Author: Reynold Xin <rxin@apache.org>

Closes #443 from rxin/readme and squashes the following commits:

16853de [Reynold Xin] Updated SBT and Scala instructions.
3ac3ceb [Reynold Xin] README update

...AsNewAPIHadoopDataset

`writer.close` should be put in the `finally` block to avoid potential resource leaks.

JIRA: https://issues.apache.org/jira/browse/SPARK-1482

Author: zsxwing <zsxwing@gmail.com>

Closes #400 from zsxwing/SPARK-1482 and squashes the following commits:

06b197a [zsxwing] SPARK-1482: Fix potential resource leaks in saveAsHadoopDataset and saveAsNewAPIHadoopDataset

This doesn't require creating new Ordering objects per row.  Additionally, [view bounds are going to be deprecated](https://issues.scala-lang.org/browse/SI-7629), so we should get rid of them while APIs are still flexible.

Author: Michael Armbrust <michael@databricks.com>

Closes #410 from marmbrus/viewBounds and squashes the following commits:

c574221 [Michael Armbrust] fix example.
812008e [Michael Armbrust] Update Java API.
1b9b85c [Michael Armbrust] Update scala doc.
35798a8 [Michael Armbrust] Remove view bounds on Ordered in favor of a context bound on Ordering.

Author: Reynold Xin <rxin@apache.org>

Closes #444 from rxin/pyspark and squashes the following commits:

fc11356 [Reynold Xin] Made the PySpark shell version checking compatible with Python 2.6.
571830b [Reynold Xin] Fixed broken pyspark shell.

Actually it is separated from https://github.com/apache/spark/pull/85 as suggested by @rxin

compare

https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/util/AkkaUtils.scala#L122

and

https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/util/AkkaUtils.scala#L117

the first one use get and then toLong, the second one getLong....better to make them consistent

very very small fix........

Author: CodingCat <zhunansjtu@gmail.com>

Closes #434 from CodingCat/SPARK-1523 and squashes the following commits:

0e86f3f [CodingCat] improve the readability of code in AkkaUtil

Per discussion, this is my suggestion to make ALS Rating, ClassificationModel, RegressionModel experimental for now, to reserve the right to possibly change after 1.0. See what you think of this much.

Author: Sean Owen <sowen@cloudera.com>

Closes #372 from srowen/SPARK-1357Addendum and squashes the following commits:

17cf1ea [Sean Owen] Remove (another) blank line after ":: Experimental ::"
6800e4c [Sean Owen] Remove blank line after ":: Experimental ::"
b3a88d2 [Sean Owen] Make ALS Rating, ClassificationModel, RegressionModel experimental for now, to reserve the right to possibly change after 1.0

A quick fix for https://issues.apache.org/jira/browse/SPARK-1520

By excluding fastutil, we bring the number of files in the assembly jar back under 65536, so Java 7 won't create the assembly jar in zip64 format, which cannot be read by Java 6.

With this change, the assembly jar now has about 60000 entries (58000 files), tested with both sbt and maven.

Author: Xiangrui Meng <meng@databricks.com>

Closes #437 from mengxr/remove-fastutil and squashes the following commits:

00f9beb [Xiangrui Meng] remove fastutil from dependencies

Author: Cheng Lian <lian.cs.zju@gmail.com>

Closes #432 from liancheng/reuseRow and squashes the following commits:

9e6d083 [Cheng Lian] Simplified code with BufferedIterator
52acec9 [Cheng Lian] Reuses Row object in ExistingRdd.productToRowRdd()

https://issues.apache.org/jira/browse/SPARK-1483

From the original JIRA: " The parameter name is part of the public API in Scala and Python, since you can pass named parameters to a method, so we should name it to this more descriptive term. Everywhere else we refer to "splits" as partitions." - @mateiz

Author: CodingCat <zhunansjtu@gmail.com>

Closes #430 from CodingCat/SPARK-1483 and squashes the following commits:

4b60541 [CodingCat] deprecate defaultMinSplits
ba2c663 [CodingCat] Rename minSplits to minPartitions in public APIs

This is currently causing many builds to hang.

https://issues.apache.org/jira/browse/SPARK-1530

Author: Patrick Wendell <pwendell@gmail.com>

Closes #440 from pwendell/uitest-fix and squashes the following commits:

9a143dc [Patrick Wendell] Ignore streaming UI test

This will make the tests more stable when not running SQL tests.

