├── data ├── people.txt └── inputfile.txt ├── renovate.json ├── README.md ├── src ├── main │ ├── scala │ │ ├── com │ │ │ └── javachen │ │ │ │ └── spark │ │ │ │ └── examples │ │ │ │ ├── rdd │ │ │ │ ├── Pipe.scala │ │ │ │ ├── ActionTest.scala │ │ │ │ ├── TransformTest.scala │ │ │ │ ├── FlatMap.scala │ │ │ │ ├── Lookup.scala │ │ │ │ ├── CollectAsMap.scala │ │ │ │ ├── MapPartitions.scala │ │ │ │ ├── ReduceByKey.scala │ │ │ │ ├── AggregateOrder.scala │ │ │ │ ├── MapValues.scala │ │ │ │ ├── Aggregate.scala │ │ │ │ ├── PartitionBy.scala │ │ │ │ ├── Join.scala │ │ │ │ ├── ScalaWordCount.scala │ │ │ │ ├── GroupWith.scala │ │ │ │ ├── GroupByKey.scala │ │ │ │ ├── Cartesian.scala │ │ │ │ └── GroupByAction.scala │ │ │ │ ├── sparksql │ │ │ │ ├── ScalaSparkSQLByReflection.scala │ │ │ │ └── ScalaSparkSQLBySchema.scala │ │ │ │ └── mllib │ │ │ │ ├── EvaluateResult.scala │ │ │ │ ├── ScalaMovieLensALS.scala │ │ │ │ ├── MovieSimilarities.scala │ │ │ │ ├── MovieLensALS.scala │ │ │ │ └── ScalaLocalALS.scala │ │ └── org │ │ │ └── apache │ │ │ └── spark │ │ │ └── examples │ │ │ ├── LocalPi.scala │ │ │ ├── ExceptionHandlingTest.scala │ │ │ ├── SparkPi.scala │ │ │ ├── HdfsTest.scala │ │ │ ├── SparkTachyonPi.scala │ │ │ ├── DriverSubmissionTest.scala │ │ │ ├── MultiBroadcastTest.scala │ │ │ ├── GroupByTest.scala │ │ │ ├── BroadcastTest.scala │ │ │ ├── SkewedGroupByTest.scala │ │ │ ├── SparkTC.scala │ │ │ ├── LocalFileLR.scala │ │ │ ├── SparkPageRank.scala │ │ │ ├── LocalLR.scala │ │ │ ├── SimpleSkewedGroupByTest.scala │ │ │ ├── SparkLR.scala │ │ │ ├── SparkTachyonHdfsLR.scala │ │ │ ├── SparkKMeans.scala │ │ │ ├── SparkHdfsLR.scala │ │ │ ├── LocalKMeans.scala │ │ │ ├── LogQuery.scala │ │ │ ├── LocalALS.scala │ │ │ └── SparkALS.scala │ ├── python │ │ ├── PythonWordCount.py │ │ ├── PythonALS.py │ │ ├── PythonSparkSQLByReflection.py │ │ └── PythonSparkSQLBySchema.py │ └── java │ │ └── com │ │ └── javachen │ │ └── spark │ │ └── examples │ │ ├── sparksql │ │ ├── JavaSparkSQLBySchema.java │ │ └── JavaSparkSQLByReflection.java │ │ ├── mllib │ │ └── JavaALS.java │ │ └── rdd │ │ └── JavaWordCount.java └── test │ └── java │ └── com │ └── javachen │ └── spark │ └── AppTest.java ├── .gitignore ├── LICENSE └── pom.xml /data/people.txt: -------------------------------------------------------------------------------- 1 | Michael, 29 2 | Andy, 30 3 | Justin, 19 -------------------------------------------------------------------------------- /renovate.json: -------------------------------------------------------------------------------- 1 | { 2 | "$schema": "https://docs.renovatebot.com/renovate-schema.json", 3 | "extends": [ 4 | "config:recommended" 5 | ] 6 | } 7 | -------------------------------------------------------------------------------- /data/inputfile.txt: -------------------------------------------------------------------------------- 1 | apple 2 | banana counter 3 | counter one two three 4 | three one 5 | five seven eight 6 | twenty one three five counter six 7 | one siz helga 8 | apple banana fiver -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # learning-spark 2 | 3 | Learning to write Spark examples 4 | 5 | # Links 6 | 7 | - https://github.com/JerryLead/SparkLearning 8 | - https://github.com/ceteri/spark-exercises 9 | - https://github.com/databricks/reference-apps -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/Pipe.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | /** 4 | * 5 | * @author june. 6 | * @date 2015-05-12 17:21. 7 | */ 8 | object Pipe { 9 | 10 | } 11 | -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/ActionTest.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | /** 4 | * 5 | * @author june. 6 | * @date 2015-05-12 17:25. 7 | */ 8 | object ActionTest { 9 | 10 | } 11 | -------------------------------------------------------------------------------- /.gitignore: -------------------------------------------------------------------------------- 1 | *.iml 2 | .idea/ 3 | target/ 4 | *.class 5 | 6 | # Mobile Tools for Java (J2ME) 7 | .mtj.tmp/ 8 | 9 | # Package Files # 10 | *.jar 11 | *.war 12 | *.ear 13 | 14 | # virtual machine crash logs, see http://www.java.com/en/download/help/error_hotspot.xml 15 | -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/TransformTest.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | /** 4 | * 5 | * @author june. 6 | * @date 2015-05-12 17:25. 7 | */ 8 | object TransformTest { 9 | 10 | } 11 | -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/FlatMap.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.SparkContext 4 | 5 | object FlatMap { 6 | def main(args: Array[String]) { 7 | 8 | val sc = new SparkContext("local", "FlatMap Test") 9 | val data = Array[(String, Int)](("A", 1), ("B", 2), 10 | ("B", 3), ("C", 4), 11 | ("C", 5), ("C", 6) 12 | ) 13 | val pairs = sc.makeRDD(data, 3) 14 | 15 | val result = pairs.flatMap(T => (T._1 + T._2)) 16 | 17 | result.foreach(println) 18 | 19 | } 20 | } -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/Lookup.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.SparkContext 4 | import org.apache.spark.SparkContext._ 5 | 6 | object Lookup { 7 | def main(args: Array[String]) { 8 | 9 | val sc = new SparkContext("local", "LookUp Test") 10 | 11 | val data = Array[(String, Int)](("A", 1), ("B", 2), 12 | ("B", 3), ("C", 4), 13 | ("C", 5), ("C", 6)) 14 | 15 | val pairs = sc.parallelize(data, 3) 16 | 17 | val finalRDD = pairs.lookup("B") 18 | 19 | finalRDD.foreach(println) 20 | // output: 21 | // 2 22 | // 3 23 | } 24 | } -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/CollectAsMap.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.SparkContext 4 | import org.apache.spark.SparkContext._ 5 | 6 | object CollectAsMap { 7 | def main(args: Array[String]) { 8 | 9 | val sc = new SparkContext("local", "CollectAsMap Test") 10 | val data = Array[(String, Int)](("A", 1), ("B", 2), 11 | ("B", 3), ("C", 4), 12 | ("C", 5), ("C", 6)) 13 | 14 | // as same as "val pairs = sc.parallelize(data, 3)" 15 | val pairs = sc.makeRDD(data, 3) 16 | 17 | val result = pairs.collectAsMap 18 | 19 | // output Map(A -> 1, C -> 6, B -> 3) 20 | print(result) 21 | } 22 | 23 | } -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/MapPartitions.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.SparkContext 4 | 5 | object MapPartitions { 6 | 7 | def main(args: Array[String]) { 8 | val sc = new SparkContext("local", "MapPartitionsRDD Test") 9 | val data = Array[(String, Int)](("A1", 1), ("A2", 2), 10 | ("B1", 1), ("B2", 4), 11 | ("C1", 3), ("C2", 4) 12 | ) 13 | val pairs = sc.parallelize(data, 3) 14 | 15 | val finalRDD = pairs.mapPartitions(iter => iter.filter(_._2 >= 2)) 16 | // val finalRDD2 = pairs.mapPartitionsWithIndex(f, preservesPartitioning) 17 | 18 | finalRDD.toArray().foreach(println) 19 | 20 | } 21 | } -------------------------------------------------------------------------------- /src/main/python/PythonWordCount.py: -------------------------------------------------------------------------------- 1 | import sys 2 | 3 | from pyspark import SparkContext 4 | 5 | if __name__ == "__main__": 6 | file=sys.argv[1] 7 | threshold=int(sys.argv[2]) 8 | sc = SparkContext(appName="PythonWordCount") 9 | lines = sc.textFile(file, 1) 10 | counts = lines.flatMap(lambda x: x.split(' ')) \ 11 | .map(lambda x: (x, 1)) \ 12 | .reduceByKey(lambda a, b: a + b) \ 13 | .filter(lambda (a, b) : b >= threshold) \ 14 | .flatMap(lambda (a, b): list(a)) \ 15 | .map(lambda x: (x, 1)) \ 16 | .reduceByKey(lambda a, b: a + b) 17 | 18 | print ",".join(str(t) for t in counts.collect()) 19 | sc.stop() -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/ReduceByKey.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.SparkContext 4 | 5 | object ReduceByKey { 6 | 7 | def main(args: Array[String]) { 8 | 9 | val sc = new SparkContext("local", "ReduceByKeyToDriver Test") 10 | val data1 = Array[(String, Int)](("K", 1), ("U", 2), 11 | ("U", 3), ("W", 4), 12 | ("W", 5), ("W", 6)) 13 | val pairs = sc.parallelize(data1, 3) 14 | //val result = pairs.reduce((A, B) => (A._1 + "#" + B._1, A._2 + B._2)) 15 | //val result = pairs.fold(("K0",10))((A, B) => (A._1 + "#" + B._1, A._2 + B._2)) 16 | val result = pairs.reduceByKey(_ + _, 2) 17 | result.foreach(println) 18 | } 19 | 20 | } -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/AggregateOrder.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.SparkContext 4 | 5 | object AggregateOrder { 6 | 7 | def main(args: Array[String]) { 8 | 9 | val sc = new SparkContext("local", "AggregateOrder Test") 10 | val data = List("12", "23", "345", "4567") 11 | 12 | val pairs = sc.parallelize(data, 2) 13 | pairs.foreach(x => println(x.length)) 14 | 15 | //val result = pairs.aggregate("")((x,y) => math.min(x.length, y.length).toString, (x,y) => x + y) 16 | 17 | val result2 = pairs.aggregate("")((x,y) => "[" + x.length + "," + y.length + "] ", (x,y) => x + y) 18 | 19 | result2.foreach(println) 20 | println(result2) 21 | 22 | } 23 | } -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/MapValues.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.SparkContext 4 | 5 | object MapValues { 6 | def main(args: Array[String]) { 7 | 8 | val sc = new SparkContext("local", "ReduceByKeyToDriver Test") 9 | val data1 = Array[(String, Int)](("K", 1), ("T", 2), 10 | ("T", 3), ("W", 4), 11 | ("W", 5), ("W", 6) 12 | ) 13 | val pairs = sc.parallelize(data1, 3) 14 | //val result = pairs.reduce((A, B) => (A._1 + "#" + B._1, A._2 + B._2)) 15 | //val result = pairs.fold(("K0",10))((A, B) => (A._1 + "#" + B._1, A._2 + B._2)) 16 | //val result = pairs.partitionBy(new RangePartitioner(2, pairs, true)) 17 | val result = pairs.mapValues(V => 10 * V) 18 | result.foreach(println) 19 | } 20 | } -------------------------------------------------------------------------------- /src/test/java/com/javachen/spark/AppTest.java: -------------------------------------------------------------------------------- 1 | package com.javachen.spark; 2 | 3 | import junit.framework.Test; 4 | import junit.framework.TestCase; 5 | import junit.framework.TestSuite; 6 | 7 | /** 8 | * Unit test for simple App. 9 | */ 10 | public class AppTest 11 | extends TestCase { 12 | /** 13 | * Create the test case 14 | * 15 | * @param testName name of the test case 16 | */ 17 | public AppTest(String testName) { 18 | super(testName); 19 | } 20 | 21 | /** 22 | * @return the suite of tests being tested 23 | */ 24 | public static Test suite() { 25 | return new TestSuite(AppTest.class); 26 | } 27 | 28 | /** 29 | * Rigourous Test :-) 30 | */ 31 | public void testApp() { 32 | assertTrue(true); 33 | } 34 | } 35 | -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/Aggregate.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.SparkContext 4 | 5 | object Aggregate { 6 | 7 | def main(args: Array[String]) { 8 | 9 | val sc = new SparkContext("local", "AggregateAction Test") 10 | val data = Array[(String, Int)](("A1", 1), ("A2", 2), 11 | ("B1", 3), ("B2", 4), 12 | ("C1", 5), ("C2", 6)) 13 | 14 | val pairs = sc.parallelize(data, 3) 15 | 16 | // output: 17 | // (A1,1)(A2,2) 18 | // (B1,3)(B2,4) 19 | // (C1,5)(C2,6) 20 | pairs.foreach(print) 21 | 22 | val result = pairs.aggregate(("", 0))((U, T) => (U._1 + T._1, U._2 + T._2), (U, T) => 23 | ("[" + U._1 + T._1 + "] ", U._2 + T._2)) 24 | 25 | // output ([[[A1A2] B1B2] C1C2] ,21) 26 | println(result) 27 | } 28 | } -------------------------------------------------------------------------------- /src/main/python/PythonALS.py: -------------------------------------------------------------------------------- 1 | from pyspark.mllib.recommendation import ALS 2 | from numpy import array 3 | 4 | # Load and parse the data 5 | data = sc.textFile("data/mllib/als/test.data") 6 | ratings = data.map(lambda line: array([float(x) for x in line.split(',')])) 7 | 8 | # Build the recommendation model using Alternating Least Squares 9 | rank = 10 10 | numIterations = 20 11 | model = ALS.train(ratings, rank, numIterations) 12 | 13 | # Evaluate the model on training data 14 | testdata = ratings.map(lambda p: (int(p[0]), int(p[1]))) 15 | predictions = model.predictAll(testdata).map(lambda r: ((r[0], r[1]), r[2])) 16 | ratesAndPreds = ratings.map(lambda r: ((r[0], r[1]), r[2])).join(predictions) 17 | MSE = ratesAndPreds.map(lambda r: (r[1][0] - r[1][1])**2).reduce(lambda x, y: x + y)/ratesAndPreds.count() 18 | print("Mean Squared Error = " + str(MSE)) -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/PartitionBy.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.{RangePartitioner,HashPartitioner, SparkContext} 4 | 5 | object PartitionBy { 6 | def main(args: Array[String]) { 7 | 8 | val sc = new SparkContext("local", "ReduceByKeyToDriver Test") 9 | val data1 = Array[(String, Int)](("K", 1), ("T", 2), 10 | ("T", 3), ("W", 4), 11 | ("W", 5), ("W", 6) 12 | ) 13 | val pairs = sc.parallelize(data1, 3) 14 | //val result = pairs.reduce((A, B) => (A._1 + "#" + B._1, A._2 + B._2)) 15 | //val result = pairs.fold(("K0",10))((A, B) => (A._1 + "#" + B._1, A._2 + B._2)) 16 | var result = pairs.partitionBy(new RangePartitioner(2, pairs, true)) 17 | result = pairs.partitionBy(new HashPartitioner(2)) 18 | result.foreach(println) 19 | } 20 | } -------------------------------------------------------------------------------- /src/main/python/PythonSparkSQLByReflection.py: -------------------------------------------------------------------------------- 1 | # sc is an existing SparkContext. 2 | from pyspark.sql import SQLContext, Row 3 | 4 | sqlContext = SQLContext(sc) 5 | 6 | # Load a text file and convert each line to a Row. 7 | lines = sc.textFile("people.txt") 8 | parts = lines.map(lambda l: l.split(",")) 9 | people = parts.map(lambda p: Row(name=p[0], age=int(p[1]))) 10 | 11 | # Infer the schema, and register the DataFrame as a table. 12 | schemaPeople = sqlContext.inferSchema(people) 13 | schemaPeople.registerTempTable("people") 14 | 15 | # SQL can be run over DataFrames that have been registered as a table. 16 | teenagers = sqlContext.sql("SELECT name FROM people WHERE age >= 13 AND age <= 19") 17 | 18 | # The results of SQL queries are RDDs and support all the normal RDD operations. 19 | teenNames = teenagers.map(lambda p: "Name: " + p.name) 20 | for teenName in teenNames.collect(): 21 | print teenName -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/Join.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.SparkContext 4 | import org.apache.spark.SparkContext._ 5 | 6 | object Join { 7 | def main(args: Array[String]) { 8 | 9 | val sc = new SparkContext("local[2]", "JoinAction Test") 10 | 11 | val data1 = Array[(String, Int)](("A1", 1), ("A2", 2), 12 | ("B1", 3), ("B2", 4), 13 | ("C1", 5), ("C1", 6) 14 | ) 15 | 16 | val data2 = Array[(String, Int)](("A1", 7), ("A2", 8), 17 | ("B1", 9), ("C1", 0) 18 | ) 19 | val pairs1 = sc.parallelize(data1, 3) 20 | val pairs2 = sc.parallelize(data2, 2) 21 | 22 | 23 | val result = pairs1.join(pairs2) 24 | 25 | // output: 26 | // (A1,(1,7)) 27 | // (B1,(3,9)) 28 | // (A2,(2,8)) 29 | // 30 | // (C1,(5,0)) 31 | // (C1,(6,0)) 32 | result.foreach(println) 33 | } 34 | 35 | } -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/ScalaWordCount.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.{SparkConf, SparkContext} 4 | 5 | object ScalaWordCount { 6 | def main(args: Array[String]) { 7 | val sc = new SparkContext(new SparkConf().setAppName("ScalaWordCount")) 8 | val threshold = args(1).toInt 9 | 10 | // split each document into words 11 | val tokenized = sc.textFile(args(0)).flatMap(_.split(" ")) 12 | 13 | // count the occurrence of each word 14 | val wordCounts = tokenized.map((_, 1)).reduceByKey(_ + _) 15 | 16 | // filter out words with less than threshold occurrences 17 | val filtered = wordCounts.filter(_._2 >= threshold) 18 | 19 | // count characters 20 | val charCounts = filtered.flatMap(_._1.toCharArray).map((_, 1)).reduceByKey(_ + _) 21 | 22 | System.out.println(charCounts.collect().mkString(", ")) 23 | } 24 | } -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/GroupWith.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.SparkContext 4 | import org.apache.spark.SparkContext._ 5 | 6 | object GroupWith { 7 | def main(args: Array[String]) { 8 | 9 | val sc = new SparkContext("local[2]", "GroupWith Test") 10 | 11 | val data1 = Array[(String, Int)](("A1", 1), ("A2", 2), 12 | ("B1", 3), ("B2", 4), 13 | ("C1", 5), ("C1", 6) 14 | ) 15 | 16 | val data2 = Array[(String, Int)](("A1", 7), ("A2", 8), 17 | ("B1", 9), ("C1", 0) 18 | ) 19 | val pairs1 = sc.parallelize(data1, 3) 20 | val pairs2 = sc.parallelize(data2, 2) 21 | 22 | val result = pairs1.groupWith(pairs2) 23 | result.foreach(println) 24 | 25 | // output: 26 | // (B1,(ArrayBuffer(3),ArrayBuffer(9))) 27 | // (A1,(ArrayBuffer(1),ArrayBuffer(7))) 28 | // (A2,(ArrayBuffer(2),ArrayBuffer(8))) 29 | // 30 | // (C1,(ArrayBuffer(5, 6),ArrayBuffer(0))) 31 | // (B2,(ArrayBuffer(4),ArrayBuffer())) 32 | 33 | 34 | } 35 | } -------------------------------------------------------------------------------- /src/main/python/PythonSparkSQLBySchema.py: -------------------------------------------------------------------------------- 1 | # Import SQLContext and data types 2 | from pyspark.sql import * 3 | 4 | # sc is an existing SparkContext. 5 | sqlContext = SQLContext(sc) 6 | 7 | # Load a text file and convert each line to a tuple. 8 | lines = sc.textFile("people.txt") 9 | parts = lines.map(lambda l: l.split(",")) 10 | people = parts.map(lambda p: (p[0], p[1].strip())) 11 | 12 | # The schema is encoded in a string. 