4b736dbab3
This PR adds package "org.apache.spark.ml" with pipeline and parameters, as discussed on the JIRA. This is a joint work of jkbradley etrain shivaram and many others who helped on the design, also with help from marmbrus and liancheng on the Spark SQL side. The design doc can be found at: https://docs.google.com/document/d/1rVwXRjWKfIb-7PI6b86ipytwbUH7irSNLF1_6dLmh8o/edit?usp=sharing **org.apache.spark.ml** This is a new package with new set of ML APIs that address practical machine learning pipelines. (Sorry for taking so long!) It will be an alpha component, so this is definitely not something set in stone. The new set of APIs, inspired by the MLI project from AMPLab and scikit-learn, takes leverage on Spark SQL's schema support and execution plan optimization. It introduces the following components that help build a practical pipeline: 1. Transformer, which transforms a dataset into another 2. Estimator, which fits models to data, where models are transformers 3. Evaluator, which evaluates model output and returns a scalar metric 4. Pipeline, a simple pipeline that consists of transformers and estimators Parameters could be supplied at fit/transform or embedded with components. 1. Param: a strong-typed parameter key with self-contained doc 2. ParamMap: a param -> value map 3. Params: trait for components with parameters For any component that implements `Params`, user can easily check the doc by calling `explainParams`: ~~~ > val lr = new LogisticRegression > lr.explainParams maxIter: max number of iterations (default: 100) regParam: regularization constant (default: 0.1) labelCol: label column name (default: label) featuresCol: features column name (default: features) ~~~ or user can check individual param: ~~~ > lr.maxIter maxIter: max number of iterations (default: 100) ~~~ **Please start with the example code in test suites and under `org.apache.spark.examples.ml`, where I put several examples:** 1. run a simple logistic regression job ~~~ val lr = new LogisticRegression() .setMaxIter(10) .setRegParam(1.0) val model = lr.fit(dataset) model.transform(dataset, model.threshold -> 0.8) // overwrite threshold .select('label, 'score, 'prediction).collect() .foreach(println) ~~~ 2. run logistic regression with cross-validation and grid search using areaUnderROC (default) as the metric ~~~ val lr = new LogisticRegression val lrParamMaps = new ParamGridBuilder() .addGrid(lr.regParam, Array(0.1, 100.0)) .addGrid(lr.maxIter, Array(0, 5)) .build() val eval = new BinaryClassificationEvaluator val cv = new CrossValidator() .setEstimator(lr) .setEstimatorParamMaps(lrParamMaps) .setEvaluator(eval) .setNumFolds(3) val bestModel = cv.fit(dataset) ~~~ 3. run a pipeline that consists of a standard scaler and a logistic regression component ~~~ val scaler = new StandardScaler() .setInputCol("features") .setOutputCol("scaledFeatures") val lr = new LogisticRegression() .setFeaturesCol(scaler.getOutputCol) val pipeline = new Pipeline() .setStages(Array(scaler, lr)) val model = pipeline.fit(dataset) val predictions = model.transform(dataset) .select('label, 'score, 'prediction) .collect() .foreach(println) ~~~ 4. a simple text classification pipeline, which recognizes "spark": ~~~ val training = sparkContext.parallelize(Seq( LabeledDocument(0L, "a b c d e spark", 1.0), LabeledDocument(1L, "b d", 0.0), LabeledDocument(2L, "spark f g h", 1.0), LabeledDocument(3L, "hadoop mapreduce", 0.0))) val tokenizer = new Tokenizer() .setInputCol("text") .setOutputCol("words") val hashingTF = new HashingTF() .setInputCol(tokenizer.getOutputCol) .setOutputCol("features") val lr = new LogisticRegression() .setMaxIter(10) val pipeline = new Pipeline() .setStages(Array(tokenizer, hashingTF, lr)) val model = pipeline.fit(training) val test = sparkContext.parallelize(Seq( Document(4L, "spark i j k"), Document(5L, "l m"), Document(6L, "mapreduce spark"), Document(7L, "apache hadoop"))) model.transform(test) .select('id, 'text, 'prediction, 'score) .collect() .foreach(println) ~~~ Java examples are very similar. I put example code that creates a simple text classification pipeline in Scala and Java, where a simple tokenizer is defined as a transformer outside `org.apache.spark.ml`. **What are missing now and will be added soon:** 1. ~~Runtime check of schemas. So before we touch the data, we will go through the schema and make sure column names and types match the input parameters.~~ 2. ~~Java examples.~~ 3. ~~Store training parameters in trained models.~~ 4. (later) Serialization and Python API. Author: Xiangrui Meng <meng@databricks.com> Closes #3099 from mengxr/SPARK-3530 and squashes the following commits: 2cc93fd [Xiangrui Meng] hide APIs as much as I can 34319ba [Xiangrui Meng] use local instead local[2] for unit tests 2524251 [Xiangrui Meng] rename PipelineStage.transform to transformSchema c9daab4 [Xiangrui Meng] remove mockito version 1397ab5 [Xiangrui Meng] use sqlContext from LocalSparkContext instead of TestSQLContext 6ffc389 [Xiangrui Meng] try to fix unit test a59d8b7 [Xiangrui Meng] doc updates 977fd9d [Xiangrui Meng] add scala ml package object 6d97fe6 [Xiangrui Meng] add AlphaComponent annotation 731f0e4 [Xiangrui Meng] update package doc 0435076 [Xiangrui Meng] remove ;this from setters fa21d9b [Xiangrui Meng] update extends indentation f1091b3 [Xiangrui Meng] typo 228a9f4 [Xiangrui Meng] do not persist before calling binary classification metrics f51cd27 [Xiangrui Meng] rename default to defaultValue b3be094 [Xiangrui Meng] refactor schema transform in lr 8791e8e [Xiangrui Meng] rename copyValues to inheritValues and make it do the right thing 51f1c06 [Xiangrui Meng] remove leftover code in Transformer 494b632 [Xiangrui Meng] compure score once ad678e9 [Xiangrui Meng] more doc for Transformer 4306ed4 [Xiangrui Meng] org imports in text pipeline 6e7c1c7 [Xiangrui Meng] update pipeline 4f9e34f [Xiangrui Meng] more doc for pipeline aa5dbd4 [Xiangrui Meng] fix typo 11be383 [Xiangrui Meng] fix unit tests 3df7952 [Xiangrui Meng] clean up 986593e [Xiangrui Meng] re-org java test suites 2b11211 [Xiangrui Meng] remove external data deps 9fd4933 [Xiangrui Meng] add unit test for pipeline 2a0df46 [Xiangrui Meng] update tests 2d52e4d [Xiangrui Meng] add @AlphaComponent to package-info 27582a4 [Xiangrui Meng] doc changes 73a000b [Xiangrui Meng] add schema transformation layer 6736e87 [Xiangrui Meng] more doc / remove HasMetricName trait 80a8b5e [Xiangrui Meng] rename SimpleTransformer to UnaryTransformer 62ca2bb [Xiangrui Meng] check param parent in set/get 1622349 [Xiangrui Meng] add getModel to PipelineModel a0e0054 [Xiangrui Meng] update StandardScaler to use SimpleTransformer d0faa04 [Xiangrui Meng] remove implicit mapping from ParamMap c7f6921 [Xiangrui Meng] move ParamGridBuilder test to ParamGridBuilderSuite e246f29 [Xiangrui Meng] re-org: 7772430 [Xiangrui Meng] remove modelParams add a simple text classification pipeline b95c408 [Xiangrui Meng] remove implicits add unit tests to params bab3e5b [Xiangrui Meng] update params fe0ee92 [Xiangrui Meng] Merge remote-tracking branch 'apache/master' into SPARK-3530 6e86d98 [Xiangrui Meng] some code clean-up 2d040b3 [Xiangrui Meng] implement setters inside each class, add Params.copyValues [ci skip] fd751fc [Xiangrui Meng] add java-friendly versions of fit and tranform 3f810cd [Xiangrui Meng] use multi-model training api in cv 5b8f413 [Xiangrui Meng] rename model to modelParams 9d2d35d [Xiangrui Meng] test varargs and chain model params f46e927 [Xiangrui Meng] Merge remote-tracking branch 'apache/master' into SPARK-3530 1ef26e0 [Xiangrui Meng] specialize methods/types for Java df293ed [Xiangrui Meng] switch to setter/getter 376db0a [Xiangrui Meng] pipeline and parameters |