Author: Patrick Wendell <pwendell@gmail.com>

Closes #439 from pwendell/hive-tests and squashes the following commits:

88a6032 [Patrick Wendell] FIX: Don't build Hive in assembly unless running Hive tests.

...finishes

Note this is dependent on https://github.com/apache/spark/pull/204 to have a working history server, but there are no code dependencies.

This also fixes SPARK-1288 yarn stable finishApplicationMaster incomplete. Since I was in there I made the diagnostic message be passed properly.

Author: Thomas Graves <tgraves@apache.org>

Closes #362 from tgravescs/SPARK-1408 and squashes the following commits:

ec89705 [Thomas Graves] Fix typo.
446122d [Thomas Graves] Make config yarn specific
f5d5373 [Thomas Graves] SPARK-1408 Modify Spark on Yarn to point to the history server when app finishes

This only works for the three paths defined in the environment
(SPARK_JAR, SPARK_YARN_APP_JAR and SPARK_LOG4J_CONF).

Tested by running SparkPi with local: and file: URIs against Yarn cluster (no "upload" shows up in logs in the local case).

Author: Marcelo Vanzin <vanzin@cloudera.com>

Closes #303 from vanzin/yarn-local and squashes the following commits:

82219c1 [Marcelo Vanzin] [SPARK-1395] Allow "local:" URIs to work on Yarn.

Python alternative for https://github.com/apache/spark/pull/392; managed from shell.py

Author: AbhishekKr <abhikumar163@gmail.com>

Closes #399 from abhishekkr/pyspark_shell and squashes the following commits:

134bdc9 [AbhishekKr] pyspark require Python2, failing if system default is Py3 from shell.py

This will also fix SPARK-1464: Update MLLib Examples to Use Breeze.

Author: Sandeep <sandeep@techaddict.me>

Closes #416 from techaddict/1462 and squashes the following commits:

a43638e [Sandeep] Some Style Changes
3ce69c3 [Sandeep] Fix Ordering and Naming of Imports in Examples
6c7e543 [Sandeep] SPARK-1462: Examples of ML algorithms are using deprecated APIs

It is very confusing when your code throws an exception, but the only stack trace show is in the DAGScheduler.  This is a simple patch to include the stack trace for the actual failure in the error message.  Suggestions on formatting welcome.

Before:
```
scala> sc.parallelize(1 :: Nil).map(_ => sys.error("Ahh!")).collect()
org.apache.spark.SparkException: Job aborted due to stage failure: Task 0.0:3 failed 1 times (most recent failure: Exception failure in TID 3 on host localhost: java.lang.RuntimeException: Ahh!)
	at org.apache.spark.scheduler.DAGScheduler.org$apache$spark$scheduler$DAGScheduler$$failJobAndIndependentStages(DAGScheduler.scala:1055)
	at org.apache.spark.scheduler.DAGScheduler$$anonfun$org$apache$spark$scheduler$DAGScheduler$$abortStage$1.apply(DAGScheduler.scala:1039)
	at org.apache.spark.scheduler.DAGScheduler$$anonfun$org$apache$spark$scheduler$DAGScheduler$$abortStage$1.apply(DAGScheduler.scala:1037)
...
```