13 | schemaString = "name age" 14 | 15 | fields = [StructField(field_name, StringType(), True) for field_name in schemaString.split()] 16 | schema = StructType(fields) 17 | 18 | # Apply the schema to the RDD. 19 | schemaPeople = sqlContext.createDataFrame(people, schema) 20 | 21 | # Register the DataFrame as a table. 22 | schemaPeople.registerTempTable("people") 23 | 24 | # SQL can be run over DataFrames that have been registered as a table. 25 | results = sqlContext.sql("SELECT name FROM people") 26 | 27 | # The results of SQL queries are RDDs and support all the normal RDD operations. 28 | names = results.map(lambda p: "Name: " + p.name) 29 | for name in names.collect(): 30 | print name -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/GroupByKey.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import java.util.Random 4 | 5 | import org.apache.spark.{SparkConf, SparkContext} 6 | import org.apache.spark.SparkContext._ 7 | 8 | /** 9 | * Usage: GroupByTest [numMappers] [numKVPairs] [KeySize] [numReducers] 10 | */ 11 | object GroupByKey { 12 | def main(args: Array[String]) { 13 | val sparkConf = new SparkConf().setAppName("GroupBy Test").setMaster("local[2]") 14 | var numMappers = 10 15 | var numKVPairs = 100 16 | var valSize = 100 17 | var numReducers = 3 18 | 19 | val sc = new SparkContext(sparkConf) 20 | 21 | val pairs1 = sc.parallelize(0 until numMappers, numMappers).flatMap { p => 22 | val ranGen = new Random 23 | var arr1 = new Array[(Int, Array[Byte])](numKVPairs) 24 | for (i <- 0 until numKVPairs) { 25 | val byteArr = new Array[Byte](valSize) 26 | ranGen.nextBytes(byteArr) 27 | arr1(i) = (ranGen.nextInt(10), byteArr) 28 | } 29 | arr1 30 | }.cache 31 | // Enforce that everything has been calculated and in cache 32 | pairs1.count 33 | 34 | val result = pairs1.groupByKey(numReducers) 35 | println(result.count) 36 | println(result.toDebugString) 37 | 38 | sc.stop() 39 | } 40 | } -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/LocalPi.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import scala.math.random 21 | 22 | import org.apache.spark._ 23 | import org.apache.spark.SparkContext._ 24 | 25 | object LocalPi { 26 | def main(args: Array[String]) { 27 | var count = 0 28 | for (i <- 1 to 100000) { 29 | val x = random * 2 - 1 30 | val y = random * 2 - 1 31 | if (x*x + y*y < 1) count += 1 32 | } 33 | println("Pi is roughly " + 4 * count / 100000.0) 34 | } 35 | } 36 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/ExceptionHandlingTest.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import org.apache.spark.{SparkConf, SparkContext} 21 | 22 | object ExceptionHandlingTest { 23 | def main(args: Array[String]) { 24 | val sparkConf = new SparkConf().setAppName("ExceptionHandlingTest") 25 | val sc = new SparkContext(sparkConf) 26 | sc.parallelize(0 until sc.defaultParallelism).foreach { i => 27 | if (math.random > 0.75) { 28 | throw new Exception("Testing exception handling") 29 | } 30 | } 31 | 32 | sc.stop() 33 | } 34 | } 35 | -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/Cartesian.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.SparkContext 4 | 5 | object Cartesian { 6 | def main(args: Array[String]) { 7 | val sc = new SparkContext("local", "Cartesian Test") 8 | val data1 = Array[(String, Int)](("A1", 1), ("A2", 2), 9 | ("B1", 3), ("B2", 4), 10 | ("C1", 5), ("C1", 6)) 11 | 12 | val data2 = Array[(String, Int)](("A1", 7), ("A2", 8), 13 | ("B1", 9), ("C1", 0)) 14 | val pairs1 = sc.parallelize(data1, 3) 15 | val pairs2 = sc.parallelize(data2, 2) 16 | 17 | val resultRDD = pairs1.cartesian(pairs2) 18 | 19 | resultRDD.foreach(println) 20 | 21 | /* 22 | * Output of task1: 23 | * ((A1,1),(A1,7)) 24 | * ((A1,1),(A2,8)) 25 | * ((A2,2),(A1,7)) 26 | * ((A2,2),(A2,8)) 27 | * Output of task2: 28 | * ((A1,1),(B1,9)) 29 | * ((A1,1),(C1,0)) 30 | * ((A2,2),(B1,9)) 31 | * ((A2,2),(C1,0)) 32 | * Output of task3: 33 | * ((B1,3),(A1,7)) 34 | * ((B1,3),(A2,8)) 35 | * ((B2,4),(A1,7)) 36 | * ((B2,4),(A2,8)) 37 | * Output of task4: 38 | * ((B1,3),(B1,9)) 39 | * ((B1,3),(C1,0)) 40 | * ((B2,4),(B1,9)) 41 | * ((B2,4),(C1,0)) 42 | * Output of task5: 43 | * ((C1,5),(A1,7)) 44 | * ((C1,5),(A2,8)) 45 | * ((C1,6),(A1,7)) 46 | * ((C1,6),(A2,8)) 47 | * Output of task6: 48 | * ((C1,5),(B1,9)) 49 | * ((C1,5),(C1,0)) 50 | * ((C1,6),(B1,9)) 51 | * ((C1,6),(C1,0)) 52 | */ 53 | 54 | } 55 | } -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/rdd/GroupByAction.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd 2 | 3 | import org.apache.spark.SparkContext 4 | import org.apache.spark.RangePartitioner 5 | 6 | object GroupByAction { 7 | def main(args: Array[String]) { 8 | 9 | val sc = new SparkContext("local", "GroupByAction Test") 10 | 11 | val data = Array[(String, Int)](("A1", 1), ("A2", 2), 12 | ("B1", 6), ("A2", 4), 13 | ("B1", 3), ("B1", 5)) 14 | 15 | val pairs = sc.parallelize(data, 3) 16 | 17 | // output: 18 | // (A1,1) 19 | // (A2,2) 20 | // 21 | // (B1,6) 22 | // (A2,4) 23 | // 24 | // (B1,3) 25 | // (B1,5) 26 | pairs.foreach(println) 27 | 28 | val result1 = pairs.groupBy(K => K._1) 29 | val result2 = pairs.groupBy((K : (String, Int)) => K._1, 1) 30 | val result3 = pairs.groupBy((K : (String, Int)) => K._1, new RangePartitioner(3, pairs)) 31 | 32 | // output of result1: 33 | // (A1,ArrayBuffer((A1,1))) 34 | // 35 | // (B1,ArrayBuffer((B1,6), (B1,3), (B1,5))) 36 | // (A2,ArrayBuffer((A2,2), (A2,4))) 37 | result1.foreach(println) 38 | 39 | // output of result2: 40 | // (A1,ArrayBuffer((A1,1))) 41 | // (B1,ArrayBuffer((B1,6), (B1,3), (B1,5))) 42 | // (A2,ArrayBuffer((A2,2), (A2,4))) 43 | result2.foreach(println) 44 | 45 | // output of result3: 46 | // (A1,ArrayBuffer((A1,1))) 47 | // (A2,ArrayBuffer((A2,2), (A2,4))) 48 | // 49 | // (B1,ArrayBuffer((B1,6), (B1,3), (B1,5))) 50 | result3.foreach(println) 51 | 52 | } 53 | 54 | } -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/sparksql/ScalaSparkSQLByReflection.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.sparksql 2 | 3 | import org.apache.spark.{SparkConf, SparkContext} 4 | 5 | object ScalaSparkSQLByReflection { 6 | 7 | // Define the schema using a case class. 8 | // Note: Case classes in Scala 2.10 can support only up to 22 fields. To work around this limit, 9 | // you can use custom classes that implement the Product interface. 10 | case class People(name: String, age: Int) 11 | 12 | def main(args: Array[String]) { 13 | val sc = new SparkContext(new SparkConf().setAppName("ScalaSparkSQL")) 14 | val sqlContext = new org.apache.spark.sql.SQLContext(sc) 15 | 16 | // this is used to implicitly convert an RDD to a DataFrame. 17 | import sqlContext.implicits._ 18 | 19 | // Create an RDD of People objects and register it as a table. 20 | val people = sc.textFile("people.txt").map(_.split(",")).map(p => People(p(0), p(1).trim.toInt)).toDF() 21 | people.registerTempTable("people") 22 | 23 | // SQL statements can be run by using the sql methods provided by sqlContext. 24 | val teenagers = sqlContext.sql("SELECT name FROM people WHERE age >= 13 AND age <= 19") 25 | 26 | // The results of SQL queries are DataFrames and support all the normal RDD operations. 27 | // The columns of a row in the result can be accessed by ordinal. 28 | teenagers.map(t => "Name: " + t(0)).collect().foreach(println) 29 | 30 | people.saveAsParquetFile("people.parquet") 31 | 32 | val parquetFile = sqlContext.parquetFile("people.parquet") 33 | } 34 | } -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/SparkPi.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import scala.math.random 21 | 22 | import org.apache.spark._ 23 | 24 | /** Computes an approximation to pi */ 25 | object SparkPi { 26 | def main(args: Array[String]) { 27 | val conf = new SparkConf().setAppName("Spark Pi") 28 | val spark = new SparkContext(conf) 29 | val slices = if (args.length > 0) args(0).toInt else 2 30 | val n = math.min(100000L * slices, Int.MaxValue).toInt // avoid overflow 31 | val count = spark.parallelize(1 until n, slices).map { i => 32 | val x = random * 2 - 1 33 | val y = random * 2 - 1 34 | if (x*x + y*y < 1) 1 else 0 35 | }.reduce(_ + _) 36 | println("Pi is roughly " + 4.0 * count / n) 37 | spark.stop() 38 | } 39 | } 40 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/HdfsTest.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import org.apache.spark._ 21 | 22 | 23 | object HdfsTest { 24 | 25 | /** Usage: HdfsTest [file] */ 26 | def main(args: Array[String]) { 27 | if (args.length < 1) { 28 | System.err.println("Usage: HdfsTest ") 29 | System.exit(1) 30 | } 31 | val sparkConf = new SparkConf().setAppName("HdfsTest") 32 | val sc = new SparkContext(sparkConf) 33 | val file = sc.textFile(args(0)) 34 | val mapped = file.map(s => s.length).cache() 35 | for (iter <- 1 to 10) { 36 | val start = System.currentTimeMillis() 37 | for (x <- mapped) { x + 2 } 38 | val end = System.currentTimeMillis() 39 | println("Iteration " + iter + " took " + (end-start) + " ms") 40 | } 41 | sc.stop() 42 | } 43 | } 44 | -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/sparksql/ScalaSparkSQLBySchema.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.sparksql 2 | 3 | import org.apache.spark.sql.types.{StringType, StructField, StructType} 4 | import org.apache.spark.{SparkConf, SparkContext} 5 | 6 | object ScalaSparkSQLBySchema { 7 | 8 | def main(args: Array[String]) { 9 | val sc = new SparkContext(new SparkConf().setAppName("ScalaSparkSQL")) 10 | val sqlContext = new org.apache.spark.sql.SQLContext(sc) 11 | 12 | // Create an RDD 13 | val people = sc.textFile("people.txt") 14 | 15 | // The schema is encoded in a string 16 | val schemaString = "name age" 17 | 18 | // Import Spark SQL data types and Row. 19 | import org.apache.spark.sql._ 20 | 21 | // Generate the schema based on the string of schema 22 | val schema = 23 | StructType( 24 | schemaString.split(" ").map(fieldName => StructField(fieldName, StringType, true))) 25 | 26 | // Convert records of the RDD (people) to Rows. 27 | val rowRDD = people.map(_.split(",")).map(p => Row(p(0), p(1).trim)) 28 | 29 | // Apply the schema to the RDD. 30 | val peopleDataFrame = sqlContext.createDataFrame(rowRDD, schema) 31 | 32 | // Register the DataFrames as a table. 33 | peopleDataFrame.registerTempTable("people") 34 | 35 | // SQL statements can be run by using the sql methods provided by sqlContext. 36 | val results = sqlContext.sql("SELECT name FROM people") 37 | 38 | // The results of SQL queries are DataFrames and support all the normal RDD operations. 39 | // The columns of a row in the result can be accessed by ordinal. 40 | results.map(t => "Name: " + t(0)).collect().foreach(println) 41 | } 42 | } -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/SparkTachyonPi.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import scala.math.random 21 | 22 | import org.apache.spark._ 23 | import org.apache.spark.storage.StorageLevel 24 | 25 | /** 26 | * Computes an approximation to pi 27 | * This example uses Tachyon to persist rdds during computation. 28 | */ 29 | object SparkTachyonPi { 30 | def main(args: Array[String]) { 31 | val sparkConf = new SparkConf().setAppName("SparkTachyonPi") 32 | val spark = new SparkContext(sparkConf) 33 | 34 | val slices = if (args.length > 0) args(0).toInt else 2 35 | val n = 100000 * slices 36 | 37 | val rdd = spark.parallelize(1 to n, slices) 38 | rdd.persist(StorageLevel.OFF_HEAP) 39 | val count = rdd.map { i => 40 | val x = random * 2 - 1 41 | val y = random * 2 - 1 42 | if (x * x + y * y < 1) 1 else 0 43 | }.reduce(_ + _) 44 | println("Pi is roughly " + 4.0 * count / n) 45 | 46 | spark.stop() 47 | } 48 | } 49 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/DriverSubmissionTest.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import scala.collection.JavaConversions._ 21 | 22 | import org.apache.spark.util.Utils 23 | 24 | /** Prints out environmental information, sleeps, and then exits. Made to 25 | * test driver submission in the standalone scheduler. */ 26 | object DriverSubmissionTest { 27 | def main(args: Array[String]) { 28 | if (args.size < 1) { 29 | println("Usage: DriverSubmissionTest ") 30 | System.exit(0) 31 | } 32 | val numSecondsToSleep = args(0).toInt 33 | 34 | val env = System.getenv() 35 | val properties = Utils.getSystemProperties 36 | 37 | println("Environment variables containing SPARK_TEST:") 38 | env.filter{case (k, v) => k.contains("SPARK_TEST")}.foreach(println) 39 | 40 | println("System properties containing spark.test:") 41 | properties.filter{case (k, v) => k.toString.contains("spark.test")}.foreach(println) 42 | 43 | for (i <- 1 until numSecondsToSleep) { 44 | println(s"Alive for $i out of $numSecondsToSleep seconds") 45 | Thread.sleep(1000) 46 | } 47 | } 48 | } 49 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/MultiBroadcastTest.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import org.apache.spark.rdd.RDD 21 | import org.apache.spark.{SparkConf, SparkContext} 22 | 23 | /** 24 | * Usage: MultiBroadcastTest [slices] [numElem] 25 | */ 26 | object MultiBroadcastTest { 27 | def main(args: Array[String]) { 28 | 29 | val sparkConf = new SparkConf().setAppName("Multi-Broadcast Test") 30 | val sc = new SparkContext(sparkConf) 31 | 32 | val slices = if (args.length > 0) args(0).toInt else 2 33 | val num = if (args.length > 1) args(1).toInt else 1000000 34 | 35 | val arr1 = new Array[Int](num) 36 | for (i <- 0 until arr1.length) { 37 | arr1(i) = i 38 | } 39 | 40 | val arr2 = new Array[Int](num) 41 | for (i <- 0 until arr2.length) { 42 | arr2(i) = i 43 | } 44 | 45 | val barr1 = sc.broadcast(arr1) 46 | val barr2 = sc.broadcast(arr2) 47 | val observedSizes: RDD[(Int, Int)] = sc.parallelize(1 to 10, slices).map { _ => 48 | (barr1.value.size, barr2.value.size) 49 | } 50 | // Collect the small RDD so we can print the observed sizes locally. 51 | observedSizes.collect().foreach(i => println(i)) 52 | 53 | sc.stop() 54 | } 55 | } 56 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/GroupByTest.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import java.util.Random 21 | 22 | import org.apache.spark.{SparkConf, SparkContext} 23 | import org.apache.spark.SparkContext._ 24 | 25 | /** 26 | * Usage: GroupByTest [numMappers] [numKVPairs] [KeySize] [numReducers] 27 | */ 28 | object GroupByTest { 29 | def main(args: Array[String]) { 30 | val sparkConf = new SparkConf().setAppName("GroupBy Test") 31 | var numMappers = if (args.length > 0) args(0).toInt else 2 32 | var numKVPairs = if (args.length > 1) args(1).toInt else 1000 33 | var valSize = if (args.length > 2) args(2).toInt else 1000 34 | var numReducers = if (args.length > 3) args(3).toInt else numMappers 35 | 36 | val sc = new SparkContext(sparkConf) 37 | 38 | val pairs1 = sc.parallelize(0 until numMappers, numMappers).flatMap { p => 39 | val ranGen = new Random 40 | var arr1 = new Array[(Int, Array[Byte])](numKVPairs) 41 | for (i <- 0 until numKVPairs) { 42 | val byteArr = new Array[Byte](valSize) 43 | ranGen.nextBytes(byteArr) 44 | arr1(i) = (ranGen.nextInt(Int.MaxValue), byteArr) 45 | } 46 | arr1 47 | }.cache() 48 | // Enforce that everything has been calculated and in cache 49 | pairs1.count() 50 | 51 | println(pairs1.groupByKey(numReducers).count()) 52 | 53 | sc.stop() 54 | } 55 | } 56 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/BroadcastTest.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import org.apache.spark.{SparkConf, SparkContext} 21 | 22 | /** 23 | * Usage: BroadcastTest [slices] [numElem] [broadcastAlgo] [blockSize] 24 | */ 25 | object BroadcastTest { 26 | def main(args: Array[String]) { 27 | 28 | val bcName = if (args.length > 2) args(2) else "Http" 29 | val blockSize = if (args.length > 3) args(3) else "4096" 30 | 31 | val sparkConf = new SparkConf().setAppName("Broadcast Test") 32 | .set("spark.broadcast.factory", s"org.apache.spark.broadcast.${bcName}BroadcastFactory") 33 | .set("spark.broadcast.blockSize", blockSize) 34 | val sc = new SparkContext(sparkConf) 35 | 36 | val slices = if (args.length > 0) args(0).toInt else 2 37 | val num = if (args.length > 1) args(1).toInt else 1000000 38 | 39 | val arr1 = (0 until num).toArray 40 | 41 | for (i <- 0 until 3) { 42 | println("Iteration " + i) 43 | println("===========") 44 | val startTime = System.nanoTime 45 | val barr1 = sc.broadcast(arr1) 46 | val observedSizes = sc.parallelize(1 to 10, slices).map(_ => barr1.value.size) 47 | // Collect the small RDD so we can print the observed sizes locally. 48 | observedSizes.collect().foreach(i => println(i)) 49 | println("Iteration %d took %.0f milliseconds".format(i, (System.nanoTime - startTime) / 1E6)) 50 | } 51 | 52 | sc.stop() 53 | } 54 | } 55 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/SkewedGroupByTest.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import java.util.Random 21 | 22 | import org.apache.spark.{SparkConf, SparkContext} 23 | import org.apache.spark.SparkContext._ 24 | 25 | /** 26 | * Usage: GroupByTest [numMappers] [numKVPairs] [KeySize] [numReducers] 27 | */ 28 | object SkewedGroupByTest { 29 | def main(args: Array[String]) { 30 | val sparkConf = new SparkConf().setAppName("GroupBy Test") 31 | var numMappers = if (args.length > 0) args(0).toInt else 2 32 | var numKVPairs = if (args.length > 1) args(1).toInt else 1000 33 | var valSize = if (args.length > 2) args(2).toInt else 1000 34 | var numReducers = if (args.length > 3) args(3).toInt else numMappers 35 | 36 | val sc = new SparkContext(sparkConf) 37 | 38 | val pairs1 = sc.parallelize(0 until numMappers, numMappers).flatMap { p => 39 | val ranGen = new Random 40 | 41 | // map output sizes lineraly increase from the 1st to the last 42 | numKVPairs = (1.0 * (p + 1) / numMappers * numKVPairs).toInt 43 | 44 | var arr1 = new Array[(Int, Array[Byte])](numKVPairs) 45 | for (i <- 0 until numKVPairs) { 46 | val byteArr = new Array[Byte](valSize) 47 | ranGen.nextBytes(byteArr) 48 | arr1(i) = (ranGen.nextInt(Int.MaxValue), byteArr) 49 | } 50 | arr1 51 | }.cache() 52 | // Enforce that everything has been calculated and in cache 53 | pairs1.count() 54 | 55 | println(pairs1.groupByKey(numReducers).count()) 56 | 57 | sc.stop() 58 | } 59 | } 60 | -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/mllib/EvaluateResult.