
- class RegexTokenizer(override val uid: String)
- class VectorAssembler(override val uid: String)
- class VectorIndexer(override val uid: String) extends Estimator[VectorIndexerModel]
- final class Word2Vec(override val uid: String) extends Estimator[Word2VecModel] with Word2VecBase
- class Param[T](val parent: String, val name: String, val doc: String, val isValid: T => Boolean)
- class DoubleParam(parent: String, name: String, doc: String, isValid: Double => Boolean)
- class IntParam(parent: String, name: String, doc: String, isValid: Int => Boolean)
- class FloatParam(parent: String, name: String, doc: String, isValid: Float => Boolean)
- class LongParam(parent: String, name: String, doc: String, isValid: Long => Boolean)
- class BooleanParam(parent: String, name: String, doc: String) // No need for isValid
- class ALS(override val uid: String) extends Estimator[ALSModel] with ALSParams
- final class DecisionTreeRegressor(override val uid: String)
- final class GBTRegressor(override val uid: String)
- class LinearRegression(override val uid: String)
- final class RandomForestRegressor(override val uid: String)
- class CrossValidator(override val uid: String) extends Estimator[CrossValidatorModel]
- trait Identifiable

AmplabJenkins · 2015-05-14T10:58:16Z

Merged build finished. Test PASSed.

AmplabJenkins · 2015-05-17T18:58:13Z

Merged build triggered.

AmplabJenkins · 2015-05-17T19:01:29Z

Merged build finished. Test FAILed.

AmplabJenkins · 2015-05-17T19:01:29Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/32942/
Test FAILed.

AmplabJenkins · 2015-05-17T21:18:05Z

Merged build triggered.

AmplabJenkins · 2015-05-17T21:18:15Z

Merged build started.

SparkQA · 2015-05-17T21:21:28Z

Test build #32945 has started for PR 6088 at commit 611c719.

AmplabJenkins · 2015-05-17T21:32:13Z

Merged build triggered.

AmplabJenkins · 2015-05-17T21:32:19Z

Merged build started.

mengxr · 2015-05-18T04:19:30Z

test this please

AmplabJenkins · 2015-05-18T04:22:11Z

Merged build triggered.

AmplabJenkins · 2015-05-18T04:22:19Z

Merged build started.

SparkQA · 2015-05-18T04:22:39Z

Test build #32966 has started for PR 6088 at commit 413c463.

SparkQA · 2015-05-18T06:05:50Z

Test build #32966 has finished for PR 6088 at commit 413c463.

This patch passes all tests.
This patch merges cleanly.
This patch adds the following public classes (experimental):
- class VectorIndexerModel(JavaModel):

AmplabJenkins · 2015-05-18T06:05:55Z

Merged build finished. Test PASSed.

AmplabJenkins · 2015-05-18T06:05:56Z

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/32966/
Test PASSed.

jkbradley · 2015-05-18T17:45:11Z

LGTM

This PR makes pipeline stages in Python copyable and hence simplifies some implementations. It also includes the following changes: 1. Rename `paramMap` and `defaultParamMap` to `_paramMap` and `_defaultParamMap`, respectively. 2. Accept a list of param maps in `fit`. 3. Use parent uid and name to identify param. jkbradley Author: Xiangrui Meng <meng@databricks.com> Author: Joseph K. Bradley <joseph@databricks.com> Closes #6088 from mengxr/SPARK-7380 and squashes the following commits: 413c463 [Xiangrui Meng] remove unnecessary doc 4159f35 [Xiangrui Meng] Merge remote-tracking branch 'apache/master' into SPARK-7380 611c719 [Xiangrui Meng] fix python style 68862b8 [Xiangrui Meng] update _java_obj initialization 927ad19 [Xiangrui Meng] fix ml/tests.py 0138fc3 [Xiangrui Meng] update feature transformers and fix a bug in RegexTokenizer 9ca44fb [Xiangrui Meng] simplify Java wrappers and add tests c7d84ef [Xiangrui Meng] update ml/tests.py to test copy params 7e0d27f [Xiangrui Meng] merge master 46840fb [Xiangrui Meng] update wrappers b6db1ed [Xiangrui Meng] update all self.paramMap to self._paramMap 46cb6ed [Xiangrui Meng] merge master a163413 [Xiangrui Meng] fix style 1042e80 [Xiangrui Meng] Merge remote-tracking branch 'apache/master' into SPARK-7380 9630eae [Xiangrui Meng] fix Identifiable._randomUID 13bd70a [Xiangrui Meng] update ml/tests.py 64a536c [Xiangrui Meng] use _fit/_transform/_evaluate to simplify the impl 02abf13 [Xiangrui Meng] Merge remote-tracking branch 'apache/master' into copyable-python 66ce18c [Joseph K. Bradley] some cleanups before sending to Xiangrui 7431272 [Joseph K. Bradley] Rebased with master (cherry picked from commit 9c7e802) Signed-off-by: Xiangrui Meng <meng@databricks.com>

mengxr · 2015-05-18T19:02:48Z

Merged into master and branch-1.4.