After:
```
org.apache.spark.SparkException: Job aborted due to stage failure: Task 0.0:3 failed 1 times, most recent failure: Exception failure in TID 3 on host localhost: java.lang.RuntimeException: Ahh!
        scala.sys.package$.error(package.scala:27)
        $iwC$$iwC$$iwC$$iwC$$anonfun$1.apply(<console>:13)
        $iwC$$iwC$$iwC$$iwC$$anonfun$1.apply(<console>:13)
        scala.collection.Iterator$$anon$11.next(Iterator.scala:328)
        scala.collection.Iterator$class.foreach(Iterator.scala:727)
        scala.collection.AbstractIterator.foreach(Iterator.scala:1157)
        scala.collection.generic.Growable$class.$plus$plus$eq(Growable.scala:48)
        scala.collection.mutable.ArrayBuffer.$plus$plus$eq(ArrayBuffer.scala:103)
        scala.collection.mutable.ArrayBuffer.$plus$plus$eq(ArrayBuffer.scala:47)
        scala.collection.TraversableOnce$class.to(TraversableOnce.scala:273)
        scala.collection.AbstractIterator.to(Iterator.scala:1157)
        scala.collection.TraversableOnce$class.toBuffer(TraversableOnce.scala:265)
        scala.collection.AbstractIterator.toBuffer(Iterator.scala:1157)
        scala.collection.TraversableOnce$class.toArray(TraversableOnce.scala:252)
        scala.collection.AbstractIterator.toArray(Iterator.scala:1157)
        org.apache.spark.rdd.RDD$$anonfun$6.apply(RDD.scala:676)
        org.apache.spark.rdd.RDD$$anonfun$6.apply(RDD.scala:676)
        org.apache.spark.SparkContext$$anonfun$runJob$5.apply(SparkContext.scala:1048)
        org.apache.spark.SparkContext$$anonfun$runJob$5.apply(SparkContext.scala:1048)
        org.apache.spark.scheduler.ResultTask.runTask(ResultTask.scala:110)
        org.apache.spark.scheduler.Task.run(Task.scala:50)
        org.apache.spark.executor.Executor$TaskRunner$$anonfun$run$1.apply$mcV$sp(Executor.scala:211)
        org.apache.spark.deploy.SparkHadoopUtil.runAsUser(SparkHadoopUtil.scala:46)
        org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:176)
        java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
        java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
        java.lang.Thread.run(Thread.java:744)
Driver stacktrace:
	at org.apache.spark.scheduler.DAGScheduler.org$apache$spark$scheduler$DAGScheduler$$failJobAndIndependentStages(DAGScheduler.scala:1055)
	at org.apache.spark.scheduler.DAGScheduler$$anonfun$org$apache$spark$scheduler$DAGScheduler$$abortStage$1.apply(DAGScheduler.scala:1039)
	at org.apache.spark.scheduler.DAGScheduler$$anonfun$org$apache$spark$scheduler$DAGScheduler$$abortStage$1.apply(DAGScheduler.scala:1037)
	at scala.collection.mutable.ResizableArray$class.foreach(ResizableArray.scala:59)
	at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:47)
	at org.apache.spark.scheduler.DAGScheduler.org$apache$spark$scheduler$DAGScheduler$$abortStage(DAGScheduler.scala:1037)
	at org.apache.spark.scheduler.DAGScheduler$$anonfun$processEvent$10.apply(DAGScheduler.scala:614)
	at org.apache.spark.scheduler.DAGScheduler$$anonfun$processEvent$10.apply(DAGScheduler.scala:614)
	at scala.Option.foreach(Option.scala:236)
	at org.apache.spark.scheduler.DAGScheduler.processEvent(DAGScheduler.scala:614)
	at org.apache.spark.scheduler.DAGScheduler$$anonfun$start$1$$anon$2$$anonfun$receive$1.applyOrElse(DAGScheduler.scala:143)
	at akka.actor.ActorCell.receiveMessage(ActorCell.scala:498)
	at akka.actor.ActorCell.invoke(ActorCell.scala:456)
	at akka.dispatch.Mailbox.processMailbox(Mailbox.scala:237)
	at akka.dispatch.Mailbox.run(Mailbox.scala:219)
	at akka.dispatch.ForkJoinExecutorConfigurator$AkkaForkJoinTask.exec(AbstractDispatcher.scala:386)
	at scala.concurrent.forkjoin.ForkJoinTask.doExec(ForkJoinTask.java:260)
	at scala.concurrent.forkjoin.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:1339)
	at scala.concurrent.forkjoin.ForkJoinPool.runWorker(ForkJoinPool.java:1979)
	at scala.concurrent.forkjoin.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:107)
```

Author: Michael Armbrust <michael@databricks.com>

Closes #409 from marmbrus/stacktraces and squashes the following commits:

3e4eb65 [Michael Armbrust] indent. include header for driver stack trace.
018b06b [Michael Armbrust] Include stack trace for exceptions in user code.

change  _slideDuration  to   _windowDuration

Author: baishuo(白硕) <vc_java@hotmail.com>

Closes #425 from baishuo/master and squashes the following commits:

6f09ea1 [baishuo(白硕)] Update ReducedWindowedDStream.scala

"By default, this uses only 8 parallel tasks to do the grouping." is a big misleading. Please refer to https://github.com/apache/spark/pull/389

detail is as following code :

  def defaultPartitioner(rdd: RDD[_], others: RDD[_]*): Partitioner = {
    val bySize = (Seq(rdd) ++ others).sortBy(_.partitions.size).reverse
    for (r <- bySize if r.partitioner.isDefined) {
      return r.partitioner.get
    }
    if (rdd.context.conf.contains("spark.default.parallelism")) {
      new HashPartitioner(rdd.context.defaultParallelism)
    } else {
      new HashPartitioner(bySize.head.partitions.size)
    }
  }

Author: Chen Chao <crazyjvm@gmail.com>

Closes #403 from CrazyJvm/patch-4 and squashes the following commits:

42f6c9e [Chen Chao] fix format
829a995 [Chen Chao] fix format
1568336 [Chen Chao] misleading task number of groupByKey

Author: Kan Zhang <kzhang@apache.org>

Closes #401 from kanzhang/fix-1475 and squashes the following commits:

c6058bd [Kan Zhang] Fixing a race condition in event listener unit test

delete semicolon

Author: Chen Chao <crazyjvm@gmail.com>

Closes #411 from CrazyJvm/patch-5 and squashes the following commits:

72333a3 [Chen Chao] remove unnecessary brace
de5d9a7 [Chen Chao] style fix

Each vertex partition is co-located with a pid2vid array created in RoutingTable.scala. This array maps edge partition IDs to the list of vertices in the current vertex partition that are mentioned by edges in that partition. Therefore the pid2vid array should have one entry per edge partition.

GraphX currently creates one entry per *vertex* partition, which is a bug that leads to an ArrayIndexOutOfBoundsException when there are more edge partitions than vertex partitions. This commit fixes the bug and adds a test for this case.

Resolves SPARK-1329. Thanks to Daniel Darabos for reporting this bug.

Author: Ankur Dave <ankurdave@gmail.com>

Closes #368 from ankurdave/fix-pid2vid-size and squashes the following commits:

5a5c52a [Ankur Dave] SPARK-1329: Create pid2vid with correct number of partitions

GraphImpl.reverse used to reverse edges in each partition of the edge RDD but preserve the routing table and replicated vertex view, since reversing should not affect partitioning.

However, the old routing table would then have incorrect information for srcAttrOnly and dstAttrOnly. These RDDs should be switched.

A simple fix is for Graph.reverse to rebuild the routing table and replicated vertex view.

Thanks to Bogdan Ghidireac for reporting this issue on the [mailing list](http://apache-spark-user-list.1001560.n3.nabble.com/graph-reverse-amp-Pregel-API-td4338.html).

Author: Ankur Dave <ankurdave@gmail.com>

Closes #431 from ankurdave/fix-reverse-bug and squashes the following commits:

75d63cb [Ankur Dave] Rebuild routing table after Graph.reverse

JIRA issue:[SPARK-1511](https://issues.apache.org/jira/browse/SPARK-1511)

TestUtils.createCompiledClass method use renameTo() to move files which fails when the src and dest files are in different disks or partitions. This pr uses Files.move() instead. The move method will try to use renameTo() and then fall back to copy() and delete(). I think this should handle this issue.

I didn't found a test suite for this file, so I add file existence detection after file moving.

Author: Ye Xianjin <advancedxy@gmail.com>

Closes #427 from advancedxy/SPARK-1511 and squashes the following commits:

a2b97c7 [Ye Xianjin] Based on @srowen's comment, assert file existence.
6f95550 [Ye Xianjin] use Files.move instead of renameTo to handle the src and dest files are in different disks or partitions.

YARN-1824 changes the APIs (addToEnvironment, setEnvFromInputString) in Apps, which causes the spark build to break if built against a version 2.4.0. To fix this, create the spark own function to do that functionality which will not break compiling against 2.3 and other 2.x versions.

Author: xuan <xuan@MacBook-Pro.local>
Author: xuan <xuan@macbook-pro.home>

Closes #396 from xgong/master and squashes the following commits:

42b5984 [xuan] Remove two extra imports
bc0926f [xuan] Remove usage of org.apache.hadoop.util.Shell
be89fa7 [xuan] fix Spark compilation is broken with the latest hadoop-2.4.0 release

... nicer error messages

There are  two improvements to Scheduler Mode:
1. Made the built in ones case insensitive (fair/FAIR, fifo/FIFO).
2. If an invalid mode is given we should print a better error message.