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.grab 2 | 3 | import org.apache.spark.mllib.recommendation.{ALS, MatrixFactorizationModel, Rating} 4 | import org.apache.spark.rdd.RDD 5 | 6 | /** 7 | * 8 | * Created by june on 2015-05-27 09:13. 9 | */ 10 | object EvaluateResult { 11 | def coverage(training: RDD[Rating],userRecommends:RDD[(Int, List[Int])])={ 12 | userRecommends.flatMap(_._2).distinct().count.toDouble / training.map(_.product).distinct().count 13 | } 14 | 15 | def popularity(training: RDD[Rating],userRecommends:RDD[(Int, List[Int])])={ 16 | var ret = 0.0 17 | var n=0 18 | val item_popularity=training.map{ case Rating(user, product, rate) => 19 | (product,(user, rate)) 20 | }.groupByKey(4).map{case (product,list)=> 21 | (product,list.size) 22 | }.collectAsMap() 23 | 24 | userRecommends.flatMap(_._2).collect().foreach { p => 25 | ret = ret + math.log(1 + item_popularity.get(p).get) 26 | n = n + 1 27 | } 28 | 29 | ret/n 30 | } 31 | 32 | def recallAndPrecisionAndF1(training: RDD[Rating],userRecommends:RDD[(Int, List[Int])]):(Double, Double,Double) = { 33 | val usersProducts: RDD[(Int, Int)] = training.map { case Rating(user, product, rate) => 34 | (user, product) 35 | } 36 | 37 | val groupData=userRecommends.join(usersProducts.groupByKey().map {case (k,v) => (k,v.toList)}) 38 | 39 | val (hit, testNum, recNum) = groupData.map{ case (user, (mItems, tItems)) => 40 | var count = 0 41 | // 计算准确率：推荐命中商品数/实际推荐商品数, topN为推荐上限值 42 | val precNum = mItems.length 43 | for (i <- 0 until precNum) 44 | if (tItems.contains(mItems(i))) 45 | count += 1 46 | (count, tItems.length, precNum) }.reduce( (t1, t2) => (t1._1 + t2._1, t1._2 + t2._2, t1._3 + t2._3) ) 47 | 48 | val recall: Double = hit * 1.0 / testNum 49 | val precision: Double = hit * 1.0 / recNum 50 | val f1: Double = 2 * recall * precision / (recall + precision) 51 | 52 | println(s"$hit,$testNum,$recNum") 53 | (recall,precision,f1) 54 | } 55 | 56 | def recallAndPrecision(test:RDD[Rating],result:RDD[Rating]):Double = { 57 | val numHit: Long = result.intersection(test).count 58 | val recall: Double = numHit * 1.0 / test.count 59 | val precision: Double = numHit * 1.0 / result.count 60 | val f1: Double = 2 * recall * precision / (recall + precision) 61 | System.out.println("recall : " + recall + "\nprecision : " + precision + "\nf1 : " + f1) 62 | f1 63 | } 64 | } 65 | -------------------------------------------------------------------------------- /src/main/java/com/javachen/spark/examples/sparksql/JavaSparkSQLBySchema.java: -------------------------------------------------------------------------------- 1 | //package com.javachen.spark.examples.sparksql; 2 | // 3 | //import org.apache.spark.SparkConf; 4 | //import org.apache.spark.api.java.JavaRDD; 5 | //import org.apache.spark.api.java.JavaSparkContext; 6 | //import org.apache.spark.api.java.function.Function; 7 | //import org.apache.spark.sql.DataFrame; 8 | //import org.apache.spark.sql.Row; 9 | //import org.apache.spark.sql.SQLContext; 10 | // 11 | //import java.util.List; 12 | // 13 | //public class JavaSparkSQLBySchema { 14 | // public static void main(String[] args) throws Exception { 15 | // SparkConf sparkConf = new SparkConf().setAppName("JavaSparkSQLBySchema"); 16 | // JavaSparkContext ctx = new JavaSparkContext(sparkConf); 17 | // SQLContext sqlContext = new SQLContext(sc); 18 | // 19 | // // Load a text file and convert each line to a JavaBean. 20 | // JavaRDD people = sc.textFile("people.txt"); 21 | // 22 | // // The schema is encoded in a string 23 | // String schemaString = "name age"; 24 | // 25 | // // Generate the schema based on the string of schema 26 | // List fields = new ArrayList(); 27 | // for (String fieldName : schemaString.split(" ")) { 28 | // fields.add(DataType.createStructField(fieldName, DataType.StringType, true)); 29 | // } 30 | // StructType schema = DataType.createStructType(fields); 31 | // 32 | // // Convert records of the RDD (people) to Rows. 33 | // JavaRDD rowRDD = people.map( 34 | // new Function() { 35 | // public Row call(String record) throws Exception { 36 | // String[] fields = record.split(","); 37 | // return Row.create(fields[0], fields[1].trim()); 38 | // } 39 | // }); 40 | // 41 | // // Apply the schema to the RDD. 42 | // DataFrame peopleDataFrame = sqlContext.createDataFrame(rowRDD, schema); 43 | // 44 | // // Register the DataFrame as a table. 45 | // peopleDataFrame.registerTempTable("people"); 46 | // 47 | // // SQL can be run over RDDs that have been registered as tables. 48 | // DataFrame results = sqlContext.sql("SELECT name FROM people"); 49 | // 50 | // // The results of SQL queries are DataFrames and support all the normal RDD operations. 51 | // // The columns of a row in the result can be accessed by ordinal. 52 | // List names = results.map(new Function() { 53 | // public String call(Row row) { 54 | // return "Name: " + row.getString(0); 55 | // } 56 | // }).collect(); 57 | // } 58 | // 59 | //} -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/SparkTC.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import scala.util.Random 21 | import scala.collection.mutable 22 | 23 | import org.apache.spark.{SparkConf, SparkContext} 24 | import org.apache.spark.SparkContext._ 25 | 26 | /** 27 | * Transitive closure on a graph. 28 | */ 29 | object SparkTC { 30 | val numEdges = 200 31 | val numVertices = 100 32 | val rand = new Random(42) 33 | 34 | def generateGraph = { 35 | val edges: mutable.Set[(Int, Int)] = mutable.Set.empty 36 | while (edges.size < numEdges) { 37 | val from = rand.nextInt(numVertices) 38 | val to = rand.nextInt(numVertices) 39 | if (from != to) edges.+=((from, to)) 40 | } 41 | edges.toSeq 42 | } 43 | 44 | def main(args: Array[String]) { 45 | val sparkConf = new SparkConf().setAppName("SparkTC") 46 | val spark = new SparkContext(sparkConf) 47 | val slices = if (args.length > 0) args(0).toInt else 2 48 | var tc = spark.parallelize(generateGraph, slices).cache() 49 | 50 | // Linear transitive closure: each round grows paths by one edge, 51 | // by joining the graph's edges with the already-discovered paths. 52 | // e.g. join the path (y, z) from the TC with the edge (x, y) from 53 | // the graph to obtain the path (x, z). 54 | 55 | // Because join() joins on keys, the edges are stored in reversed order. 56 | val edges = tc.map(x => (x._2, x._1)) 57 | 58 | // This join is iterated until a fixed point is reached. 59 | var oldCount = 0L 60 | var nextCount = tc.count() 61 | do { 62 | oldCount = nextCount 63 | // Perform the join, obtaining an RDD of (y, (z, x)) pairs, 64 | // then project the result to obtain the new (x, z) paths. 65 | tc = tc.union(tc.join(edges).map(x => (x._2._2, x._2._1))).distinct().cache() 66 | nextCount = tc.count() 67 | } while (nextCount != oldCount) 68 | 69 | println("TC has " + tc.count() + " edges.") 70 | spark.stop() 71 | } 72 | } 73 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/LocalFileLR.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import java.util.Random 21 | 22 | import breeze.linalg.{Vector, DenseVector} 23 | 24 | /** 25 | * Logistic regression based classification. 26 | * 27 | * This is an example implementation for learning how to use Spark. For more conventional use, 28 | * please refer to either org.apache.spark.mllib.classification.LogisticRegressionWithSGD or 29 | * org.apache.spark.mllib.classification.LogisticRegressionWithLBFGS based on your needs. 30 | */ 31 | object LocalFileLR { 32 | val D = 10 // Numer of dimensions 33 | val rand = new Random(42) 34 | 35 | case class DataPoint(x: Vector[Double], y: Double) 36 | 37 | def parsePoint(line: String): DataPoint = { 38 | val nums = line.split(' ').map(_.toDouble) 39 | DataPoint(new DenseVector(nums.slice(1, D + 1)), nums(0)) 40 | } 41 | 42 | def showWarning() { 43 | System.err.println( 44 | """WARN: This is a naive implementation of Logistic Regression and is given as an example! 45 | |Please use either org.apache.spark.mllib.classification.LogisticRegressionWithSGD or 46 | |org.apache.spark.mllib.classification.LogisticRegressionWithLBFGS 47 | |for more conventional use. 48 | """.stripMargin) 49 | } 50 | 51 | def main(args: Array[String]) { 52 | 53 | showWarning() 54 | 55 | val lines = scala.io.Source.fromFile(args(0)).getLines().toArray 56 | val points = lines.map(parsePoint _) 57 | val ITERATIONS = args(1).toInt 58 | 59 | // Initialize w to a random value 60 | var w = DenseVector.fill(D){2 * rand.nextDouble - 1} 61 | println("Initial w: " + w) 62 | 63 | for (i <- 1 to ITERATIONS) { 64 | println("On iteration " + i) 65 | var gradient = DenseVector.zeros[Double](D) 66 | for (p <- points) { 67 | val scale = (1 / (1 + math.exp(-p.y * (w.dot(p.x)))) - 1) * p.y 68 | gradient += p.x * scale 69 | } 70 | w -= gradient 71 | } 72 | 73 | println("Final w: " + w) 74 | } 75 | } 76 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/SparkPageRank.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import org.apache.spark.SparkContext._ 21 | import org.apache.spark.{SparkConf, SparkContext} 22 | 23 | /** 24 | * Computes the PageRank of URLs from an input file. Input file should 25 | * be in format of: 26 | * URL neighbor URL 27 | * URL neighbor URL 28 | * URL neighbor URL 29 | * ... 30 | * where URL and their neighbors are separated by space(s). 31 | * 32 | * This is an example implementation for learning how to use Spark. For more conventional use, 33 | * please refer to org.apache.spark.graphx.lib.PageRank 34 | */ 35 | object SparkPageRank { 36 | 37 | def showWarning() { 38 | System.err.println( 39 | """WARN: This is a naive implementation of PageRank and is given as an example! 40 | |Please use the PageRank implementation found in org.apache.spark.graphx.lib.PageRank 41 | |for more conventional use. 42 | """.stripMargin) 43 | } 44 | 45 | def main(args: Array[String]) { 46 | if (args.length < 1) { 47 | System.err.println("Usage: SparkPageRank ") 48 | System.exit(1) 49 | } 50 | 51 | showWarning() 52 | 53 | val sparkConf = new SparkConf().setAppName("PageRank") 54 | val iters = if (args.length > 0) args(1).toInt else 10 55 | val ctx = new SparkContext(sparkConf) 56 | val lines = ctx.textFile(args(0), 1) 57 | val links = lines.map{ s => 58 | val parts = s.split("\\s+") 59 | (parts(0), parts(1)) 60 | }.distinct().groupByKey().cache() 61 | var ranks = links.mapValues(v => 1.0) 62 | 63 | for (i <- 1 to iters) { 64 | val contribs = links.join(ranks).values.flatMap{ case (urls, rank) => 65 | val size = urls.size 66 | urls.map(url => (url, rank / size)) 67 | } 68 | ranks = contribs.reduceByKey(_ + _).mapValues(0.15 + 0.85 * _) 69 | } 70 | 71 | val output = ranks.collect() 72 | output.foreach(tup => println(tup._1 + " has rank: " + tup._2 + ".")) 73 | 74 | ctx.stop() 75 | } 76 | } 77 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/LocalLR.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import java.util.Random 21 | 22 | import breeze.linalg.{Vector, DenseVector} 23 | 24 | /** 25 | * Logistic regression based classification. 26 | * 27 | * This is an example implementation for learning how to use Spark. For more conventional use, 28 | * please refer to either org.apache.spark.mllib.classification.LogisticRegressionWithSGD or 29 | * org.apache.spark.mllib.classification.LogisticRegressionWithLBFGS based on your needs. 30 | */ 31 | object LocalLR { 32 | val N = 10000 // Number of data points 33 | val D = 10 // Number of dimensions 34 | val R = 0.7 // Scaling factor 35 | val ITERATIONS = 5 36 | val rand = new Random(42) 37 | 38 | case class DataPoint(x: Vector[Double], y: Double) 39 | 40 | def generateData = { 41 | def generatePoint(i: Int) = { 42 | val y = if(i % 2 == 0) -1 else 1 43 | val x = DenseVector.fill(D){rand.nextGaussian + y * R} 44 | DataPoint(x, y) 45 | } 46 | Array.tabulate(N)(generatePoint) 47 | } 48 | 49 | def showWarning() { 50 | System.err.println( 51 | """WARN: This is a naive implementation of Logistic Regression and is given as an example! 52 | |Please use either org.apache.spark.mllib.classification.LogisticRegressionWithSGD or 53 | |org.apache.spark.mllib.classification.LogisticRegressionWithLBFGS 54 | |for more conventional use. 55 | """.stripMargin) 56 | } 57 | 58 | def main(args: Array[String]) { 59 | 60 | showWarning() 61 | 62 | val data = generateData 63 | // Initialize w to a random value 64 | var w = DenseVector.fill(D){2 * rand.nextDouble - 1} 65 | println("Initial w: " + w) 66 | 67 | for (i <- 1 to ITERATIONS) { 68 | println("On iteration " + i) 69 | var gradient = DenseVector.zeros[Double](D) 70 | for (p <- data) { 71 | val scale = (1 / (1 + math.exp(-p.y * (w.dot(p.x)))) - 1) * p.y 72 | gradient += p.x * scale 73 | } 74 | w -= gradient 75 | } 76 | 77 | println("Final w: " + w) 78 | } 79 | } 80 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/SimpleSkewedGroupByTest.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import java.util.Random 21 | 22 | import org.apache.spark.{SparkConf, SparkContext} 23 | import org.apache.spark.SparkContext._ 24 | 25 | /** 26 | * Usage: SimpleSkewedGroupByTest [numMappers] [numKVPairs] [valSize] [numReducers] [ratio] 27 | */ 28 | object SimpleSkewedGroupByTest { 29 | def main(args: Array[String]) { 30 | 31 | val sparkConf = new SparkConf().setAppName("SimpleSkewedGroupByTest") 32 | var numMappers = if (args.length > 0) args(0).toInt else 2 33 | var numKVPairs = if (args.length > 1) args(1).toInt else 1000 34 | var valSize = if (args.length > 2) args(2).toInt else 1000 35 | var numReducers = if (args.length > 3) args(3).toInt else numMappers 36 | var ratio = if (args.length > 4) args(4).toInt else 5.0 37 | 38 | val sc = new SparkContext(sparkConf) 39 | 40 | val pairs1 = sc.parallelize(0 until numMappers, numMappers).flatMap { p => 41 | val ranGen = new Random 42 | var result = new Array[(Int, Array[Byte])](numKVPairs) 43 | for (i <- 0 until numKVPairs) { 44 | val byteArr = new Array[Byte](valSize) 45 | ranGen.nextBytes(byteArr) 46 | val offset = ranGen.nextInt(1000) * numReducers 47 | if (ranGen.nextDouble < ratio / (numReducers + ratio - 1)) { 48 | // give ratio times higher chance of generating key 0 (for reducer 0) 49 | result(i) = (offset, byteArr) 50 | } else { 51 | // generate a key for one of the other reducers 52 | val key = 1 + ranGen.nextInt(numReducers-1) + offset 53 | result(i) = (key, byteArr) 54 | } 55 | } 56 | result 57 | }.cache 58 | // Enforce that everything has been calculated and in cache 59 | pairs1.count 60 | 61 | println("RESULT: " + pairs1.groupByKey(numReducers).count) 62 | // Print how many keys each reducer got (for debugging) 63 | // println("RESULT: " + pairs1.groupByKey(numReducers) 64 | // .map{case (k,v) => (k, v.size)} 65 | // .collectAsMap) 66 | 67 | sc.stop() 68 | } 69 | } 70 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/SparkLR.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import java.util.Random 21 | 22 | import scala.math.exp 23 | 24 | import breeze.linalg.{Vector, DenseVector} 25 | 26 | import org.apache.spark._ 27 | 28 | /** 29 | * Logistic regression based classification. 30 | * Usage: SparkLR [slices] 31 | * 32 | * This is an example implementation for learning how to use Spark. For more conventional use, 33 | * please refer to either org.apache.spark.mllib.classification.LogisticRegressionWithSGD or 34 | * org.apache.spark.mllib.classification.LogisticRegressionWithLBFGS based on your needs. 35 | */ 36 | object SparkLR { 37 | val N = 10000 // Number of data points 38 | val D = 10 // Numer of dimensions 39 | val R = 0.7 // Scaling factor 40 | val ITERATIONS = 5 41 | val rand = new Random(42) 42 | 43 | case class DataPoint(x: Vector[Double], y: Double) 44 | 45 | def generateData = { 46 | def generatePoint(i: Int) = { 47 | val y = if(i % 2 == 0) -1 else 1 48 | val x = DenseVector.fill(D){rand.nextGaussian + y * R} 49 | DataPoint(x, y) 50 | } 51 | Array.tabulate(N)(generatePoint) 52 | } 53 | 54 | def showWarning() { 55 | System.err.println( 56 | """WARN: This is a naive implementation of Logistic Regression and is given as an example! 57 | |Please use either org.apache.spark.mllib.classification.LogisticRegressionWithSGD or 58 | |org.apache.spark.mllib.classification.LogisticRegressionWithLBFGS 59 | |for more conventional use. 60 | """.stripMargin) 61 | } 62 | 63 | def main(args: Array[String]) { 64 | 65 | showWarning() 66 | 67 | val sparkConf = new SparkConf().setAppName("SparkLR") 68 | val sc = new SparkContext(sparkConf) 69 | val numSlices = if (args.length > 0) args(0).toInt else 2 70 | val points = sc.parallelize(generateData, numSlices).cache() 71 | 72 | // Initialize w to a random value 73 | var w = DenseVector.fill(D){2 * rand.nextDouble - 1} 74 | println("Initial w: " + w) 75 | 76 | for (i <- 1 to ITERATIONS) { 77 | println("On iteration " + i) 78 | val gradient = points.map { p => 79 | p.x * (1 / (1 + exp(-p.y * (w.dot(p.x)))) - 1) * p.y 80 | }.reduce(_ + _) 81 | w -= gradient 82 | } 83 | 84 | println("Final w: " + w) 85 | 86 | sc.stop() 87 | } 88 | } 89 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/SparkTachyonHdfsLR.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import java.util.Random 21 | 22 | import scala.math.exp 23 | 24 | import breeze.linalg.