This PR makes pipeline stages in Python copyable and hence simplifies some implementations. It also includes the following changes: 1. Rename `paramMap` and `defaultParamMap` to `_paramMap` and `_defaultParamMap`, respectively. 2. Accept a list of param maps in `fit`. 3. Use parent uid and name to identify param. jkbradley Author: Xiangrui Meng <meng@databricks.com> Author: Joseph K. Bradley <joseph@databricks.com> Closes apache#6088 from mengxr/SPARK-7380 and squashes the following commits: 413c463 [Xiangrui Meng] remove unnecessary doc 4159f35 [Xiangrui Meng] Merge remote-tracking branch 'apache/master' into SPARK-7380 611c719 [Xiangrui Meng] fix python style 68862b8 [Xiangrui Meng] update _java_obj initialization 927ad19 [Xiangrui Meng] fix ml/tests.py 0138fc3 [Xiangrui Meng] update feature transformers and fix a bug in RegexTokenizer 9ca44fb [Xiangrui Meng] simplify Java wrappers and add tests c7d84ef [Xiangrui Meng] update ml/tests.py to test copy params 7e0d27f [Xiangrui Meng] merge master 46840fb [Xiangrui Meng] update wrappers b6db1ed [Xiangrui Meng] update all self.paramMap to self._paramMap 46cb6ed [Xiangrui Meng] merge master a163413 [Xiangrui Meng] fix style 1042e80 [Xiangrui Meng] Merge remote-tracking branch 'apache/master' into SPARK-7380 9630eae [Xiangrui Meng] fix Identifiable._randomUID 13bd70a [Xiangrui Meng] update ml/tests.py 64a536c [Xiangrui Meng] use _fit/_transform/_evaluate to simplify the impl 02abf13 [Xiangrui Meng] Merge remote-tracking branch 'apache/master' into copyable-python 66ce18c [Joseph K. Bradley] some cleanups before sending to Xiangrui 7431272 [Joseph K. Bradley] Rebased with master

jkbradley and others added 6 commits May 11, 2015 12:51

Rebased with master

7431272

some cleanups before sending to Xiangrui

66ce18c

Merge remote-tracking branch 'apache/master' into copyable-python

02abf13

use _fit/_transform/_evaluate to simplify the impl

64a536c

update ml/tests.py

13bd70a

fix Identifiable._randomUID

9630eae

Merge remote-tracking branch 'apache/master' into SPARK-7380

1042e80

mengxr changed the title ~~[SPARK-7380][MLLIB] pipeline stages should be copyable in Python~~ [WIP][SPARK-7380][MLLIB] pipeline stages should be copyable in Python May 12, 2015

fix style

a163413

jkbradley reviewed May 13, 2015
View reviewed changes

merge master

46cb6ed

mengxr added 4 commits May 17, 2015 09:51

simplify Java wrappers and add tests

9ca44fb

update feature transformers and fix a bug in RegexTokenizer

0138fc3

fix ml/tests.py

927ad19

update _java_obj initialization

68862b8

mengxr changed the title ~~[WIP][SPARK-7380][MLLIB] pipeline stages should be copyable in Python~~ [SPARK-7380][MLLIB] pipeline stages should be copyable in Python May 17, 2015

fix python style

611c719

mengxr added 2 commits May 17, 2015 14:27

Merge remote-tracking branch 'apache/master' into SPARK-7380

4159f35

remove unnecessary doc

413c463

asfgit closed this in 9c7e802 May 18, 2015

Conversation

mengxr commented May 12, 2015

Uh oh!

AmplabJenkins commented May 12, 2015

Uh oh!

AmplabJenkins commented May 12, 2015

Uh oh!

SparkQA commented May 12, 2015

Uh oh!

SparkQA commented May 12, 2015

Uh oh!

AmplabJenkins commented May 12, 2015

Uh oh!

AmplabJenkins commented May 12, 2015

Uh oh!

AmplabJenkins commented May 12, 2015

Uh oh!

AmplabJenkins commented May 12, 2015

Uh oh!

SparkQA commented May 12, 2015

Uh oh!

SparkQA commented May 12, 2015

Uh oh!

AmplabJenkins commented May 12, 2015

Uh oh!

AmplabJenkins commented May 12, 2015

Uh oh!

jkbradley May 13, 2015

Choose a reason for hiding this comment

Uh oh!

jkbradley commented May 13, 2015

Uh oh!

mengxr commented May 13, 2015

Uh oh!

AmplabJenkins commented May 14, 2015

Uh oh!

AmplabJenkins commented May 14, 2015

Uh oh!

SparkQA commented May 14, 2015

Uh oh!

SparkQA commented May 14, 2015

Uh oh!

AmplabJenkins commented May 14, 2015

Uh oh!

AmplabJenkins commented May 17, 2015

Uh oh!

AmplabJenkins commented May 17, 2015

Uh oh!

AmplabJenkins commented May 17, 2015

Uh oh!

AmplabJenkins commented May 17, 2015

Uh oh!

AmplabJenkins commented May 17, 2015

Uh oh!

SparkQA commented May 17, 2015

Uh oh!

AmplabJenkins commented May 17, 2015

Uh oh!

AmplabJenkins commented May 17, 2015

Uh oh!

mengxr commented May 18, 2015

Uh oh!

AmplabJenkins commented May 18, 2015

Uh oh!

AmplabJenkins commented May 18, 2015

Uh oh!

SparkQA commented May 18, 2015

Uh oh!

SparkQA commented May 18, 2015

Uh oh!

AmplabJenkins commented May 18, 2015

Uh oh!

AmplabJenkins commented May 18, 2015

Uh oh!

jkbradley commented May 18, 2015

Uh oh!

mengxr commented May 18, 2015

Uh oh!

Reviewers

Assignees