Author: Sandeep <sandeep@techaddict.me>

Closes #388 from techaddict/1469 and squashes the following commits:

a31bbd5 [Sandeep] SPARK-1469: Scheduler mode should accept lower-case definitions and have nicer error messages There are  two improvements to Scheduler Mode: 1. Made the built in ones case insensitive (fair/FAIR, fifo/FIFO). 2. If an invalid mode is given we should print a better error message.

(I wasn't sure how to automatically set `SPARK_YARN=true` and `SPARK_HIVE=true` when running scalastyle, but these are the errors that turn up.)

Author: Sean Owen <sowen@cloudera.com>

Closes #413 from srowen/SPARK-1497 and squashes the following commits:

f0c9318 [Sean Owen] Fix more scalastyle warnings in yarn
80bf4c3 [Sean Owen] Add YARN alpha / YARN profile to scalastyle check
026319c [Sean Owen] Fix scalastyle warnings in YARN, Hive code

SPARK-1310: Start adding k-fold cross validation to MLLib [adds kFold to MLUtils & fixes bug in BernoulliSampler]

Author: Holden Karau <holden@pigscanfly.ca>

Closes #18 from holdenk/addkfoldcrossvalidation and squashes the following commits:

208db9b [Holden Karau] Fix a bad space
e84f2fc [Holden Karau] Fix the test, we should be looking at the second element instead
6ddbf05 [Holden Karau] swap training and validation order
7157ae9 [Holden Karau] CR feedback
90896c7 [Holden Karau] New line
150889c [Holden Karau] Fix up error messages in the MLUtilsSuite
2cb90b3 [Holden Karau] Fix the names in kFold
c702a96 [Holden Karau] Fix imports in MLUtils
e187e35 [Holden Karau] Move { up to same line as whenExecuting(random) in RandomSamplerSuite.scala
c5b723f [Holden Karau] clean up
7ebe4d5 [Holden Karau] CR feedback, remove unecessary learners (came back during merge mistake) and insert an empty line
bb5fa56 [Holden Karau] extra line sadness
163c5b1 [Holden Karau] code review feedback 1.to -> 1 to and folds -> numFolds
5a33f1d [Holden Karau] Code review follow up.
e8741a7 [Holden Karau] CR feedback
b78804e [Holden Karau] Remove cross validation [TODO in another pull request]
91eae64 [Holden Karau] Consolidate things in mlutils
264502a [Holden Karau] Add a test for the bug that was found with BernoulliSampler not copying the complement param
dd0b737 [Holden Karau] Wrap long lines (oops)
c0b7fa4 [Holden Karau] Switch FoldedRDD to use BernoulliSampler and PartitionwiseSampledRDD
08f8e4d [Holden Karau] Fix BernoulliSampler to respect complement
a751ec6 [Holden Karau] Add k-fold cross validation to MLLib

actually, the value 8 is only valid in mesos fine-grained mode :
<code>
  override def defaultParallelism() = sc.conf.getInt("spark.default.parallelism", 8)
</code>

while in coarse-grained model including mesos coares-grained, the value of the property depending on core numbers!
<code>
override def defaultParallelism(): Int = {
   conf.getInt("spark.default.parallelism", math.max(totalCoreCount.get(), 2))
  }
</code>

Author: Chen Chao <crazyjvm@gmail.com>

Closes #389 from CrazyJvm/patch-2 and squashes the following commits:

84a7fe4 [Chen Chao] miss </li> at the end of every single line
04a9796 [Chen Chao] change format
ee0fae0 [Chen Chao] update spark.default.parallelism

When running Hive tests, the working directory is `$SPARK_HOME/sql/hive`, while when running `sbt hive/console`, it becomes `$SPARK_HOME`, and test tables are not loaded if `HIVE_DEV_HOME` is not defined.

Author: Cheng Lian <lian.cs.zju@gmail.com>

Closes #417 from liancheng/loadTestTables and squashes the following commits:

7cea8d6 [Cheng Lian] Loads test tables when running "sbt hive/console" without HIVE_DEV_HOME

This is not an issue with the driver UI, but when you fire
up the history server, there's currently no way to go back to
the app listing page without editing the browser's location
field (since the logo's link points to the root of the
application's own UI - i.e. the "stages" tab).

The change just points the logo link to "/", which is the app
listing for the history server, and the stages tab for the
driver's UI.