{Vector, DenseVector} 25 | import org.apache.hadoop.conf.Configuration 26 | 27 | import org.apache.spark._ 28 | import org.apache.spark.scheduler.InputFormatInfo 29 | import org.apache.spark.storage.StorageLevel 30 | 31 | 32 | /** 33 | * Logistic regression based classification. 34 | * This example uses Tachyon to persist rdds during computation. 35 | * 36 | * This is an example implementation for learning how to use Spark. For more conventional use, 37 | * please refer to either org.apache.spark.mllib.classification.LogisticRegressionWithSGD or 38 | * org.apache.spark.mllib.classification.LogisticRegressionWithLBFGS based on your needs. 39 | */ 40 | object SparkTachyonHdfsLR { 41 | val D = 10 // Numer of dimensions 42 | val rand = new Random(42) 43 | 44 | def showWarning() { 45 | System.err.println( 46 | """WARN: This is a naive implementation of Logistic Regression and is given as an example! 47 | |Please use either org.apache.spark.mllib.classification.LogisticRegressionWithSGD or 48 | |org.apache.spark.mllib.classification.LogisticRegressionWithLBFGS 49 | |for more conventional use. 50 | """.stripMargin) 51 | } 52 | 53 | case class DataPoint(x: Vector[Double], y: Double) 54 | 55 | def parsePoint(line: String): DataPoint = { 56 | val tok = new java.util.StringTokenizer(line, " ") 57 | var y = tok.nextToken.toDouble 58 | var x = new Array[Double](D) 59 | var i = 0 60 | while (i < D) { 61 | x(i) = tok.nextToken.toDouble; i += 1 62 | } 63 | DataPoint(new DenseVector(x), y) 64 | } 65 | 66 | def main(args: Array[String]) { 67 | 68 | showWarning() 69 | 70 | val inputPath = args(0) 71 | val sparkConf = new SparkConf().setAppName("SparkTachyonHdfsLR") 72 | val conf = new Configuration() 73 | val sc = new SparkContext(sparkConf, 74 | InputFormatInfo.computePreferredLocations( 75 | Seq(new InputFormatInfo(conf, classOf[org.apache.hadoop.mapred.TextInputFormat], inputPath)) 76 | )) 77 | val lines = sc.textFile(inputPath) 78 | val points = lines.map(parsePoint _).persist(StorageLevel.OFF_HEAP) 79 | val ITERATIONS = args(1).toInt 80 | 81 | // Initialize w to a random value 82 | var w = DenseVector.fill(D){2 * rand.nextDouble - 1} 83 | println("Initial w: " + w) 84 | 85 | for (i <- 1 to ITERATIONS) { 86 | println("On iteration " + i) 87 | val gradient = points.map { p => 88 | p.x * (1 / (1 + exp(-p.y * (w.dot(p.x)))) - 1) * p.y 89 | }.reduce(_ + _) 90 | w -= gradient 91 | } 92 | 93 | println("Final w: " + w) 94 | sc.stop() 95 | } 96 | } 97 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/SparkKMeans.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import breeze.linalg.{Vector, DenseVector, squaredDistance} 21 | 22 | import org.apache.spark.{SparkConf, SparkContext} 23 | import org.apache.spark.SparkContext._ 24 | 25 | /** 26 | * K-means clustering. 27 | * 28 | * This is an example implementation for learning how to use Spark. For more conventional use, 29 | * please refer to org.apache.spark.mllib.clustering.KMeans 30 | */ 31 | object SparkKMeans { 32 | 33 | def parseVector(line: String): Vector[Double] = { 34 | DenseVector(line.split(' ').map(_.toDouble)) 35 | } 36 | 37 | def closestPoint(p: Vector[Double], centers: Array[Vector[Double]]): Int = { 38 | var bestIndex = 0 39 | var closest = Double.PositiveInfinity 40 | 41 | for (i <- 0 until centers.length) { 42 | val tempDist = squaredDistance(p, centers(i)) 43 | if (tempDist < closest) { 44 | closest = tempDist 45 | bestIndex = i 46 | } 47 | } 48 | 49 | bestIndex 50 | } 51 | 52 | def showWarning() { 53 | System.err.println( 54 | """WARN: This is a naive implementation of KMeans Clustering and is given as an example! 55 | |Please use the KMeans method found in org.apache.spark.mllib.clustering 56 | |for more conventional use. 57 | """.stripMargin) 58 | } 59 | 60 | def main(args: Array[String]) { 61 | 62 | if (args.length < 3) { 63 | System.err.println("Usage: SparkKMeans ") 64 | System.exit(1) 65 | } 66 | 67 | showWarning() 68 | 69 | val sparkConf = new SparkConf().setAppName("SparkKMeans") 70 | val sc = new SparkContext(sparkConf) 71 | val lines = sc.textFile(args(0)) 72 | val data = lines.map(parseVector _).cache() 73 | val K = args(1).toInt 74 | val convergeDist = args(2).toDouble 75 | 76 | val kPoints = data.takeSample(withReplacement = false, K, 42).toArray 77 | var tempDist = 1.0 78 | 79 | while(tempDist > convergeDist) { 80 | val closest = data.map (p => (closestPoint(p, kPoints), (p, 1))) 81 | 82 | val pointStats = closest.reduceByKey{case ((x1, y1), (x2, y2)) => (x1 + x2, y1 + y2)} 83 | 84 | val newPoints = pointStats.map {pair => 85 | (pair._1, pair._2._1 * (1.0 / pair._2._2))}.collectAsMap() 86 | 87 | tempDist = 0.0 88 | for (i <- 0 until K) { 89 | tempDist += squaredDistance(kPoints(i), newPoints(i)) 90 | } 91 | 92 | for (newP <- newPoints) { 93 | kPoints(newP._1) = newP._2 94 | } 95 | println("Finished iteration (delta = " + tempDist + ")") 96 | } 97 | 98 | println("Final centers:") 99 | kPoints.foreach(println) 100 | sc.stop() 101 | } 102 | } 103 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/SparkHdfsLR.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import java.util.Random 21 | 22 | import scala.math.exp 23 | 24 | import breeze.linalg.{Vector, DenseVector} 25 | import org.apache.hadoop.conf.Configuration 26 | 27 | import org.apache.spark._ 28 | import org.apache.spark.scheduler.InputFormatInfo 29 | 30 | 31 | /** 32 | * Logistic regression based classification. 33 | * 34 | * This is an example implementation for learning how to use Spark. For more conventional use, 35 | * please refer to either org.apache.spark.mllib.classification.LogisticRegressionWithSGD or 36 | * org.apache.spark.mllib.classification.LogisticRegressionWithLBFGS based on your needs. 37 | */ 38 | object SparkHdfsLR { 39 | val D = 10 // Numer of dimensions 40 | val rand = new Random(42) 41 | 42 | case class DataPoint(x: Vector[Double], y: Double) 43 | 44 | def parsePoint(line: String): DataPoint = { 45 | val tok = new java.util.StringTokenizer(line, " ") 46 | var y = tok.nextToken.toDouble 47 | var x = new Array[Double](D) 48 | var i = 0 49 | while (i < D) { 50 | x(i) = tok.nextToken.toDouble; i += 1 51 | } 52 | DataPoint(new DenseVector(x), y) 53 | } 54 | 55 | def showWarning() { 56 | System.err.println( 57 | """WARN: This is a naive implementation of Logistic Regression and is given as an example! 58 | |Please use either org.apache.spark.mllib.classification.LogisticRegressionWithSGD or 59 | |org.apache.spark.mllib.classification.LogisticRegressionWithLBFGS 60 | |for more conventional use. 61 | """.stripMargin) 62 | } 63 | 64 | def main(args: Array[String]) { 65 | 66 | if (args.length < 2) { 67 | System.err.println("Usage: SparkHdfsLR ") 68 | System.exit(1) 69 | } 70 | 71 | showWarning() 72 | 73 | val sparkConf = new SparkConf().setAppName("SparkHdfsLR") 74 | val inputPath = args(0) 75 | val conf = new Configuration() 76 | val sc = new SparkContext(sparkConf, 77 | InputFormatInfo.computePreferredLocations( 78 | Seq(new InputFormatInfo(conf, classOf[org.apache.hadoop.mapred.TextInputFormat], inputPath)) 79 | )) 80 | val lines = sc.textFile(inputPath) 81 | val points = lines.map(parsePoint _).cache() 82 | val ITERATIONS = args(1).toInt 83 | 84 | // Initialize w to a random value 85 | var w = DenseVector.fill(D){2 * rand.nextDouble - 1} 86 | println("Initial w: " + w) 87 | 88 | for (i <- 1 to ITERATIONS) { 89 | println("On iteration " + i) 90 | val gradient = points.map { p => 91 | p.x * (1 / (1 + exp(-p.y * (w.dot(p.x)))) - 1) * p.y 92 | }.reduce(_ + _) 93 | w -= gradient 94 | } 95 | 96 | println("Final w: " + w) 97 | sc.stop() 98 | } 99 | } 100 | -------------------------------------------------------------------------------- /src/main/java/com/javachen/spark/examples/mllib/JavaALS.java: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.mllib; 2 | 3 | import org.apache.spark.SparkConf; 4 | import org.apache.spark.api.java.JavaDoubleRDD; 5 | import org.apache.spark.api.java.JavaPairRDD; 6 | import org.apache.spark.api.java.JavaRDD; 7 | import org.apache.spark.api.java.JavaSparkContext; 8 | import org.apache.spark.api.java.function.Function; 9 | import org.apache.spark.mllib.recommendation.ALS; 10 | import org.apache.spark.mllib.recommendation.MatrixFactorizationModel; 11 | import org.apache.spark.mllib.recommendation.Rating; 12 | import scala.Tuple2; 13 | 14 | public class JavaALS { 15 | public static void main(String[] args) { 16 | SparkConf conf = new SparkConf().setAppName("Java Collaborative Filtering Example"); 17 | JavaSparkContext sc = new JavaSparkContext(conf); 18 | 19 | // Load and parse the data 20 | String path = "data/mllib/als/test.data"; 21 | JavaRDD data = sc.textFile(path); 22 | JavaRDD ratings = data.map( 23 | new Function() { 24 | public Rating call(String s) { 25 | String[] sarray = s.split(","); 26 | return new Rating(Integer.parseInt(sarray[0]), Integer.parseInt(sarray[1]), 27 | Double.parseDouble(sarray[2])); 28 | } 29 | } 30 | ); 31 | 32 | // Build the recommendation model using ALS 33 | int rank = 10; 34 | int numIterations = 20; 35 | MatrixFactorizationModel model = ALS.train(JavaRDD.toRDD(ratings), rank, numIterations, 0.01); 36 | 37 | // Evaluate the model on rating data 38 | JavaRDD> userProducts = ratings.map( 39 | new Function>() { 40 | public Tuple2 call(Rating r) { 41 | return new Tuple2(r.user(), r.product()); 42 | } 43 | } 44 | ); 45 | JavaPairRDD, Double> predictions = JavaPairRDD.fromJavaRDD( 46 | model.predict(JavaRDD.toRDD(userProducts)).toJavaRDD().map( 47 | new Function, Double>>() { 48 | public Tuple2, Double> call(Rating r) { 49 | return new Tuple2, Double>( 50 | new Tuple2(r.user(), r.product()), r.rating()); 51 | } 52 | } 53 | )); 54 | JavaRDD> ratesAndPreds = 55 | JavaPairRDD.fromJavaRDD(ratings.map( 56 | new Function, Double>>() { 57 | public Tuple2, Double> call(Rating r) { 58 | return new Tuple2, Double>( 59 | new Tuple2(r.user(), r.product()), r.rating()); 60 | } 61 | } 62 | )).join(predictions).values(); 63 | 64 | double MSE = JavaDoubleRDD.fromRDD(ratesAndPreds.map( 65 | new Function, Object>() { 66 | public Object call(Tuple2 pair) { 67 | Double err = pair._1() - pair._2(); 68 | return err * err; 69 | } 70 | } 71 | ).rdd()).mean(); 72 | System.out.println("Mean Squared Error = " + MSE); 73 | } 74 | } -------------------------------------------------------------------------------- /src/main/java/com/javachen/spark/examples/rdd/JavaWordCount.java: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.rdd; 2 | 3 | import org.apache.spark.SparkConf; 4 | import org.apache.spark.api.java.JavaPairRDD; 5 | import org.apache.spark.api.java.JavaRDD; 6 | import org.apache.spark.api.java.JavaSparkContext; 7 | import org.apache.spark.api.java.function.FlatMapFunction; 8 | import org.apache.spark.api.java.function.Function; 9 | import org.apache.spark.api.java.function.Function2; 10 | import org.apache.spark.api.java.function.PairFunction; 11 | import scala.Tuple2; 12 | 13 | import java.util.ArrayList; 14 | import java.util.Arrays; 15 | import java.util.Collection; 16 | 17 | /** 18 | * @author june. 19 | * @date 2015-05-06 16:20. 20 | */ 21 | public class JavaWordCount { 22 | public static void main(String[] args) { 23 | JavaSparkContext sc = new JavaSparkContext(new SparkConf().setAppName("JavaWordCount")); 24 | final int threshold = Integer.parseInt(args[1]); 25 | 26 | // split each document into words 27 | JavaRDD tokenized = sc.textFile(args[0]).flatMap( 28 | new FlatMapFunction() { 29 | @Override 30 | public Iterable call(String s) { 31 | return Arrays.asList(s.split(" ")); 32 | } 33 | } 34 | ); 35 | 36 | // count the occurrence of each word 37 | JavaPairRDD counts = tokenized.mapToPair( 38 | new PairFunction() { 39 | @Override 40 | public Tuple2 call(String s) { 41 | return new Tuple2(s, 1); 42 | } 43 | } 44 | ).reduceByKey( 45 | new Function2() { 46 | @Override 47 | public Integer call(Integer i1, Integer i2) { 48 | return i1 + i2; 49 | } 50 | } 51 | ); 52 | 53 | // filter out words with less than threshold occurrences 54 | JavaPairRDD filtered = counts.filter( 55 | new Function, Boolean>() { 56 | @Override 57 | public Boolean call(Tuple2 tup) { 58 | return tup._2() >= threshold; 59 | } 60 | } 61 | ); 62 | 63 | // count characters 64 | JavaPairRDD charCounts = filtered.flatMap( 65 | new FlatMapFunction, Character>() { 66 | @Override 67 | public Iterable call(Tuple2 s) { 68 | Collection chars = new ArrayList(s._1().length()); 69 | for (char c : s._1().toCharArray()) { 70 | chars.add(c); 71 | } 72 | return chars; 73 | } 74 | } 75 | ).mapToPair( 76 | new PairFunction() { 77 | @Override 78 | public Tuple2 call(Character c) { 79 | return new Tuple2(c, 1); 80 | } 81 | } 82 | ).reduceByKey( 83 | new Function2() { 84 | @Override 85 | public Integer call(Integer i1, Integer i2) { 86 | return i1 + i2; 87 | } 88 | } 89 | ); 90 | 91 | System.out.println(charCounts.collect()); 92 | } 93 | } -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/LocalKMeans.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import java.util.Random 21 | 22 | import scala.collection.mutable.HashMap 23 | import scala.collection.mutable.HashSet 24 | 25 | import breeze.linalg.{Vector, DenseVector, squaredDistance} 26 | 27 | import org.apache.spark.SparkContext._ 28 | 29 | /** 30 | * K-means clustering. 31 | * 32 | * This is an example implementation for learning how to use Spark. For more conventional use, 33 | * please refer to org.apache.spark.mllib.clustering.KMeans 34 | */ 35 | object LocalKMeans { 36 | val N = 1000 37 | val R = 1000 // Scaling factor 38 | val D = 10 39 | val K = 10 40 | val convergeDist = 0.001 41 | val rand = new Random(42) 42 | 43 | def generateData = { 44 | def generatePoint(i: Int) = { 45 | DenseVector.fill(D){rand.nextDouble * R} 46 | } 47 | Array.tabulate(N)(generatePoint) 48 | } 49 | 50 | def closestPoint(p: Vector[Double], centers: HashMap[Int, Vector[Double]]): Int = { 51 | var index = 0 52 | var bestIndex = 0 53 | var closest = Double.PositiveInfinity 54 | 55 | for (i <- 1 to centers.size) { 56 | val vCurr = centers.get(i).get 57 | val tempDist = squaredDistance(p, vCurr) 58 | if (tempDist < closest) { 59 | closest = tempDist 60 | bestIndex = i 61 | } 62 | } 63 | 64 | bestIndex 65 | } 66 | 67 | def showWarning() { 68 | System.err.println( 69 | """WARN: This is a naive implementation of KMeans Clustering and is given as an example! 70 | |Please use the KMeans method found in org.apache.spark.mllib.clustering 71 | |for more conventional use. 72 | """.stripMargin) 73 | } 74 | 75 | def main(args: Array[String]) { 76 | 77 | showWarning() 78 | 79 | val data = generateData 80 | var points = new HashSet[Vector[Double]] 81 | var kPoints = new HashMap[Int, Vector[Double]] 82 | var tempDist = 1.0 83 | 84 | while (points.size < K) { 85 | points.add(data(rand.nextInt(N))) 86 | } 87 | 88 | val iter = points.iterator 89 | for (i <- 1 to points.size) { 90 | kPoints.put(i, iter.next()) 91 | } 92 | 93 | println("Initial centers: " + kPoints) 94 | 95 | while(tempDist > convergeDist) { 96 | var closest = data.map (p => (closestPoint(p, kPoints), (p, 1))) 97 | 98 | var mappings = closest.groupBy[Int] (x => x._1) 99 | 100 | var pointStats = mappings.map { pair => 101 | pair._2.reduceLeft [(Int, (Vector[Double], Int))] { 102 | case ((id1, (x1, y1)), (id2, (x2, y2))) => (id1, (x1 + x2, y1 + y2)) 103 | } 104 | } 105 | 106 | var newPoints = pointStats.map {mapping => 107 | (mapping._1, mapping._2._1 * (1.0 / mapping._2._2))} 108 | 109 | tempDist = 0.0 110 | for (mapping <- newPoints) { 111 | tempDist += squaredDistance(kPoints.get(mapping._1).get, mapping._2) 112 | } 113 | 114 | for (newP <- newPoints) { 115 | kPoints.put(newP._1, newP._2) 116 | } 117 | } 118 | 119 | println("Final centers: " + kPoints) 120 | } 121 | } 122 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/LogQuery.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import org.apache.spark.{SparkConf, SparkContext} 21 | import org.apache.spark.SparkContext._ 22 | 23 | /** 24 | * Executes a roll up-style query against Apache logs. 25 | * 26 | * Usage: LogQuery [logFile] 27 | */ 28 | object LogQuery { 29 | val exampleApacheLogs = List( 30 | """10.10.10.10 - "FRED" [18/Jan/2013:17:56:07 +1100] "GET http://images.com/2013/Generic.jpg 31 | | HTTP/1.1" 304 315 "http://referall.com/" "Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; 32 | | GTB7.4; .NET CLR 2.0.50727; .NET CLR 3.0.04506.30; .NET CLR 3.0.04506.648; .NET CLR 33 | | 3.5.21022; .NET CLR 3.0.4506.2152; .NET CLR 1.0.3705; .NET CLR 1.1.4322; .NET CLR 34 | | 3.5.30729; Release=ARP)" "UD-1" - "image/jpeg" "whatever" 0.350 "-" - "" 265 923 934 "" 35 | | 62.24.11.25 images.com 1358492167 - Whatup""".stripMargin.lines.mkString, 36 | """10.10.10.10 - "FRED" [18/Jan/2013:18:02:37 +1100] "GET http://images.com/2013/Generic.jpg 37 | | HTTP/1.1" 304 306 "http:/referall.com" "Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; 38 | | GTB7.4; .NET CLR 2.0.50727; .NET CLR 3.0.04506.30; .NET CLR 3.0.04506.648; .NET CLR 39 | | 3.5.21022; .NET CLR 3.0.4506.2152; .NET CLR 1.0.3705; .NET CLR 1.1.4322; .NET CLR 40 | | 3.5.30729; Release=ARP)" "UD-1" - "image/jpeg" "whatever" 0.352 "-" - "" 256 977 988 "" 41 | | 0 73.23.2.15 images.com 1358492557 - Whatup""".stripMargin.lines.mkString 42 | ) 43 | 44 | def main(args: Array[String]) { 45 | 46 | val sparkConf = new SparkConf().setAppName("Log Query") 47 | val sc = new SparkContext(sparkConf) 48 | 49 | val dataSet = 50 | if (args.length == 1) sc.textFile(args(0)) else sc.parallelize(exampleApacheLogs) 51 | // scalastyle:off 52 | val apacheLogRegex = 53 | """^([\d.]+) (\S+) (\S+) \[([\w\d:/]+\s[+\-]\d{4})\] "(.