||
|---|---|---|
| assembly | ||
| bagel | ||
| bin | ||
| conf | ||
| core | ||
| data/mllib | ||
| dev | ||
| docker | ||
| docs | ||
| ec2 | ||
| examples | ||
| external | ||
| extras | ||
| graphx | ||
| mllib | ||
| network | ||
| project | ||
| python | ||
| repl | ||
| sbin | ||
| sbt | ||
| sql | ||
| streaming | ||
| tools | ||
| yarn | ||
| .gitattributes | ||
| .gitignore | ||
| .rat-excludes | ||
| CONTRIBUTING.md | ||
| LICENSE | ||
| make-distribution.sh | ||
| NOTICE | ||
| pom.xml | ||
| README.md | ||
| scalastyle-config.xml | ||
| tox.ini | ||
Apache Spark
Spark is a fast and general cluster computing system for Big Data. It provides high-level APIs in Scala, Java, and Python, and an optimized engine that supports general computation graphs for data analysis. It also supports a rich set of higher-level tools including Spark SQL for SQL and structured data processing, MLlib for machine learning, GraphX for graph processing, and Spark Streaming for stream processing.
Online Documentation
You can find the latest Spark documentation, including a programming guide, on the project web page and project wiki. This README file only contains basic setup instructions.
Building Spark
Spark is built using Apache Maven. To build Spark and its example programs, run:
mvn -DskipTests clean package
(You do not need to do this if you downloaded a pre-built package.) More detailed documentation is available from the project site, at "Building Spark with Maven".
Interactive Scala Shell
The easiest way to start using Spark is through the Scala shell:
./bin/spark-shell
Try the following command, which should return 1000:
scala> sc.parallelize(1 to 1000).count()
Interactive Python Shell
Alternatively, if you prefer Python, you can use the Python shell:
./bin/pyspark
And run the following command, which should also return 1000:
>>> sc.parallelize(range(1000)).count()
Example Programs
Spark also comes with several sample programs in the examples directory.
To run one of them, use ./bin/run-example <class> [params]. For example:
./bin/run-example SparkPi
will run the Pi example locally.
You can set the MASTER environment variable when running examples to submit
examples to a cluster. This can be a mesos:// or spark:// URL,
"yarn-cluster" or "yarn-client" to run on YARN, and "local" to run
locally with one thread, or "local[N]" to run locally with N threads. You
can also use an abbreviated class name if the class is in the examples
package. For instance:
MASTER=spark://host:7077 ./bin/run-example SparkPi
Many of the example programs print usage help if no params are given.
Running Tests
Testing first requires building Spark. Once Spark is built, tests can be run using:
./dev/run-tests
Please see the guidance on how to run all automated tests.
A Note About Hadoop Versions
Spark uses the Hadoop core library to talk to HDFS and other Hadoop-supported storage systems. Because the protocols have changed in different versions of Hadoop, you must build Spark against the same version that your cluster runs.
Please refer to the build documentation at "Specifying the Hadoop Version" for detailed guidance on building for a particular distribution of Hadoop, including building for particular Hive and Hive Thriftserver distributions. See also "Third Party Hadoop Distributions" for guidance on building a Spark application that works with a particular distribution.
Configuration
Please refer to the Configuration guide in the online documentation for an overview on how to configure Spark.