Tested with both history server and live driver.

Author: Marcelo Vanzin <vanzin@cloudera.com>

Closes #408 from vanzin/web-ui-root and squashes the following commits:

1b60cb6 [Marcelo Vanzin] Make "spark logo" link refer to "/".

JIRA issue: [SPARK-959](https://spark-project.atlassian.net/browse/SPARK-959)

SBT 0.13.2 has been officially released. This version updated Ivy 2.0 to Ivy 2.3, which fixes [IVY-899](https://issues.apache.org/jira/browse/IVY-899). This PR also removed previous workaround.

Author: Cheng Lian <lian.cs.zju@gmail.com>

Closes #426 from liancheng/updateSbt and squashes the following commits:

95e3dc8 [Cheng Lian] Updated SBT from 0.13.1 to 0.13.2 to fix SPARK-959

This makes it possible to create tables and insert into them using the DSL and SQL for the scala and java apis.

Author: Michael Armbrust <michael@databricks.com>

Closes #354 from marmbrus/insertIntoTable and squashes the following commits:

6c6f227 [Michael Armbrust] Create random temporary files in python parquet unit tests.
f5e6d5c [Michael Armbrust] Merge remote-tracking branch 'origin/master' into insertIntoTable
765c506 [Michael Armbrust] Add to JavaAPI.
77b512c [Michael Armbrust] typos.
5c3ef95 [Michael Armbrust] use names for boolean args.
882afdf [Michael Armbrust] Change createTableAs to saveAsTable.  Clean up api annotations.
d07d94b [Michael Armbrust] Add tests, support for creating parquet files and hive tables.
fa3fe81 [Michael Armbrust] Make insertInto available on JavaSchemaRDD as well.  Add createTableAs function.

This PR adds a SparseVector class in PySpark and updates all the regression, classification and clustering algorithms and models to support sparse data, similar to MLlib. I chose to add this class because SciPy is quite difficult to install in many environments (more so than NumPy), but I plan to add support for SciPy sparse vectors later too, and make the methods work transparently on objects of either type.

On the Scala side, we keep Python sparse vectors sparse and pass them to MLlib. We always return dense vectors from our models.

Some to-do items left:
- [x] Support SciPy's scipy.sparse matrix objects when SciPy is available. We can easily add a function to convert these to our own SparseVector.
- [x] MLlib currently uses a vector with one extra column on the left to represent what we call LabeledPoint in Scala. Do we really want this? It may get annoying once you deal with sparse data since you must add/subtract 1 to each feature index when training. We can remove this API in 1.0 and use tuples for labeling.
- [x] Explain how to use these in the Python MLlib docs.

CC @mengxr, @joshrosen

Author: Matei Zaharia <matei@databricks.com>

Closes #341 from mateiz/py-ml-update and squashes the following commits:

d52e763 [Matei Zaharia] Remove no-longer-needed slice code and handle review comments
ea5a25a [Matei Zaharia] Fix remaining uses of copyto() after merge
b9f97a3 [Matei Zaharia] Fix test
1e1bd0f [Matei Zaharia] Add MLlib logistic regression example in Python
88bc01f [Matei Zaharia] Clean up inheritance of LinearModel in Python, and expose its parametrs
37ab747 [Matei Zaharia] Fix some examples and docs due to changes in MLlib API
da0f27e [Matei Zaharia] Added a MLlib K-means example and updated docs to discuss sparse data
c48e85a [Matei Zaharia] Added some tests for passing lists as input, and added mllib/tests.py to run-tests script.
a07ba10 [Matei Zaharia] Fix some typos and calculation of initial weights
74eefe7 [Matei Zaharia] Added LabeledPoint class in Python
889dde8 [Matei Zaharia] Support scipy.sparse matrices in all our algorithms and models
ab244d1 [Matei Zaharia] Allow SparseVectors to be initialized using a dict
a5d6426 [Matei Zaharia] Add linalg.py to run-tests script
0e7a3d8 [Matei Zaharia] Keep vectors sparse in Java when reading LabeledPoints
eaee759 [Matei Zaharia] Update regression, classification and clustering models for sparse data
2abbb44 [Matei Zaharia] Further work to get linear models working with sparse data
154f45d [Matei Zaharia] Update docs, name some magic values
881fef7 [Matei Zaharia] Added a sparse vector in Python and made Java-Python format more compact

I saw `No "scala-library*.jar" in Scala compiler library` error in IDEA. It seems upgrading `sbt-idea` to 1.6.0 fixed the problem.