+?)" (\d{3}) ([\d\-]+) "([^"]+)" "([^"]+)".*""".r 54 | // scalastyle:on 55 | /** Tracks the total query count and number of aggregate bytes for a particular group. */ 56 | class Stats(val count: Int, val numBytes: Int) extends Serializable { 57 | def merge(other: Stats) = new Stats(count + other.count, numBytes + other.numBytes) 58 | override def toString = "bytes=%s\tn=%s".format(numBytes, count) 59 | } 60 | 61 | def extractKey(line: String): (String, String, String) = { 62 | apacheLogRegex.findFirstIn(line) match { 63 | case Some(apacheLogRegex(ip, _, user, dateTime, query, status, bytes, referer, ua)) => 64 | if (user != "\"-\"") (ip, user, query) 65 | else (null, null, null) 66 | case _ => (null, null, null) 67 | } 68 | } 69 | 70 | def extractStats(line: String): Stats = { 71 | apacheLogRegex.findFirstIn(line) match { 72 | case Some(apacheLogRegex(ip, _, user, dateTime, query, status, bytes, referer, ua)) => 73 | new Stats(1, bytes.toInt) 74 | case _ => new Stats(1, 0) 75 | } 76 | } 77 | 78 | dataSet.map(line => (extractKey(line), extractStats(line))) 79 | .reduceByKey((a, b) => a.merge(b)) 80 | .collect().foreach{ 81 | case (user, query) => println("%s\t%s".format(user, query))} 82 | 83 | sc.stop() 84 | } 85 | } 86 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/LocalALS.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import org.apache.commons.math3.linear._ 21 | 22 | /** 23 | * Alternating least squares matrix factorization. 24 | * 25 | * This is an example implementation for learning how to use Spark. For more conventional use, 26 | * please refer to org.apache.spark.mllib.recommendation.ALS 27 | */ 28 | object LocalALS { 29 | 30 | // Parameters set through command line arguments 31 | var M = 0 // Number of movies 32 | var U = 0 // Number of users 33 | var F = 0 // Number of features 34 | var ITERATIONS = 0 35 | val LAMBDA = 0.01 // Regularization coefficient 36 | 37 | def generateR(): RealMatrix = { 38 | val mh = randomMatrix(M, F) 39 | val uh = randomMatrix(U, F) 40 | mh.multiply(uh.transpose()) 41 | } 42 | 43 | def rmse(targetR: RealMatrix, ms: Array[RealVector], us: Array[RealVector]): Double = { 44 | val r = new Array2DRowRealMatrix(M, U) 45 | for (i <- 0 until M; j <- 0 until U) { 46 | r.setEntry(i, j, ms(i).dotProduct(us(j))) 47 | } 48 | val diffs = r.subtract(targetR) 49 | var sumSqs = 0.0 50 | for (i <- 0 until M; j <- 0 until U) { 51 | val diff = diffs.getEntry(i, j) 52 | sumSqs += diff * diff 53 | } 54 | math.sqrt(sumSqs / (M.toDouble * U.toDouble)) 55 | } 56 | 57 | def updateMovie(i: Int, m: RealVector, us: Array[RealVector], R: RealMatrix) : RealVector = { 58 | var XtX: RealMatrix = new Array2DRowRealMatrix(F, F) 59 | var Xty: RealVector = new ArrayRealVector(F) 60 | // For each user that rated the movie 61 | for (j <- 0 until U) { 62 | val u = us(j) 63 | // Add u * u^t to XtX 64 | XtX = XtX.add(u.outerProduct(u)) 65 | // Add u * rating to Xty 66 | Xty = Xty.add(u.mapMultiply(R.getEntry(i, j))) 67 | } 68 | // Add regularization coefficients to diagonal terms 69 | for (d <- 0 until F) { 70 | XtX.addToEntry(d, d, LAMBDA * U) 71 | } 72 | // Solve it with Cholesky 73 | new CholeskyDecomposition(XtX).getSolver.solve(Xty) 74 | } 75 | 76 | def updateUser(j: Int, u: RealVector, ms: Array[RealVector], R: RealMatrix) : RealVector = { 77 | var XtX: RealMatrix = new Array2DRowRealMatrix(F, F) 78 | var Xty: RealVector = new ArrayRealVector(F) 79 | // For each movie that the user rated 80 | for (i <- 0 until M) { 81 | val m = ms(i) 82 | // Add m * m^t to XtX 83 | XtX = XtX.add(m.outerProduct(m)) 84 | // Add m * rating to Xty 85 | Xty = Xty.add(m.mapMultiply(R.getEntry(i, j))) 86 | } 87 | // Add regularization coefficients to diagonal terms 88 | for (d <- 0 until F) { 89 | XtX.addToEntry(d, d, LAMBDA * M) 90 | } 91 | // Solve it with Cholesky 92 | new CholeskyDecomposition(XtX).getSolver.solve(Xty) 93 | } 94 | 95 | def showWarning() { 96 | System.err.println( 97 | """WARN: This is a naive implementation of ALS and is given as an example! 98 | |Please use the ALS method found in org.apache.spark.mllib.recommendation 99 | |for more conventional use. 100 | """.stripMargin) 101 | } 102 | 103 | def main(args: Array[String]) { 104 | 105 | args match { 106 | case Array(m, u, f, iters) => { 107 | M = m.toInt 108 | U = u.toInt 109 | F = f.toInt 110 | ITERATIONS = iters.toInt 111 | } 112 | case _ => { 113 | System.err.println("Usage: LocalALS ") 114 | System.exit(1) 115 | } 116 | } 117 | 118 | showWarning() 119 | 120 | println(s"Running with M=$M, U=$U, F=$F, iters=$ITERATIONS") 121 | 122 | val R = generateR() 123 | 124 | // Initialize m and u randomly 125 | var ms = Array.fill(M)(randomVector(F)) 126 | var us = Array.fill(U)(randomVector(F)) 127 | 128 | // Iteratively update movies then users 129 | for (iter <- 1 to ITERATIONS) { 130 | println(s"Iteration $iter:") 131 | ms = (0 until M).map(i => updateMovie(i, ms(i), us, R)).toArray 132 | us = (0 until U).map(j => updateUser(j, us(j), ms, R)).toArray 133 | println("RMSE = " + rmse(R, ms, us)) 134 | println() 135 | } 136 | } 137 | 138 | private def randomVector(n: Int): RealVector = 139 | new ArrayRealVector(Array.fill(n)(math.random)) 140 | 141 | private def randomMatrix(rows: Int, cols: Int): RealMatrix = 142 | new Array2DRowRealMatrix(Array.fill(rows, cols)(math.random)) 143 | 144 | } 145 | -------------------------------------------------------------------------------- /src/main/scala/org/apache/spark/examples/SparkALS.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package org.apache.spark.examples 19 | 20 | import org.apache.commons.math3.linear._ 21 | 22 | import org.apache.spark._ 23 | 24 | /** 25 | * Alternating least squares matrix factorization. 26 | * 27 | * This is an example implementation for learning how to use Spark. For more conventional use, 28 | * please refer to org.apache.spark.mllib.recommendation.ALS 29 | */ 30 | object SparkALS { 31 | 32 | // Parameters set through command line arguments 33 | var M = 0 // Number of movies 34 | var U = 0 // Number of users 35 | var F = 0 // Number of features 36 | var ITERATIONS = 0 37 | val LAMBDA = 0.01 // Regularization coefficient 38 | 39 | def generateR(): RealMatrix = { 40 | val mh = randomMatrix(M, F) 41 | val uh = randomMatrix(U, F) 42 | mh.multiply(uh.transpose()) 43 | } 44 | 45 | def rmse(targetR: RealMatrix, ms: Array[RealVector], us: Array[RealVector]): Double = { 46 | val r = new Array2DRowRealMatrix(M, U) 47 | for (i <- 0 until M; j <- 0 until U) { 48 | r.setEntry(i, j, ms(i).dotProduct(us(j))) 49 | } 50 | val diffs = r.subtract(targetR) 51 | var sumSqs = 0.0 52 | for (i <- 0 until M; j <- 0 until U) { 53 | val diff = diffs.getEntry(i, j) 54 | sumSqs += diff * diff 55 | } 56 | math.sqrt(sumSqs / (M.toDouble * U.toDouble)) 57 | } 58 | 59 | def update(i: Int, m: RealVector, us: Array[RealVector], R: RealMatrix) : RealVector = { 60 | val U = us.size 61 | val F = us(0).getDimension 62 | var XtX: RealMatrix = new Array2DRowRealMatrix(F, F) 63 | var Xty: RealVector = new ArrayRealVector(F) 64 | // For each user that rated the movie 65 | for (j <- 0 until U) { 66 | val u = us(j) 67 | // Add u * u^t to XtX 68 | XtX = XtX.add(u.outerProduct(u)) 69 | // Add u * rating to Xty 70 | Xty = Xty.add(u.mapMultiply(R.getEntry(i, j))) 71 | } 72 | // Add regularization coefs to diagonal terms 73 | for (d <- 0 until F) { 74 | XtX.addToEntry(d, d, LAMBDA * U) 75 | } 76 | // Solve it with Cholesky 77 | new CholeskyDecomposition(XtX).getSolver.solve(Xty) 78 | } 79 | 80 | def showWarning() { 81 | System.err.println( 82 | """WARN: This is a naive implementation of ALS and is given as an example! 83 | |Please use the ALS method found in org.apache.spark.mllib.recommendation 84 | |for more conventional use. 85 | """.stripMargin) 86 | } 87 | 88 | def main(args: Array[String]) { 89 | 90 | var slices = 0 91 | 92 | val options = (0 to 4).map(i => if (i < args.length) Some(args(i)) else None) 93 | 94 | options.toArray match { 95 | case Array(m, u, f, iters, slices_) => 96 | M = m.getOrElse("100").toInt 97 | U = u.getOrElse("500").toInt 98 | F = f.getOrElse("10").toInt 99 | ITERATIONS = iters.getOrElse("5").toInt 100 | slices = slices_.getOrElse("2").toInt 101 | case _ => 102 | System.err.println("Usage: SparkALS [M] [U] [F] [iters] [slices]") 103 | System.exit(1) 104 | } 105 | 106 | showWarning() 107 | 108 | println(s"Running with M=$M, U=$U, F=$F, iters=$ITERATIONS") 109 | 110 | val sparkConf = new SparkConf().setAppName("SparkALS") 111 | val sc = new SparkContext(sparkConf) 112 | 113 | val R = generateR() 114 | 115 | // Initialize m and u randomly 116 | var ms = Array.fill(M)(randomVector(F)) 117 | var us = Array.fill(U)(randomVector(F)) 118 | 119 | // Iteratively update movies then users 120 | val Rc = sc.broadcast(R) 121 | var msb = sc.broadcast(ms) 122 | var usb = sc.broadcast(us) 123 | for (iter <- 1 to ITERATIONS) { 124 | println(s"Iteration $iter:") 125 | ms = sc.parallelize(0 until M, slices) 126 | .map(i => update(i, msb.value(i), usb.value, Rc.value)) 127 | .collect() 128 | msb = sc.broadcast(ms) // Re-broadcast ms because it was updated 129 | us = sc.parallelize(0 until U, slices) 130 | .map(i => update(i, usb.value(i), msb.value, Rc.value.transpose())) 131 | .collect() 132 | usb = sc.broadcast(us) // Re-broadcast us because it was updated 133 | println("RMSE = " + rmse(R, ms, us)) 134 | println() 135 | } 136 | 137 | sc.stop() 138 | } 139 | 140 | private def randomVector(n: Int): RealVector = 141 | new ArrayRealVector(Array.fill(n)(math.random)) 142 | 143 | private def randomMatrix(rows: Int, cols: Int): RealMatrix = 144 | new Array2DRowRealMatrix(Array.fill(rows, cols)(math.random)) 145 | 146 | } 147 | -------------------------------------------------------------------------------- /src/main/java/com/javachen/spark/examples/sparksql/JavaSparkSQLByReflection.java: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.sparksql; 2 | 3 | import org.apache.spark.SparkConf; 4 | import org.apache.spark.api.java.JavaRDD; 5 | import org.apache.spark.api.java.JavaSparkContext; 6 | import org.apache.spark.api.java.function.Function; 7 | import org.apache.spark.sql.DataFrame; 8 | import org.apache.spark.sql.Row; 9 | import org.apache.spark.sql.SQLContext; 10 | 11 | import java.io.Serializable; 12 | import java.util.Arrays; 13 | import java.util.List; 14 | 15 | public class JavaSparkSQLByReflection { 16 | public static void main(String[] args) throws Exception { 17 | SparkConf sparkConf = new SparkConf().setAppName("JavaSparkSQLByReflection"); 18 | JavaSparkContext ctx = new JavaSparkContext(sparkConf); 19 | SQLContext sqlCtx = new SQLContext(ctx); 20 | 21 | System.out.println("=== Data source: RDD ==="); 22 | // Load a text file and convert each line to a Java Bean. 23 | JavaRDD people = ctx.textFile("people.txt").map( 24 | new Function() { 25 | @Override 26 | public People call(String line) { 27 | String[] parts = line.split(","); 28 | 29 | People people = new People(); 30 | people.setName(parts[0]); 31 | people.setAge(Integer.parseInt(parts[1].trim())); 32 | return people; 33 | } 34 | }); 35 | 36 | // Apply a schema to an RDD of Java Beans and register it as a table. 37 | DataFrame schemaPeople = sqlCtx.createDataFrame(people, People.class); 38 | schemaPeople.registerTempTable("people"); 39 | 40 | // SQL can be run over RDDs that have been registered as tables. 41 | DataFrame teenagers = sqlCtx.sql("SELECT name FROM people WHERE age >= 13 AND age <= 19"); 42 | 43 | // The results of SQL queries are DataFrames and support all the normal RDD operations. 44 | // The columns of a row in the result can be accessed by ordinal. 45 | List teenagerNames = teenagers.toJavaRDD().map(new Function() { 46 | @Override 47 | public String call(Row row) { 48 | return "Name: " + row.getString(0); 49 | } 50 | }).collect(); 51 | 52 | for (String name : teenagerNames) { 53 | System.out.println(name); 54 | } 55 | 56 | 57 | System.out.println("=== Data source: Parquet File ==="); 58 | // DataFrames can be saved as parquet files, maintaining the schema information. 59 | schemaPeople.saveAsParquetFile("people.parquet"); 60 | 61 | // Read in the parquet file created above. 62 | // Parquet files are self-describing so the schema is preserved. 63 | // The result of loading a parquet file is also a DataFrame. 64 | DataFrame parquetFile = sqlCtx.parquetFile("people.parquet"); 65 | 66 | //Parquet files can also be registered as tables and then used in SQL statements. 67 | parquetFile.registerTempTable("parquetFile"); 68 | DataFrame teenagers2 = 69 | sqlCtx.sql("SELECT name FROM parquetFile WHERE age >= 13 AND age <= 19"); 70 | teenagerNames = teenagers2.toJavaRDD().map(new Function() { 71 | @Override 72 | public String call(Row row) { 73 | return "Name: " + row.getString(0); 74 | } 75 | }).collect(); 76 | for (String name : teenagerNames) { 77 | System.out.println(name); 78 | } 79 | 80 | System.out.println("=== Data source: JSON Dataset ==="); 81 | // A JSON dataset is pointed by path. 82 | // The path can be either a single text file or a directory storing text files. 83 | String path = "people.json"; 84 | // Create a DataFrame from the file(s) pointed by path 85 | DataFrame peopleFromJsonFile = sqlCtx.jsonFile(path); 86 | 87 | // Because the schema of a JSON dataset is automatically inferred, to write queries, 88 | // it is better to take a look at what is the schema. 89 | peopleFromJsonFile.printSchema(); 90 | // The schema of people is ... 91 | // root 92 | // |-- age: IntegerType 93 | // |-- name: StringType 94 | 95 | // Register this DataFrame as a table. 96 | peopleFromJsonFile.registerTempTable("people"); 97 | 98 | // SQL statements can be run by using the sql methods provided by sqlCtx. 99 | DataFrame teenagers3 = sqlCtx.sql("SELECT name FROM people WHERE age >= 13 AND age <= 19"); 100 | 101 | // The results of SQL queries are DataFrame and support all the normal RDD operations. 102 | // The columns of a row in the result can be accessed by ordinal. 103 | teenagerNames = teenagers3.toJavaRDD().map(new Function() { 104 | @Override 105 | public String call(Row row) { 106 | return "Name: " + row.getString(0); 107 | } 108 | }).collect(); 109 | for (String name : teenagerNames) { 110 | System.out.println(name); 111 | } 112 | 113 | // Alternatively, a DataFrame can be created for a JSON dataset represented by 114 | // a RDD[String] storing one JSON object per string. 115 | List jsonData = Arrays.asList( 116 | "{\"name\":\"Yin\",\"address\":{\"city\":\"Columbus\",\"state\":\"Ohio\"}}"); 117 | JavaRDD anotherPeopleRDD = ctx.parallelize(jsonData); 118 | DataFrame peopleFromJsonRDD = sqlCtx.jsonRDD(anotherPeopleRDD.rdd()); 119 | 120 | 121 | // Take a look at the schema of this new DataFrame. 122 | peopleFromJsonRDD.printSchema(); 123 | // The schema of anotherPeople is ... 124 | // root 125 | // |-- address: StructType 126 | // | |-- city: StringType 127 | // | |-- state: StringType 128 | // |-- name: StringType 129 | 130 | peopleFromJsonRDD.registerTempTable("people2"); 131 | 132 | DataFrame peopleWithCity = sqlCtx.sql("SELECT name, address.city FROM people2"); 133 | List nameAndCity = peopleWithCity.toJavaRDD().map(new Function() { 134 | @Override 135 | public String call(Row row) { 136 | return "Name: " + row.getString(0) + ", City: " + row.getString(1); 137 | } 138 | }).collect(); 139 | for (String name : nameAndCity) { 140 | System.out.println(name); 141 | } 142 | 143 | ctx.stop(); 144 | } 145 | 146 | public static class People implements Serializable { 147 | private String name; 148 | private int age; 149 | 150 | public String getName() { 151 | return name; 152 | } 153 | 154 | public void setName(String name) { 155 | this.name = name; 156 | } 157 | 158 | public int getAge() { 159 | return age; 160 | } 161 | 162 | public void setAge(int age) { 163 | this.age = age; 164 | } 165 | } 166 | } -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/mllib/ScalaMovieLensALS.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.mllib 2 | 3 | import java.util.Random 4 | 5 | import org.apache.spark.mllib.recommendation.{ALS, MatrixFactorizationModel, Rating} 6 | import org.apache.spark.rdd._ 7 | import org.apache.spark.{SparkConf, SparkContext} 8 | 9 | /** 10 | * see:https://github.com/mohit-shrma/RandomSamples/blob/d9f1117bc21bb09d9fa858bc6d95e08e753e6fa0/SparkScala/CollabFilter/src/main/scala/MovieLensALS.scala 11 | */ 12 | object ScalaMovieLensALS { 13 | 14 | def main(args: Array[String]) { 15 | 16 | //import org.apache.log4j.{Logger,Level} 17 | //Logger.getLogger("org.apache.spark").setLevel(Level.WARN) 18 | //Logger.getLogger("org.eclipse.jetty.server").setLevel(Level.OFF) 19 | 20 | if (args.length != 2) { 21 | println("Usage: /path/to/spark/bin/spark-submit --driver-memory 2g --class com.javachen.spark.examples.mllib.ScalaMovieLensALS " + 22 | "target/scala-*/movielens-als-ssembly-*.jar movieLensHomeDir personalRatingsFile") 23 | sys.exit(1) 24 | } 25 | 26 | // set up environment 27 | val conf = new SparkConf().setAppName("ScalaMovieLensALS") 28 | val sc = new SparkContext(conf) 29 | 30 | // load ratings and movie titles 31 | val ratings = sc.textFile(args(0) + "/ratings.dat").map { line => 32 | val fields = line.split("::") 33 | // format: (timestamp % 10, Rating(userId, movieId, rating)) 34 | (fields(3).toLong % 10, Rating(fields(0).toInt, fields(1).toInt, fields(2).toDouble)) 35 | } 36 | 37 | val movies = sc.textFile(args(0) + "/movies.dat").map { line => 38 | val fields = line.split("::") 39 | // format: (movieId, movieName) 40 | (fields(0).toInt, fields(1)) 41 | }.collect().toMap 42 | 43 | val numRatings = ratings.count() 44 | val numUsers = ratings.map(_._2.user).distinct().count() 45 | val numMovies = ratings.map(_._2.product).distinct().count() 46 | 47 | println(s"Got $numRatings ratings from $numUsers users on $numMovies movies.") 48 | 49 | //get ratings of user on top 50 popular movies 50 | val mostRatedMovieIds = ratings.map(_._2.product) //extract movieId 51 | .countByValue //count ratings per movie 52 | .toSeq //convert map to seq 53 | .sortBy(-_._2) //sort by rating count in decreasing order 54 | .take(50) //take 50 most rated 55 | .map(_._1) //get movie ids 56 | 57 | val random = new Random(0) 58 | val selectedMovies = mostRatedMovieIds.filter(x => random.nextDouble() < 0.2) 59 | .map(x => (x, movies(x))) 60 | .toSeq 61 | val myRatings = elicitateRatings(selectedMovies) 62 | //convert received ratings to RDD[Rating], now this can be worked in parallel 63 | val myRatingsRDD = sc.parallelize(myRatings) 64 | 65 | // split ratings into train (60%), validation (20%), and test (20%) based on the 66 | // last digit of the timestamp, add myRatings to train, and cache them 67 | 68 | val numPartitions = 4 69 | val training = ratings.filter(x => x._1 < 6).values.union(myRatingsRDD).repartition(numPartitions).cache() 70 | val validation = ratings.filter(x => x._1 >= 6 && x._1 < 8).values.repartition(numPartitions).cache() 71 | val test = ratings.filter(x => x._1 >= 8).values.cache() 72 | 73 | val numTraining = training.count() 74 | val numValidation = validation.count() 75 | val numTest = test.count() 76 | 77 | println(s"Training: $numTraining, validation: $numValidation, test: $numTest") 78 | 79 | // train models and evaluate them on the validation set 80 | val ranks = List(8, 10, 12) 81 | val lambdas = List(0.1, 1.0, 10.0) 82 | val numIterations = List(10, 20) 83 | var bestModel: Option[MatrixFactorizationModel] = None 84 | var bestValidationRmse = Double.MaxValue 85 | var bestRank = 0 86 | var bestLambda = -1.0 87 | var bestNumIter = -1 88 | for (rank <- ranks; lambda <- lambdas; numIter <- numIterations) { 89 | //learn model for these parameter 90 | val model = ALS.train(training, rank, numIter, lambda) 91 | val validationRmse = computeRmse(model, validation) 92 | println(s"RMSE (validation) = $validationRmse for the model trained with rank = $rank , lambda = $lambda , and numIter = $numIter .") 93 | if (validationRmse < bestValidationRmse) { 94 | bestModel = Some(model) 95 | bestValidationRmse = validationRmse 96 | bestRank = rank 97 | bestLambda = lambda 98 | bestNumIter = numIter 99 | } 100 | } 101 | 102 | // evaluate the best model on the test set 103 | val testRmse = computeRmse(bestModel.get, test) 104 | println(s"The best model was trained with rank = $bestRank and lambda = $bestLambda , and numIter = $bestNumIter , and its RMSE on the test set is $testRmse .") 105 | 106 | //find best movies for the user 107 | val myRatedMovieIds = myRatings.map(_.product).toSet 108 | //generate candidates after taking out already rated movies 109 | val candidates = sc.parallelize(movies.keys.filter(!myRatedMovieIds.contains(_)).toSeq) 110 | val recommendations = bestModel.get.predict(candidates.map((0, _))).collect.sortBy(-_.rating).take(50) 111 | var i = 1 112 | println("Movies recommendation for you: ") 113 | recommendations.foreach { r => 114 | println("%2d".format(i) + ": " + movies(r.product)) 115 | i += 1 116 | } 117 | 118 | // create a naive baseline and compare it with the best model 119 | val meanRating = training.union(validation).map(_.rating).mean 120 | val baselineRmse = math.sqrt(test.map(x => (meanRating - x.rating) * (meanRating - x.rating)).mean) 121 | val improvement = (baselineRmse - testRmse) / baselineRmse * 100 122 | println("The best model improves the baseline by " + "%1.2f".format(improvement) + "%.") 123 | 124 | // clean up 125 | sc.stop() 126 | } 127 | 128 | /** Compute RMSE (Root Mean Squared Error). */ 129 | def computeRmse(model: MatrixFactorizationModel, data: RDD[Rating]) = { 130 | val usersProducts = data.map { case Rating(user, product, rate) => 131 | (user, product) 132 | } 133 | 134 | val predictions = model.predict(usersProducts).map { case Rating(user, product, rate) => 135 | ((user, product), rate) 136 | } 137 | 138 | val ratesAndPreds = data.map { case Rating(user, product, rate) => 139 | ((user, product), rate) 140 | }.join(predictions).sortByKey() 141 | 142 | math.sqrt(ratesAndPreds.map { case ((user, product), (r1, r2)) => 143 | val err = (r1 - r2) 144 | err * err 145 | }.mean()) 146 | } 147 | 148 | /** Elicitate ratings from commandline **/ 149 | def elicitateRatings(movies: Seq[(Int, String)]) = { 150 | val prompt = "Please rate following movie (1-5(best), or 0 if not seen):" 151 | println(prompt) 152 | val ratings = movies.flatMap { x => 153 | 154 | var rating: Option[Rating] = None 155 | var valid = false 156 | 157 | while (!valid) { 158 | print(x._2 + ": ") 159 | try { 160 | val r = Console.readInt 161 | if (r < 0 || r > 5) { 162 | println(prompt) 163 | } else { 164 | valid = true 165 | if (r > 0) { 166 | rating = Some(Rating(0, x._1, r)) 167 | } 168 | } 169 | } catch { 170 | case e: Exception => println(prompt) 171 | } 172 | } 173 | 174 | rating match { 175 | case Some(r) => Iterator(r) 176 | case None => Iterator.empty 177 | } 178 | 179 | } //end flatMap 180 | 181 | if (ratings.isEmpty) { 182 | error("No rating provided") 183 | } else { 184 | ratings 185 | } 186 | 187 | } 188 | 189 | } 190 | -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/mllib/MovieSimilarities.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.spark.examples.mllib 2 | 3 | import org.apache.spark.SparkContext 4 | 5 | /** 6 | * https://gist.github.com/MLnick/5286475 7 | * A port of [[http://blog.echen.me/2012/02/09/movie-recommendations-and-more-via-mapreduce-and-scalding/]] 8 | * to Spark. 9 | * Uses movie ratings data from MovieLens 100k dataset found at [[http://www.grouplens.org/node/73]] 10 | */ 11 | object MovieSimilarities { 12 | 13 | def main(args: Array[String]) { 14 | /** 15 | * Parameters to regularize correlation. 16 | */ 17 | val PRIOR_COUNT = 10 18 | val PRIOR_CORRELATION = 0 19 | 20 | val TRAIN_FILENAME = "data/ml-100k/ua.base" 21 | val TEST_FIELNAME = "data/ml-100k/ua.test" 22 | val MOVIES_FILENAME = "data/ml-100k/u.item" 23 | 24 | /** 25 | * Spark programs require a SparkContext to be initialized 26 | */ 27 | val master = "local[*]" 28 | val sc = new SparkContext(master, "MovieSimilarities") 29 | 30 | // get movie names keyed on id 31 | val movies = sc.textFile(MOVIES_FILENAME) 32 | .map(line => { 33 | val fields = line.split("\\|") 34 | (fields(0).toInt, fields(1)) 35 | }) 36 | val movieNames = movies.collectAsMap() // for local use to map id <-> movie name for pretty-printing 37 | 38 | // extract (userid, movieid, rating) from ratings data 39 | val ratings = sc.textFile(TRAIN_FILENAME) 40 | .map(line => { 41 | val fields = line.split("\t") 42 | (fields(0).toInt, fields(1).toInt, fields(2).toInt) 43 | }) 44 | 45 | // get num raters per movie, keyed on movie id 46 | val numRatersPerMovie = ratings 47 | .groupBy(tup => tup._2) 48 | .map(grouped => (grouped._1, grouped._2.size)) 49 | 50 | // join ratings with num raters on movie id 51 | val ratingsWithSize = ratings 52 | .groupBy(tup => tup._2) 53 | .join(numRatersPerMovie) 54 | .flatMap(joined => { 55 | joined._2._1.map(f => (f._1, f._2, f._3, joined._2._2)) 56 | }) 57 | 58 | // ratingsWithSize now contains the following fields: (user, movie, rating, numRaters). 59 | 60 | // dummy copy of ratings for self join 61 | val ratings2 = ratingsWithSize.keyBy(tup => tup._1) 62 | 63 | // join on userid and filter movie pairs such that we don't double-count and exclude self-pairs 64 | val ratingPairs = 65 | ratingsWithSize 66 | .keyBy(tup => tup._1) 67 | .join(ratings2) 68 | .filter(f => f._2._1._2 < f._2._2._2) 69 | 70 | // compute raw inputs to similarity metrics for each movie pair 71 | val vectorCalcs = 72 | ratingPairs 73 | .map(data => { 74 | val key = (data._2._1._2, data._2._2._2) 75 | val stats = 76 | (data._2._1._3 * data._2._2._3, // rating 1 * rating 2 77 | data._2._1._3, // rating movie 1 78 | data._2._2._3, // rating movie 2 79 | math.pow(data._2._1._3, 2), // square of rating movie 1 80 | math.pow(data._2._2._3, 2), // square of rating movie 2 81 | data._2._1._4, // number of raters movie 1 82 | data._2._2._4) // number of raters movie 2 83 | (key, stats) 84 | }) 85 | .groupByKey() 86 | .map(data => { 87 | val key = data._1 88 | val vals = data._2 89 | val size = vals.size 90 | val dotProduct = vals.map(f => f._1).sum 91 | val ratingSum = vals.map(f => f._2).sum 92 | val rating2Sum = vals.map(f => f._3).sum 93 | val ratingSq = vals.map(f => f._4).sum 94 | val rating2Sq = vals.map(f => f._5).sum 95 | val numRaters = vals.map(f => f._6).max 96 | val numRaters2 = vals.map(f => f._7).max 97 | (key, (size, dotProduct, ratingSum, rating2Sum, ratingSq, rating2Sq, numRaters, numRaters2)) 98 | }) 99 | 100 | // compute similarity metrics for each movie pair 101 | val similarities = 102 | vectorCalcs 103 | .map(fields => { 104 | val key = fields._1 105 | val (size, dotProduct, ratingSum, rating2Sum, ratingNormSq, rating2NormSq, numRaters, numRaters2) = fields._2 106 | val corr = correlation(size, dotProduct, ratingSum, rating2Sum, ratingNormSq, rating2NormSq) 107 | val regCorr = regularizedCorrelation(size, dotProduct, ratingSum, rating2Sum, 108 | ratingNormSq, rating2NormSq, PRIOR_COUNT, PRIOR_CORRELATION) 109 | val cosSim = cosineSimilarity(dotProduct, scala.math.sqrt(ratingNormSq), scala.math.sqrt(rating2NormSq)) 110 | val jaccard = jaccardSimilarity(size, numRaters, numRaters2) 111 | 112 | (key, (corr, regCorr, cosSim, jaccard)) 113 | }) 114 | 115 | // test a few movies out (substitute the contains call with the relevant movie name 116 | val sample = similarities.filter(m => { 117 | val movies = m._1 118 | (movieNames(movies._1).contains("Star Wars (1977)")) 119 | }) 120 | 121 | // collect results, excluding NaNs if applicable 122 | val result = sample.map(v => { 123 | val m1 = v._1._1 124 | val m2 = v._1._2 125 | val corr = v._2._1 126 | val rcorr = v._2._2 127 | val cos = v._2._3 128 | val j = v._2._4 129 | (movieNames(m1), movieNames(m2), corr, rcorr, cos, j) 130 | }).collect().filter(e => !(e._4 equals Double.NaN)) // test for NaNs must use equals rather than == 131 | .sortBy(elem => elem._4).take(10) 132 | 133 | // print the top 10 out 134 | result.foreach(r => println(r._1 + " | " + r._2 + " | " + r._3.formatted("%2.4f") + " | " + r._4.formatted("%2.4f") 135 | + " | " + r._5.formatted("%2.4f") + " | " + r._6.formatted("%2.4f"))) 136 | } 137 | 138 | // ************************* 139 | // * SIMILARITY MEASURES 140 | // ************************* 141 | 142 | /** 143 | * The correlation between two vectors A, B is 144 | * cov(A, B) / (stdDev(A) * stdDev(B)) 145 | * 146 | * This is equivalent to 147 | * [n * dotProduct(A, B) - sum(A) * sum(B)] / 148 | * sqrt{ [n * norm(A)^2 - sum(A)^2] [n * norm(B)^2 - sum(B)^2] } 149 | */ 150 | def correlation(size: Double, dotProduct: Double, ratingSum: Double, 151 | rating2Sum: Double, ratingNormSq: Double, rating2NormSq: Double) = { 152 | 153 | val numerator = size * dotProduct - ratingSum * rating2Sum 154 | val denominator = scala.math.sqrt(size * ratingNormSq - ratingSum * ratingSum) * 155 | scala.math.sqrt(size * rating2NormSq - rating2Sum * rating2Sum) 156 | 157 | numerator / denominator 158 | } 159 | 160 | /** 161 | * Regularize correlation by adding virtual pseudocounts over a prior: 162 | * RegularizedCorrelation = w * ActualCorrelation + (1 - w) * PriorCorrelation 163 | * where w = # actualPairs / (# actualPairs + # virtualPairs). 164 | */ 165 | def regularizedCorrelation(size: Double, dotProduct: Double, ratingSum: Double, 166 | rating2Sum: Double, ratingNormSq: Double, rating2NormSq: Double, 167 | virtualCount: Double, priorCorrelation: Double) = { 168 | 169 | val unregularizedCorrelation = correlation(size, dotProduct, ratingSum, rating2Sum, ratingNormSq, rating2NormSq) 170 | val w = size / (size + virtualCount) 171 | 172 | w * unregularizedCorrelation + (1 - w) * priorCorrelation 173 | } 174 | 175 | /** 176 | * The cosine similarity between two vectors A, B is 177 | * dotProduct(A, B) / (norm(A) * norm(B)) 178 | */ 179 | def cosineSimilarity(dotProduct: Double, ratingNorm: Double, rating2Norm: Double) = { 180 | dotProduct / (ratingNorm * rating2Norm) 181 | } 182 | 183 | /** 184 | * The Jaccard Similarity between two sets A, B is 185 | * |Intersection(A, B)| / |Union(A, B)| 186 | */ 187 | def jaccardSimilarity(usersInCommon: Double, totalUsers1: Double, totalUsers2: Double) = { 188 | val union = totalUsers1 + totalUsers2 - usersInCommon 189 | usersInCommon / union 190 | } 191 | 192 | } -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/mllib/MovieLensALS.scala: -------------------------------------------------------------------------------- 1 | /* 2 | * Licensed to the Apache Software Foundation (ASF) under one or more 3 | * contributor license agreements. See the NOTICE file distributed with 4 | * this work for additional information regarding copyright ownership. 5 | * The ASF licenses this file to You under the Apache License, Version 2.0 6 | * (the "License"); you may not use this file except in compliance with 7 | * the License. You may obtain a copy of the License at 8 | * 9 | * http://www.apache.org/licenses/LICENSE-2.0 10 | * 11 | * Unless required by applicable law or agreed to in writing, software 12 | * distributed under the License is distributed on an "AS IS" BASIS, 13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 | * See the License for the specific language governing permissions and 15 | * limitations under the License. 16 | */ 17 | 18 | package com.javachen.spark.examples.mllib 19 | 20 | import org.apache.spark.mllib.recommendation.{ALS, MatrixFactorizationModel, Rating} 21 | import org.apache.spark.rdd.RDD 22 | import org.apache.spark.{SparkConf, SparkContext} 23 | import scopt.OptionParser 24 | 25 | import scala.collection.mutable 26 | 27 | /** 28 | * An example app for ALS on MovieLens data (http://grouplens.org/datasets/movielens/). 29 | * Run with 30 | * {{{ 31 | * bin/run-example org.apache.spark.examples.mllib.MovieLensALS 32 | * }}} 33 | * A synthetic dataset in MovieLens format can be found at `data/mllib/sample_movielens_data.txt`. 34 | * If you use it as a template to create your own app, please use `spark-submit` to submit your app. 35 | */ 36 | object MovieLensALS { 37 | 38 | case class Params( 39 | input: String = "data/ml-1m/ratings.dat", 40 | userDataInput: String = "data/ml-1m/personalRatings.txt", 41 | kryo: Boolean = false, 42 | numIterations: Int = 20, 43 | lambda: Double = 1.0, 44 | rank: Int = 10, 45 | numUserBlocks: Int = -1, 46 | numProductBlocks: Int = -1, 47 | implicitPrefs: Boolean = false) 48 | 49 | def main(args: Array[String]) { 50 | val defaultParams = Params() 51 | 52 | val parser = new OptionParser[Params]("MovieLensALS") { 53 | head("MovieLensALS: an example app for ALS on MovieLens data.") 54 | opt[Int]("rank") 55 | .text(s"rank, default: ${defaultParams.rank}}") 56 | .action((x, c) => c.copy(rank = x)) 57 | opt[Int]("numIterations") 58 | .text(s"number of iterations, default: ${defaultParams.numIterations}") 59 | .action((x, c) => c.copy(numIterations = x)) 60 | opt[Double]("lambda") 61 | .text(s"lambda (smoothing constant), default: ${defaultParams.lambda}") 62 | .action((x, c) => c.copy(lambda = x)) 63 | opt[Unit]("kryo") 64 | .text("use Kryo serialization") 65 | .action((_, c) => c.copy(kryo = true)) 66 | opt[Int]("numUserBlocks") 67 | .text(s"number of user blocks, default: ${defaultParams.numUserBlocks} (auto)") 68 | .action((x, c) => c.copy(numUserBlocks = x)) 69 | opt[Int]("numProductBlocks") 70 | .text(s"number of product blocks, default: ${defaultParams.numProductBlocks} (auto)") 71 | .action((x, c) => c.copy(numProductBlocks = x)) 72 | opt[Unit]("implicitPrefs") 73 | .text("use implicit preference") 74 | .action((_, c) => c.copy(implicitPrefs = true)) 75 | opt[String]("userDataInput") 76 | .required() 77 | .text("input paths to user dataset") 78 | .action((x, c) => c.copy(userDataInput = x)) 79 | arg[String]("") 80 | .required() 81 | .text("input paths to a MovieLens dataset of ratings") 82 | .action((x, c) => c.copy(input = x)) 83 | note( 84 | """ 85 | |For example, the following command runs this app on a synthetic dataset: 86 | | 87 | | bin/spark-submit --class com.javachen.grab.examples.mllib.MovieLensALS \ 88 | | examples/target/scala-*/grab-examples-*.jar \ 89 | | --rank 5 --numIterations 20 --lambda 1.0 \ 90 | | --userDataInput data/ml-1m/personalRatings.txt \ 91 | | data/ml-1m/ratings.dat 92 | """.stripMargin) 93 | } 94 | 95 | parser.parse(args, defaultParams).map { params => 96 | run(params) 97 | } getOrElse { 98 | System.exit(1) 99 | } 100 | } 101 | 102 | def run(params: Params) { 103 | val conf = new SparkConf().setAppName(s"MovieLensALS with $params").set("spark.executor.memory", "2g") 104 | if (params.kryo) { 105 | conf.registerKryoClasses(Array(classOf[mutable.BitSet], classOf[Rating])) 106 | .set("spark.kryoserializer.buffer.mb", "8") 107 | } 108 | val sc = new SparkContext(conf) 109 | 110 | 111 | val ratings = sc.textFile(params.input).map { line => 112 | val fields = line.split("::") 113 | /* 114 | * MovieLens ratings are on a scale of 1-5: 115 | * 5: Must see 116 | * 4: Will enjoy 117 | * 3: It's okay 118 | * 2: Fairly bad 119 | * 1: Awful 120 | * So we should not recommend a movie if the predicted rating is less than 3. 