Author: Xiangrui Meng <meng@databricks.com>

Closes #419 from mengxr/idea-plugin and squashes the following commits:

fb3c35f [Xiangrui Meng] update sbt-idea to version 1.6.0

This is a simple first step towards avoiding running the Hive tests
whenever possible.

Author: Patrick Wendell <pwendell@gmail.com>

Closes #420 from pwendell/test-isolation and squashes the following commits:

350c8af [Patrick Wendell] SPARK-1455: Better isolation for unit tests.

Added documentation for user to use the decision tree algorithms for classification and regression in Spark 1.0 release.

Apart from a general review, I need specific input on the following:
* I had to move a lot of the existing documentation under the *linear methods* umbrella to accommodate decision trees. I wonder if there is a better way to organize the programming guide given we are so close to the release.
* I have not looked closely at pyspark but I am wondering new mllib algorithms are automatically plugged in or do we need to some extra work to call mllib functions from pyspark. I will add to the pyspark examples based upon the advice I get.

cc: @mengxr, @hirakendu, @etrain, @atalwalkar

Author: Manish Amde <manish9ue@gmail.com>

Closes #402 from manishamde/tree_doc and squashes the following commits:

022485a [Manish Amde] more documentation
865826e [Manish Amde] minor: grammar
dbb0e5e [Manish Amde] minor improvements to text
b9ef6c4 [Manish Amde] basic decision tree code examples
6e297d7 [Manish Amde] added subsections
f427e84 [Manish Amde] renaming sections
9c0c4be [Manish Amde] split candidate
6925275 [Manish Amde] impurity and information gain
94fd2f9 [Manish Amde] more reorg
b93125c [Manish Amde] more subsection reorg
3ecb2ad [Manish Amde] minor text addition
1537dd3 [Manish Amde] added placeholders and some doc
d06511d [Manish Amde] basic skeleton

This PR uses Breeze's L-BFGS implement, and Breeze dependency has already been introduced by Xiangrui's sparse input format work in SPARK-1212. Nice work, @mengxr !

When use with regularized updater, we need compute the regVal and regGradient (the gradient of regularized part in the cost function), and in the currently updater design, we can compute those two values by the following way.

Let's review how updater works when returning newWeights given the input parameters.

w' = w - thisIterStepSize * (gradient + regGradient(w))  Note that regGradient is function of w!
If we set gradient = 0, thisIterStepSize = 1, then
regGradient(w) = w - w'

As a result, for regVal, it can be computed by

    val regVal = updater.compute(
      weights,
      new DoubleMatrix(initialWeights.length, 1), 0, 1, regParam)._2
and for regGradient, it can be obtained by

      val regGradient = weights.sub(
        updater.compute(weights, new DoubleMatrix(initialWeights.length, 1), 1, 1, regParam)._1)

The PR includes the tests which compare the result with SGD with/without regularization.

We did a comparison between LBFGS and SGD, and often we saw 10x less
steps in LBFGS while the cost of per step is the same (just computing
the gradient).

The following is the paper by Prof. Ng at Stanford comparing different
optimizers including LBFGS and SGD. They use them in the context of
deep learning, but worth as reference.
http://cs.stanford.edu/~jngiam/papers/LeNgiamCoatesLahiriProchnowNg2011.pdf

Author: DB Tsai <dbtsai@alpinenow.com>

Closes #353 from dbtsai/dbtsai-LBFGS and squashes the following commits:

984b18e [DB Tsai] L-BFGS Optimizer based on Breeze's implementation. Also fixed indentation issue in GradientDescent optimizer.

The Graph.apply test in GraphSuite had some assertions in a closure in
a graph transformation. As a consequence, these assertions never
actually executed.  Furthermore, these closures had a reference to
(non-serializable) test harness classes because they called assert(),
which could be a problem if we proactively check closure serializability
in the future.

This commit simply changes the Graph.apply test to collect the graph
triplets so it can assert about each triplet from a map method.

Author: William Benton <willb@redhat.com>

Closes #415 from willb/graphsuite-nop-fix and squashes the following commits:

0b63658 [William Benton] Ensure assertions in Graph.apply are asserted.