121 | * To map ratings to confidence scores, we use 122 | * 5 -> 2.5, 4 -> 1.5, 3 -> 0.5, 2 -> -0.5, 1 -> -1.5. This mappings means unobserved 123 | * entries are generally between It's okay and Fairly bad. 124 | * The semantics of 0 in this expanded world of non-positive weights 125 | * are "the same as never having interacted at all". 126 | */ 127 | if (params.implicitPrefs) { 128 | // format: (timestamp % 10, Rating(userId, movieId, rating)) 129 | (fields(3).toLong % 10, Rating(fields(0).toInt, fields(1).toInt, fields(2).toDouble - 2.5)) 130 | } else { 131 | // format: (timestamp % 10, Rating(userId, movieId, rating)) 132 | (fields(3).toLong % 10, Rating(fields(0).toInt, fields(1).toInt, fields(2).toDouble)) 133 | } 134 | }.cache() 135 | 136 | 137 | val numRatings = ratings.count() 138 | val numUsers = ratings.map(_._2.user).distinct().count() 139 | val numMovies = ratings.map(_._2.product).distinct().count() 140 | 141 | println("Got " + numRatings + " ratings from " + numUsers + " users on " + numMovies + " movies.") 142 | 143 | val model = evaluateMode(params, ratings) 144 | 145 | predictMoive(params, sc, model) 146 | 147 | // clean up 148 | sc.stop() 149 | } 150 | 151 | def predictMoive(params: Params, sc: SparkContext, model: MatrixFactorizationModel): Unit = { 152 | //为用户1推荐10个 153 | var rs = model.recommendProducts(1, 10) 154 | var value = "" 155 | var key = 0 156 | 157 | //保存推荐数据到hbase中 158 | rs.foreach(r => { 159 | key = r.user 160 | value = value + r.product + ":" + r.rating + "," 161 | }) 162 | 163 | println(value) 164 | 165 | } 166 | 167 | def evaluateMode(params: Params, ratings: RDD[(Long, Rating)]): MatrixFactorizationModel = { 168 | val training = ratings.values.repartition(4) 169 | 170 | //建立模型 171 | val start = System.currentTimeMillis() 172 | val model = new ALS().setRank(params.rank).setIterations(params.numIterations).setLambda(params.lambda).setImplicitPrefs(params.implicitPrefs).setUserBlocks(params.numUserBlocks).setProductBlocks(params.numProductBlocks).run(training) 173 | println("Train Time = " + (System.currentTimeMillis() - start) * 1.0 / 1000) 174 | val testRmse = computeRmse(model, training) 175 | 176 | println("RMSE = " + testRmse) 177 | 178 | model 179 | } 180 | 181 | /** Compute RMSE (Root Mean Squared Error). */ 182 | def computeRmse(model: MatrixFactorizationModel, data: RDD[Rating]) = { 183 | val usersProducts = data.map { case Rating(user, product, rate) => 184 | (user, product) 185 | } 186 | 187 | val predictions = model.predict(usersProducts).map { case Rating(user, product, rate) => 188 | ((user, product), rate) 189 | } 190 | 191 | val ratesAndPreds = data.map { case Rating(user, product, rate) => 192 | ((user, product), rate) 193 | }.join(predictions).sortByKey() 194 | 195 | math.sqrt(ratesAndPreds.map { case ((user, product), (r1, r2)) => 196 | val err = (r1 - r2) 197 | err * err 198 | }.mean()) 199 | } 200 | } 201 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | Apache License 2 | Version 2.0, January 2004 3 | http://www.apache.org/licenses/ 4 | 5 | TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 6 | 7 | 1. Definitions. 8 | 9 | "License" shall mean the terms and conditions for use, reproduction, 10 | and distribution as defined by Sections 1 through 9 of this document. 11 | 12 | "Licensor" shall mean the copyright owner or entity authorized by 13 | the copyright owner that is granting the License. 14 | 15 | "Legal Entity" shall mean the union of the acting entity and all 16 | other entities that control, are controlled by, or are under common 17 | control with that entity. For the purposes of this definition, 18 | "control" means (i) the power, direct or indirect, to cause the 19 | direction or management of such entity, whether by contract or 20 | otherwise, or (ii) ownership of fifty percent (50%) or more of the 21 | outstanding shares, or (iii) beneficial ownership of such entity. 22 | 23 | "You" (or "Your") shall mean an individual or Legal Entity 24 | exercising permissions granted by this License. 25 | 26 | "Source" form shall mean the preferred form for making modifications, 27 | including but not limited to software source code, documentation 28 | source, and configuration files. 29 | 30 | "Object" form shall mean any form resulting from mechanical 31 | transformation or translation of a Source form, including but 32 | not limited to compiled object code, generated documentation, 33 | and conversions to other media types. 34 | 35 | "Work" shall mean the work of authorship, whether in Source or 36 | Object form, made available under the License, as indicated by a 37 | copyright notice that is included in or attached to the work 38 | (an example is provided in the Appendix below). 39 | 40 | "Derivative Works" shall mean any work, whether in Source or Object 41 | form, that is based on (or derived from) the Work and for which the 42 | editorial revisions, annotations, elaborations, or other modifications 43 | represent, as a whole, an original work of authorship. For the purposes 44 | of this License, Derivative Works shall not include works that remain 45 | separable from, or merely link (or bind by name) to the interfaces of, 46 | the Work and Derivative Works thereof. 47 | 48 | "Contribution" shall mean any work of authorship, including 49 | the original version of the Work and any modifications or additions 50 | to that Work or Derivative Works thereof, that is intentionally 51 | submitted to Licensor for inclusion in the Work by the copyright owner 52 | or by an individual or Legal Entity authorized to submit on behalf of 53 | the copyright owner. For the purposes of this definition, "submitted" 54 | means any form of electronic, verbal, or written communication sent 55 | to the Licensor or its representatives, including but not limited to 56 | communication on electronic mailing lists, source code control systems, 57 | and issue tracking systems that are managed by, or on behalf of, the 58 | Licensor for the purpose of discussing and improving the Work, but 59 | excluding communication that is conspicuously marked or otherwise 60 | designated in writing by the copyright owner as "Not a Contribution." 61 | 62 | "Contributor" shall mean Licensor and any individual or Legal Entity 63 | on behalf of whom a Contribution has been received by Licensor and 64 | subsequently incorporated within the Work. 65 | 66 | 2. Grant of Copyright License. Subject to the terms and conditions of 67 | this License, each Contributor hereby grants to You a perpetual, 68 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 69 | copyright license to reproduce, prepare Derivative Works of, 70 | publicly display, publicly perform, sublicense, and distribute the 71 | Work and such Derivative Works in Source or Object form. 72 | 73 | 3. Grant of Patent License. Subject to the terms and conditions of 74 | this License, each Contributor hereby grants to You a perpetual, 75 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 76 | (except as stated in this section) patent license to make, have made, 77 | use, offer to sell, sell, import, and otherwise transfer the Work, 78 | where such license applies only to those patent claims licensable 79 | by such Contributor that are necessarily infringed by their 80 | Contribution(s) alone or by combination of their Contribution(s) 81 | with the Work to which such Contribution(s) was submitted. If You 82 | institute patent litigation against any entity (including a 83 | cross-claim or counterclaim in a lawsuit) alleging that the Work 84 | or a Contribution incorporated within the Work constitutes direct 85 | or contributory patent infringement, then any patent licenses 86 | granted to You under this License for that Work shall terminate 87 | as of the date such litigation is filed. 88 | 89 | 4. Redistribution. You may reproduce and distribute copies of the 90 | Work or Derivative Works thereof in any medium, with or without 91 | modifications, and in Source or Object form, provided that You 92 | meet the following conditions: 93 | 94 | (a) You must give any other recipients of the Work or 95 | Derivative Works a copy of this License; and 96 | 97 | (b) You must cause any modified files to carry prominent notices 98 | stating that You changed the files; and 99 | 100 | (c) You must retain, in the Source form of any Derivative Works 101 | that You distribute, all copyright, patent, trademark, and 102 | attribution notices from the Source form of the Work, 103 | excluding those notices that do not pertain to any part of 104 | the Derivative Works; and 105 | 106 | (d) If the Work includes a "NOTICE" text file as part of its 107 | distribution, then any Derivative Works that You distribute must 108 | include a readable copy of the attribution notices contained 109 | within such NOTICE file, excluding those notices that do not 110 | pertain to any part of the Derivative Works, in at least one 111 | of the following places: within a NOTICE text file distributed 112 | as part of the Derivative Works; within the Source form or 113 | documentation, if provided along with the Derivative Works; or, 114 | within a display generated by the Derivative Works, if and 115 | wherever such third-party notices normally appear. The contents 116 | of the NOTICE file are for informational purposes only and 117 | do not modify the License. You may add Your own attribution 118 | notices within Derivative Works that You distribute, alongside 119 | or as an addendum to the NOTICE text from the Work, provided 120 | that such additional attribution notices cannot be construed 121 | as modifying the License. 122 | 123 | You may add Your own copyright statement to Your modifications and 124 | may provide additional or different license terms and conditions 125 | for use, reproduction, or distribution of Your modifications, or 126 | for any such Derivative Works as a whole, provided Your use, 127 | reproduction, and distribution of the Work otherwise complies with 128 | the conditions stated in this License. 129 | 130 | 5. Submission of Contributions. Unless You explicitly state otherwise, 131 | any Contribution intentionally submitted for inclusion in the Work 132 | by You to the Licensor shall be under the terms and conditions of 133 | this License, without any additional terms or conditions. 134 | Notwithstanding the above, nothing herein shall supersede or modify 135 | the terms of any separate license agreement you may have executed 136 | with Licensor regarding such Contributions. 137 | 138 | 6. Trademarks. This License does not grant permission to use the trade 139 | names, trademarks, service marks, or product names of the Licensor, 140 | except as required for reasonable and customary use in describing the 141 | origin of the Work and reproducing the content of the NOTICE file. 142 | 143 | 7. Disclaimer of Warranty. Unless required by applicable law or 144 | agreed to in writing, Licensor provides the Work (and each 145 | Contributor provides its Contributions) on an "AS IS" BASIS, 146 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or 147 | implied, including, without limitation, any warranties or conditions 148 | of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A 149 | PARTICULAR PURPOSE. You are solely responsible for determining the 150 | appropriateness of using or redistributing the Work and assume any 151 | risks associated with Your exercise of permissions under this License. 152 | 153 | 8. Limitation of Liability. In no event and under no legal theory, 154 | whether in tort (including negligence), contract, or otherwise, 155 | unless required by applicable law (such as deliberate and grossly 156 | negligent acts) or agreed to in writing, shall any Contributor be 157 | liable to You for damages, including any direct, indirect, special, 158 | incidental, or consequential damages of any character arising as a 159 | result of this License or out of the use or inability to use the 160 | Work (including but not limited to damages for loss of goodwill, 161 | work stoppage, computer failure or malfunction, or any and all 162 | other commercial damages or losses), even if such Contributor 163 | has been advised of the possibility of such damages. 164 | 165 | 9. Accepting Warranty or Additional Liability. While redistributing 166 | the Work or Derivative Works thereof, You may choose to offer, 167 | and charge a fee for, acceptance of support, warranty, indemnity, 168 | or other liability obligations and/or rights consistent with this 169 | License. However, in accepting such obligations, You may act only 170 | on Your own behalf and on Your sole responsibility, not on behalf 171 | of any other Contributor, and only if You agree to indemnify, 172 | defend, and hold each Contributor harmless for any liability 173 | incurred by, or claims asserted against, such Contributor by reason 174 | of your accepting any such warranty or additional liability. 175 | 176 | END OF TERMS AND CONDITIONS 177 | 178 | APPENDIX: How to apply the Apache License to your work. 179 | 180 | To apply the Apache License to your work, attach the following 181 | boilerplate notice, with the fields enclosed by brackets "{}" 182 | replaced with your own identifying information. (Don't include 183 | the brackets!) The text should be enclosed in the appropriate 184 | comment syntax for the file format. We also recommend that a 185 | file or class name and description of purpose be included on the 186 | same "printed page" as the copyright notice for easier 187 | identification within third-party archives. 188 | 189 | Copyright {yyyy} {name of copyright owner} 190 | 191 | Licensed under the Apache License, Version 2.0 (the "License"); 192 | you may not use this file except in compliance with the License. 193 | You may obtain a copy of the License at 194 | 195 | http://www.apache.org/licenses/LICENSE-2.0 196 | 197 | Unless required by applicable law or agreed to in writing, software 198 | distributed under the License is distributed on an "AS IS" BASIS, 199 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 200 | See the License for the specific language governing permissions and 201 | limitations under the License. 202 | 203 | -------------------------------------------------------------------------------- /src/main/scala/com/javachen/spark/examples/mllib/ScalaLocalALS.scala: -------------------------------------------------------------------------------- 1 | package com.javachen.grab 2 | 3 | import org.apache.spark.mllib.recommendation.{ALS, Rating} 4 | import org.apache.spark.rdd.RDD 5 | import org.apache.spark.{SparkConf, SparkContext} 6 | import org.jblas.DoubleMatrix 7 | import scala.sys.process._ 8 | 9 | import org.apache.log4j.{Level, Logger} 10 | 11 | 12 | /** 13 | * 本地模式运行 14 | */ 15 | object ScalaLocalALS { 16 | def main(args: Array[String]): Unit = { 17 | val sc = new SparkContext(new SparkConf().setAppName("Scala Collaborative Filtering Example")) 18 | 19 | Logger.getLogger("org.apache.spark").setLevel(Level.WARN) 20 | Logger.getLogger("org.apache.hadoop").setLevel(Level.WARN) 21 | Logger.getLogger("org.eclipse.jetty.server").setLevel(Level.OFF) 22 | 23 | // 1. 加载并解析数据 24 | val data = sc.textFile("data/ml-1m/ratings.dat") 25 | 26 | val ratings = data.map(_.split("::") match { case Array(user, item, rate, ts) => 27 | Rating(user.toInt, item.toInt, rate.toDouble) 28 | }).cache() 29 | 30 | val users = ratings.map(_.user).distinct() 31 | val products = ratings.map(_.product).distinct() 32 | println("Got "+ratings.count()+" ratings from "+users.count+" users on "+products.count+" products.") 33 | //Got 1000209 ratings from 6040 users on 3706 products. 34 | 35 | // 2. 训练模型 36 | val rank = 12 37 | val lambda = 0.01 38 | val numIterations = 20 39 | val model = ALS.train(ratings, rank, numIterations, lambda) 40 | 41 | model.userFeatures 42 | model.userFeatures.count 43 | 44 | model.productFeatures 45 | model.productFeatures.count 46 | 47 | // 3. 计算均方差 48 | //从 ratings 中获得只包含用户和商品的数据集 49 | val usersProducts = ratings.map { case Rating(user, product, rate) => 50 | (user, product) 51 | } 52 | 53 | usersProducts.count //Long = 1000209 54 | 55 | //使用推荐模型对用户商品进行预测评分，得到预测评分的数据集 56 | var predictions = model.predict(usersProducts).map { case Rating(user, product, rate) => 57 | ((user, product), rate) 58 | } 59 | 60 | predictions.count //Long = 1000209 61 | 62 | //将真实评分数据集与预测评分数据集进行合并 63 | val ratesAndPreds = ratings.map { case Rating(user, product, rate) => 64 | ((user, product), rate) 65 | }.join(predictions) 66 | 67 | ratesAndPreds.count //Long = 1000209 68 | 69 | val rmse= math.sqrt(ratesAndPreds.map { case ((user, product), (r1, r2)) => 70 | val err = (r1 - r2) 71 | err * err 72 | }.mean()) 73 | 74 | println(s"RMSE = $rmse") 75 | 76 | // 4.保存预测评分，确保只生成一个文件，并排序 77 | "rm -r /tmp/result".! 78 | ratesAndPreds.sortByKey().repartition(1).sortBy(_._1).map({ 79 | case ((user, product), (rate, pred)) => (user + "," + product + "," + rate + "," + pred) 80 | }).saveAsTextFile("/tmp/result") 81 | 82 | //对预测的评分结果按用户进行分组并按评分倒排序 83 | predictions.map { case ((user, product), rate) => 84 | (user, (product, rate)) 85 | }.groupByKey().map{case (user_id,list)=> 86 | (user_id,list.toList.sortBy {case (goods_id,rate)=> - rate}) 87 | } 88 | 89 | // 5. 对某一个用户推荐所有商品 90 | users.take(5) //Array[Int] = Array(384, 1084, 4904, 3702, 5618) 91 | val userId = users.take(1)(0) //384 92 | val K = 10 93 | val topKRecs = model.recommendProducts(userId, K) 94 | //topKRecs: Array[org.apache.spark.mllib.recommendation.Rating] = Array(Rating(384,1539,7.360670267591244), Rating(384,219,6.736019537477872), Rating(384,1520,6.730562698267339), Rating(384,775,6.697620546404394), Rating(384,3161,6.49555676613329), Rating(384,2711,6.445916831219404), Rating(384,2503,6.428273027496898), Rating(384,771,6.4255234943275825), Rating(384,853,6.170422982870869), Rating(384,759,6.04929517890501)) 95 | 96 | println(topKRecs.mkString("\n")) 97 | // Rating(384,1539,7.360670267591244) 98 | // Rating(384,219,6.736019537477872) 99 | // Rating(384,1520,6.730562698267339) 100 | // Rating(384,775,6.697620546404394) 101 | // Rating(384,3161,6.49555676613329) 102 | // Rating(384,2711,6.445916831219404) 103 | // Rating(384,2503,6.428273027496898) 104 | // Rating(384,771,6.4255234943275825) 105 | // Rating(384,853,6.170422982870869) 106 | // Rating(384,759,6.04929517890501) 107 | 108 | val productsForUser=ratings.keyBy(_.user).lookup(384) 109 | // Seq[org.apache.spark.mllib.recommendation.Rating] = WrappedArray(Rating(384,2055,2.0), Rating(384,1197,4.0), Rating(384,593,5.0), Rating(384,599,3.0), Rating(384,673,2.0), Rating(384,3037,4.0), Rating(384,1381,2.0), Rating(384,1610,4.0), Rating(384,3074,4.0), Rating(384,204,4.0), Rating(384,3508,3.0), Rating(384,1007,3.0), Rating(384,260,4.0), Rating(384,3487,3.0), Rating(384,3494,3.0), Rating(384,1201,5.0), Rating(384,3671,5.0), Rating(384,1207,4.0), Rating(384,2947,4.0), Rating(384,2951,4.0), Rating(384,2896,2.0), Rating(384,1304,5.0)) 110 | 111 | productsForUser.size //Int = 22 112 | productsForUser.sortBy(-_.rating).take(10).map(rating => (rating.product, rating.rating)).foreach(println) 113 | // (593,5.0) 114 | // (1201,5.0) 115 | // (3671,5.0) 116 | // (1304,5.0) 117 | // (1197,4.0) 118 | // (3037,4.0) 119 | // (1610,4.0) 120 | // (3074,4.0) 121 | // (204,4.0) 122 | // (260,4.0) 123 | 124 | /* Compute squared error between a predicted and actual rating */ 125 | // We'll take the first rating for our example user 789 126 | val actualRating = productsForUser.take(1)(0) 127 | //actualRating: org.apache.spark.mllib.recommendation.Rating = Rating(384,2055,2.0) val predictedRating = model.predict(789, actualRating.product) 128 | val predictedRating = model.predict(384, actualRating.product) 129 | //predictedRating: Double = 1.9426030777174637 130 | 131 | //找出和2055商品最相似的商品 132 | val itemId = 2055 133 | val itemFactor = model.productFeatures.lookup(itemId).head 134 | //itemFactor: Array[Double] = Array(0.3660752773284912, 0.43573060631752014, -0.3421429991722107, 0.44382765889167786, -1.4875195026397705, 0.6274569630622864, -0.3264533579349518, -0.9939845204353333, -0.8710321187973022, -0.7578890323638916, -0.14621856808662415, -0.7254264950752258) 135 | val itemVector = new DoubleMatrix(itemFactor) 136 | //itemVector: org.jblas.DoubleMatrix = [0.366075; 0.435731; -0.342143; 0.443828; -1.487520; 0.627457; -0.326453; -0.993985; -0.871032; -0.757889; -0.146219; -0.725426] 137 | 138 | cosineSimilarity(itemVector, itemVector) 139 | // res99: Double = 0.9999999999999999 140 | 141 | val sims = model.productFeatures.map{ case (id, factor) => 142 | val factorVector = new DoubleMatrix(factor) 143 | val sim = cosineSimilarity(factorVector, itemVector) 144 | (id, sim) 145 | } 146 | val sortedSims = sims.top(K)(Ordering.by[(Int, Double), Double] { case (id, similarity) => similarity }) 147 | //sortedSims: Array[(Int, Double)] = Array((2055,0.9999999999999999), (2051,0.9138311231145874), (3520,0.8739823400539756), (2190,0.8718466671129721), (2050,0.8612639515847019), (1011,0.8466911667526461), (2903,0.8455764332511272), (3121,0.8227325520485377), (3674,0.8075743004357392), (2016,0.8063817280259447)) 148 | println(sortedSims.mkString("\n")) 149 | // (2055,0.9999999999999999) 150 | // (2051,0.9138311231145874) 151 | // (3520,0.8739823400539756) 152 | // (2190,0.8718466671129721) 153 | // (2050,0.8612639515847019) 154 | // (1011,0.8466911667526461) 155 | // (2903,0.8455764332511272) 156 | // (3121,0.8227325520485377) 157 | // (3674,0.8075743004357392) 158 | // (2016,0.8063817280259447) 159 | 160 | val sortedSims2 = sims.top(K + 1)(Ordering.by[(Int, Double), Double] { case (id, similarity) => similarity }) 161 | //sortedSims2: Array[(Int, Double)] = Array((2055,0.9999999999999999), (2051,0.9138311231145874), (3520,0.8739823400539756), (2190,0.8718466671129721), (2050,0.8612639515847019), (1011,0.8466911667526461), (2903,0.8455764332511272), (3121,0.8227325520485377), (3674,0.8075743004357392), (2016,0.8063817280259447), (3672,0.8016276723120674)) 162 | 163 | sortedSims2.slice(1, 11).map{ case (id, sim) => (id, sim) }.mkString("\n") 164 | // (2051,0.9138311231145874) 165 | // (3520,0.8739823400539756) 166 | // (2190,0.8718466671129721) 167 | // (2050,0.8612639515847019) 168 | // (1011,0.8466911667526461) 169 | // (2903,0.8455764332511272) 170 | // (3121,0.8227325520485377) 171 | // (3674,0.8075743004357392) 172 | // (2016,0.8063817280259447) 173 | // (3672,0.8016276723120674) 174 | 175 | //计算给该用户推荐的前K个商品的平均准确度MAPK 176 | val actualProducts= productsForUser.map(_.product) 177 | //actualProducts: Seq[Int] = ArrayBuffer(2055, 1197, 593, 599, 673, 3037, 1381, 1610, 3074, 204, 3508, 1007, 260, 3487, 3494, 1201, 3671, 1207, 2947, 2951, 2896, 1304) 178 | val predictedProducts= topKRecs.map(_.product) 179 | //predictedProducts:Array[Int] = Array(1539, 219, 1520, 775, 3161, 2711, 2503, 771, 853, 759) 180 | val apk10 = avgPrecisionK(actualProducts, predictedProducts, 10) 181 | // apk10: Double = 0.0 182 | 183 | users.collect.flatMap { user => 184 | model.recommendProducts(user, 10) 185 | } 186 | 187 | //计算所有的推荐结果 188 | val itemFactors = model.productFeatures.map { case (prodcut, factor) => factor }.collect() 189 | val itemMatrix = new DoubleMatrix(itemFactors) 190 | println(itemMatrix.rows, itemMatrix.columns) 191 | 192 | val imBroadcast = sc.broadcast(itemMatrix) 193 | 194 | var idxProducts=model.productFeatures.map { case (prodcut, factor) => prodcut }.zipWithIndex().map{case (prodcut, idx) => (idx,prodcut)}.collectAsMap() 195 | val idxProductsBroadcast = sc.broadcast(idxProducts) 196 | 197 | val allRecs = model.userFeatures.map{ case (user, array) => 198 | val userVector = new DoubleMatrix(array) 199 | val scores = imBroadcast.value.mmul(userVector) 200 | val sortedWithId = scores.data.zipWithIndex.sortBy(-_._1) 201 | val recommendedProducts = sortedWithId.map(_._2).map{idx=>idxProductsBroadcast.value.get(idx).get} 202 | (user, recommendedProducts) //recommendedIds 为索引 203 | } 204 | 205 | //验证结果是否正确 206 | allRecs.lookup(384).head.take(10) 207 | //res50: Array[Int] = Array(1539, 219, 1520, 775, 3161, 2711, 2503, 771, 853, 759) 208 | topKRecs.map(_.product) 209 | //res49: Array[Int] = Array(1539, 219, 1520, 775, 3161, 2711, 2503, 771, 853, 759) 210 | 211 | 212 | //得到每个用户评分过的所有商品 213 | val userProducts = ratings.map{ case Rating(user, product, rating) => (user, product) }.groupBy(_._1) 214 | 215 | // finally, compute the APK for each user, and average them to find MAPK 216 | val MAPK = allRecs.join(userProducts).map{ case (userId, (predictedProducts, actualList)) => 217 | val actualProducts = actualList.map{case (user, product)=>product}.toSeq 218 | avgPrecisionK(actualProducts, predictedProducts, K) 219 | }.reduce(_ + _) / allRecs.count 220 | println("Mean Average Precision at K = " + MAPK) 221 | 222 | // MSE, RMSE and MAE 223 | import org.apache.spark.mllib.evaluation.RegressionMetrics 224 | 225 | val predictedAndTrue = ratesAndPreds.map { case ((user, product), (actual, predicted)) => (actual, predicted) } 226 | val regressionMetrics = new RegressionMetrics(predictedAndTrue) 227 | println("Mean Squared Error = " + regressionMetrics.meanSquaredError) 228 | println("Root Mean Squared Error = " + regressionMetrics.rootMeanSquaredError) 229 | // Mean Squared Error = 0.08231947642632852 230 | // Root Mean Squared Error = 0.2869137090247319 231 | 232 | // MAPK 233 | import org.apache.spark.mllib.evaluation.RankingMetrics 234 | val predictedAndTrueForRanking = allRecs.join(userProducts).map{ case (userId, (predicted, actualWithIds)) => 235 | val actual = actualWithIds.map(_._2) 236 | (predicted.toArray, actual.toArray) 237 | } 238 | val rankingMetrics = new RankingMetrics(predictedAndTrueForRanking) 239 | println("Mean Average Precision = " + rankingMetrics.meanAveragePrecision) 240 | // Mean Average Precision = 0.07171412913757183 241 | 242 | // Compare to our implementation, using K = 2000 to approximate the overall MAP 243 | val MAPK2000 = allRecs.join(userProducts).map{ case (userId, (predicted, actualWithIds)) => 244 | val actual = actualWithIds.map(_._2).toSeq 245 | avgPrecisionK(actual, predicted, 2000) 246 | }.reduce(_ + _) / allRecs.count 247 | println("Mean Average Precision = " + MAPK2000) 248 | 249 | // recommendsByUserTopN.foreachPartition(partitionOfRecords => { 250 | // partitionOfRecords.foreach(pair => { 251 | // val jedis = RedisClient.pool.getResource 252 | // jedis.set(pair._1.toString,pair._2.mkString(",")) 253 | // jedis.close() 254 | // }) 255 | // }) 256 | 257 | } 258 | 259 | /* Compute the cosine similarity between two vectors */ 260 | def cosineSimilarity(vec1: DoubleMatrix, vec2: DoubleMatrix): Double = { 261 | vec1.dot(vec2) / (vec1.norm2() * vec2.norm2()) 262 | } 263 | 264 | /* Function to compute average precision given a set of actual and predicted ratings */ 265 | // Code for this function is based on: https://github.com/benhamner/Metrics 266 | def avgPrecisionK(actual: Seq[Int], predicted: Seq[Int], k: Int): Double = { 267 | val predK = predicted.take(k) 268 | var score = 0.0 269 | var numHits = 0.0 270 | for ((p, i) <- predK.zipWithIndex) { 271 | if (actual.contains(p)) { 272 | numHits += 1.0 273 | score += numHits / (i.toDouble + 1.0) 274 | } 275 | } 276 | if (actual.isEmpty) { 277 | 1.0 278 | } else { 279 | score / scala.math.min(actual.size, k).toDouble 280 | } 281 | } 282 | } 283 | -------------------------------------------------------------------------------- /pom.xml: -------------------------------------------------------------------------------- 1 | 3 | 4.0.0 4 | com.javachen.spark 5 | learning-spark 6 | jar 7 | 1.0-SNAPSHOT 8 | learning-spark 9 | http://maven.apache.org 10 | 11 | 12 | 13 | 14 | UTF-8 15 | UTF-8 16 | 17 | 1.7 18 | 2.10 19 | ${scala.minor.version}.4 20 | 2.1.0 21 | 22 | 2.10.2 23 | 1.7.2 24 | 1.6.3.2.6.5.0-292 25 | 26 | 1.7.12 27 | 28 | 64m 29 | 512m 30 | 31 | 32 | 33 | 34 | 35 | spring-snapshots 36 | http://repo.spring.io/snapshot 37 | 38 | true 39 | 40 | 41 | 42 | spring-milestones 43 | http://repo.spring.io/milestone 44 | 45 | 46 | cloudera repository 47 | https://repository.cloudera.com/artifactory/cloudera-repos/ 48 | 49 | 50 | Sonatype Releases 51 | https://oss.sonatype.org/content/repositories/releases/ 52 | 53 | 54 | 55 | 56 | 57 | org.apache.spark 58 | spark-core_${scala.minor.version} 59 | ${spark.version} 60 | provided 61 | 62 | 63 | org.apache.hadoop 64 | hadoop-client 65 | 66 | 67 | 68 | 69 | org.apache.spark 70 | spark-streaming_${scala.minor.version} 71 | ${spark.version} 72 | 73 | 74 | org.apache.hadoop 75 | hadoop-client 76 | 77 | 78 | 79 | 80 | org.apache.spark 81 | spark-mllib_${scala.minor.version} 82 | ${spark.version} 83 | provided 84 | 85 | 86 | org.apache.hadoop 87 | hadoop-client 88 | 89 | 90 | 91 | 92 | org.apache.spark 93 | spark-sql_${scala.minor.version} 94 | ${spark.version} 95 | provided 96 | 97 | 98 | org.apache.hadoop 99 | hadoop-client 100 | 101 | 102 | org.apache.hive 103 | hive-exec 104 | 105 | 106 | 107 | 108 | org.apache.spark 109 | spark-hive_${scala.minor.version} 110 | ${spark.version} 111 | provided 112 | 113 | 114 | org.apache.hadoop 115 | hadoop-client 116 | 117 | 118 | org.apache.hive 119 | hive-exec 120 | 121 | 122 | 123 | 124 | 125 | 126 | org.apache.hadoop 127 | hadoop-hdfs 128 | ${hadoop.version} 129 | provided 130 | 131 | 132 | javax.servlet 133 | servlet-api 134 | 135 | 136 | javax.servlet.jsp 137 | jsp-api 138 | 139 | 140 | org.mortbay.jetty 141 | servlet-api-2.5 142 | 143 | 144 | com.google.guava 145 | guava 146 | 147 | 148 | 149 | 150 | org.apache.hadoop 151 | hadoop-client 152 | ${hadoop.version} 153 | provided 154 | 155 | 156 | javax.servlet 157 | servlet-api 158 | 159 | 160 | javax.servlet.jsp 161 | jsp-api 162 | 163 | 164 | org.mortbay.jetty 165 | servlet-api-2.5 166 | 167 | 168 | com.google.guava 169 | guava 170 | 171 | 172 | 173 | 174 | 175 | org.apache.hbase 176 | hbase-client 177 | ${hbase.version} 178 | 179 | 180 | org.apache.hadoop 181 | hadoop-core 182 | 183 | 184 | com.sun.jersey 185 | jersey-json 186 | 187 | 188 | org.slf4j 189 | slf4j-log4j12 190 | 191 | 192 | org.mortbay.jetty 193 | servlet-api-2.5 194 | 195 | 196 | com.google.guava 197 | guava 198 | 199 | 200 | 201 | 202 | org.apache.hbase 203 | hbase-server 204 | ${hbase.version} 205 | 206 | 207 | org.apache.hadoop 208 | hadoop-core 209 | 210 | 211 | org.mortbay.jetty 212 | jsp-2.1 213 | 214 | 215 | org.mortbay.jetty 216 | jsp-api-2.1 217 | 218 | 219 | javax.servlet.jsp 220 | jsp-api 221 | 222 | 223 | org.mortbay.jetty 224 | servlet-api-2.5 225 | 226 | 227 | com.sun.jersey 228 | jersey-json 229 | 230 | 231 | org.codehaus.jackson 232 | jackson-mapper-asl 233 | 234 | 235 | org.codehaus.jackson 236 | jackson-core-asl 237 | 238 | 239 | org.codehaus.jackson 240 | jackson-jaxrs 241 | 242 | 243 | javax.servlet 244 | servlet-api 245 | 246 | 247 | javax.servlet.jsp 248 | jsp-api 249 | 250 | 251 | org.mortbay.jetty 252 | servlet-api-2.5 253 | 254 | 255 | com.google.guava 256 | guava 257 | 258 | 259 | 260 | 261 | 262 | 263 | org.scalanlp 264 | breeze_${scala.minor.version} 265 | 266 | 0.13.2 267 | provided 268 | 269 | 270 | org.scala-lang 271 | scala-compiler 272 | ${scala.complete.version} 273 | provided 274 | 275 | 276 | org.scala-lang 277 | scala-reflect 278 | ${scala.complete.version} 279 | provided 280 | 281 | 282 | org.scala-lang 283 | jline 284 | ${scala.complete.version} 285 | provided 286 | 287 | 288 | org.scala-lang 289 | scala-library 290 | ${scala.complete.version} 291 | provided 292 | 293 | 294 | org.scala-lang 295 | scala-actors 296 | ${scala.complete.version} 297 | provided 298 | 299 | 300 | org.scala-lang 301 | scalap 302 | ${scala.complete.version} 303 | provided 304 | 305 | 306 | org.scalaj 307 | scalaj-collection_${scala.minor.version} 308 | 1.6 309 | provided 310 | 311 | 312 | 313 | com.github.scopt 314 | scopt_${scala.minor.version} 315 | 3.7.1 316 | 317 | 318 | 319 | 320 | 321 | 322 | 323 | org.apache.maven.plugins 324 | maven-compiler-plugin 325 | 3.13.0 326 | 327 | true 328 | true 329 | true 330 | ${java.version} 331 | ${java.version} 332 | 333 | -Xlint:all,-serial,-try 334 | 335 | 336 | 337 | 338 | org.apache.maven.plugins 339 | maven-clean-plugin 340 | 2.6.1 341 | 342 | 343 | org.apache.maven.plugins 344 | maven-source-plugin 345 | 2.4 346 | 347 | 348 | attach-sources 349 | 350 | jar 351 | 352 | 353 | 354 | 355 | 356 | org.apache.maven.plugins 357 | maven-jar-plugin 358 | 2.6 359 | 360 | true 361 | 362 | 363 | 364 | 365 | test-jar 366 | 367 | 368 | 369 | 370 | 371 | org.apache.maven.plugins 372 | maven-resources-plugin 373 | 2.7 374 | 375 | UTF-8 376 | 377 | 378 | 379 | 380 | org.apache.maven.plugins 381 | maven-install-plugin 382 | 2.5.2 383 | 384 | true 385 | 386 | 387 | 388 | 389 | org.apache.maven.plugins 390 | maven-eclipse-plugin 391 | 2.10 392 | 393 | true 394 | true 395 | 2.0 396 | 397 | **/*.* 398 | 399 | 400 | 401 | org.springframework.ide.eclipse.core.springbuilder 402 | 403 | 404 | org.eclipse.m2e.core.maven2Builder 405 | 406 | 407 | 408 | org.eclipse.jdt.core.javanature 409 | org.springframework.ide.eclipse.core.springnature 410 | org.eclipse.m2e.core.maven2Nature 411 | 412 | 413 | 414 | 415 | org.apache.maven.plugins 416 | maven-dependency-plugin 417 | 418 | 419 | install 420 | install 421 | 422 | sources 423 | 424 | 425 | 426 | 427 | 428 | 429 | net.alchim31.maven 430 | scala-maven-plugin 431 | 3.4.6 432 | 433 | 434 | scala-compile-first 435 | process-resources 436 | 437 | compile 438 | 439 | 440 | 441 | scala-test-compile-first 442 | process-test-resources 443 | 444 | testCompile 445 | 446 | 447 | 448 | attach-scaladocs 449 | verify 450 | 451 | doc-jar 452 | 453 | 454 | 455 | 456 | ${scala.complete.version} 457 | incremental 458 | true 459 | 460 | -unchecked 461 | -deprecation 462 | -feature 463 | -language:postfixOps 464 | 465 | 466 | -Xms1024m 467 | -Xmx1024m 468 | -XX:PermSize=${PermGen} 469 | -XX:MaxPermSize=${MaxPermGen} 470 | 471 | 472 | -source 473 | ${java.version} 474 | -target 475 | ${java.version} 476 | 477 | 479 | 480 | 481 | org.scalamacros 482 | paradise_${scala.complete.version} 483 | ${scala.macros.version} 484 | 485 | 486 | 487 | 488 | 489 | 490 | org.codehaus.mojo 491 | build-helper-maven-plugin 492 | 1.12 493 | 494 | 495 | add-scala-sources 496 | generate-sources 497 | 498 | add-source 499 | 500 | 501 | 502 | src/main/scala 503 | 504 | 505 | 506 | 507 | add-scala-test-sources 508 | generate-test-sources 509 | 510 | add-test-source 511 | 512 | 513 | 514 | src/test/scala 515 | 516 | 517 | 518 | 519 | 520 | 521 | 522 | org.apache.maven.plugins 523 | maven-shade-plugin 524 | 1.7.1 525 | 526 | 527 | package 528 | 529 | shade 530 | 531 | 532 | 533 | 534 | 535 | 537 | 538 | 539 | 540 | 541 | 542 | *:* 543 | 544 | META-INF/*.SF 545 | META-INF/*.DSA 546 | META-INF/*.RSA 547 | 548 | 549 | 550 | 551 | 552 | 553 | 554 | 555 | 556 | 557 | org.eclipse.m2e 558 | lifecycle-mapping 559 | 1.0.0 560 | 561 | 562 | 563 | 564 | 565 | org.apache.maven.plugins 566 | maven-enforcer-plugin 567 | [1.0.0,) 568 | 569 | enforce 570 | 571 | 572 | 573 | 574 | 575 | 576 | 577 | 578 | org.apache.maven.plugins 579 | maven-dependency-plugin 580 | [2.4,) 581 | 582 | unpack 583 | sources 584 | 585 | 586 | 587 | 588 | 589 | 590 | 591 | 592 | 593 | 594 | 595 | 596 | 597 | 598 | --------------------------------------------------------------------------------