Merge remote-tracking branch 'upstream/master' into topic/streaming-bp/rate-controller

Author: dragos

R/install-dev.bat

@@ -25,3 +25,8 @@ set SPARK_HOME=%~dp0..
 MKDIR %SPARK_HOME%\R\lib
 
 R.exe CMD INSTALL --library="%SPARK_HOME%\R\lib"  %SPARK_HOME%\R\pkg\
+
+rem Zip the SparkR package so that it can be distributed to worker nodes on YARN
+pushd %SPARK_HOME%\R\lib
+%JAVA_HOME%\bin\jar.exe cfM "%SPARK_HOME%\R\lib\sparkr.zip" SparkR
+popd

R/install-dev.sh

@@ -34,12 +34,16 @@ LIB_DIR="$FWDIR/lib"
 
 mkdir -p $LIB_DIR
 
-pushd $FWDIR
+pushd $FWDIR > /dev/null
 
 # Generate Rd files if devtools is installed
 Rscript -e ' if("devtools" %in% rownames(installed.packages())) { library(devtools); devtools::document(pkg="./pkg", roclets=c("rd")) }'
 
 # Install SparkR to $LIB_DIR
 R CMD INSTALL --library=$LIB_DIR $FWDIR/pkg/
 
-popd
+# Zip the SparkR package so that it can be distributed to worker nodes on YARN
+cd $LIB_DIR
+jar cfM "$LIB_DIR/sparkr.zip" SparkR
+
+popd > /dev/null

R/pkg/DESCRIPTION

@@ -29,7 +29,7 @@ Collate:
     'client.R'
     'context.R'
     'deserialize.R'
+    'mllib.R'
     'serialize.R'
     'sparkR.R'
     'utils.R'
-    'zzz.R'

R/pkg/NAMESPACE

@@ -10,6 +10,10 @@ export("sparkR.init")
 export("sparkR.stop")
 export("print.jobj")
 
+# MLlib integration
+exportMethods("glm",
+              "predict")
+
 # Job group lifecycle management methods
 export("setJobGroup",
        "clearJobGroup",
@@ -22,6 +26,7 @@ exportMethods("arrange",
               "collect",
               "columns",
               "count",
+              "crosstab",
               "describe",
               "distinct",
               "dropna",
@@ -77,6 +82,7 @@ exportMethods("abs",
               "atan",
               "atan2",
               "avg",
+              "between",
               "cast",
               "cbrt",
               "ceiling",

R/pkg/R/DataFrame.R

@@ -1314,7 +1314,7 @@ setMethod("except",
 #' write.df(df, "myfile", "parquet", "overwrite")
 #' }
 setMethod("write.df",
-          signature(df = "DataFrame", path = 'character'),
+          signature(df = "DataFrame", path = "character"),
           function(df, path, source = NULL, mode = "append", ...){
             if (is.null(source)) {
               sqlContext <- get(".sparkRSQLsc", envir = .sparkREnv)
@@ -1328,7 +1328,7 @@ setMethod("write.df",
             jmode <- callJStatic("org.apache.spark.sql.api.r.SQLUtils", "saveMode", mode)
             options <- varargsToEnv(...)
             if (!is.null(path)) {
-                options[['path']] = path
+                options[["path"]] <- path
             }
             callJMethod(df@sdf, "save", source, jmode, options)
           })
@@ -1337,7 +1337,7 @@ setMethod("write.df",
 #' @aliases saveDF
 #' @export
 setMethod("saveDF",
-          signature(df = "DataFrame", path = 'character'),
+          signature(df = "DataFrame", path = "character"),
           function(df, path, source = NULL, mode = "append", ...){
             write.df(df, path, source, mode, ...)
           })
@@ -1375,8 +1375,8 @@ setMethod("saveDF",
 #' saveAsTable(df, "myfile")
 #' }
 setMethod("saveAsTable",
-          signature(df = "DataFrame", tableName = 'character', source = 'character',
-                    mode = 'character'),
+          signature(df = "DataFrame", tableName = "character", source = "character",
+                    mode = "character"),
           function(df, tableName, source = NULL, mode="append", ...){
             if (is.null(source)) {
               sqlContext <- get(".sparkRSQLsc", envir = .sparkREnv)
@@ -1554,3 +1554,31 @@ setMethod("fillna",
             }
             dataFrame(sdf)
           })
+
+#' crosstab
+#'
+#' Computes a pair-wise frequency table of the given columns. Also known as a contingency
+#' table. The number of distinct values for each column should be less than 1e4. At most 1e6
+#' non-zero pair frequencies will be returned.
+#'
+#' @param col1 name of the first column. Distinct items will make the first item of each row.
+#' @param col2 name of the second column. Distinct items will make the column names of the output.
+#' @return a local R data.frame representing the contingency table. The first column of each row
+#'         will be the distinct values of `col1` and the column names will be the distinct values
+#'         of `col2`. The name of the first column will be `$col1_$col2`. Pairs that have no
+#'         occurrences will have `null` as their counts.
+#'
+#' @rdname statfunctions
+#' @export
+#' @examples
+#' \dontrun{
+#' df <- jsonFile(sqlCtx, "/path/to/file.json")
+#' ct = crosstab(df, "title", "gender")
+#' }
+setMethod("crosstab",
+          signature(x = "DataFrame", col1 = "character", col2 = "character"),
+          function(x, col1, col2) {
+            statFunctions <- callJMethod(x@sdf, "stat")
+            sct <- callJMethod(statFunctions, "crosstab", col1, col2)
+            collect(dataFrame(sct))
+          })

R/pkg/R/RDD.R

@@ -165,7 +165,6 @@ setMethod("getJRDD", signature(rdd = "PipelinedRDD"),
                                    serializedFuncArr,
                                    rdd@env$prev_serializedMode,
                                    packageNamesArr,
-                                   as.character(.sparkREnv[["libname"]]),
                                    broadcastArr,
                                    callJMethod(prev_jrdd, "classTag"))
             } else {
@@ -175,7 +174,6 @@ setMethod("getJRDD", signature(rdd = "PipelinedRDD"),
                                    rdd@env$prev_serializedMode,
                                    serializedMode,
                                    packageNamesArr,
-                                   as.character(.sparkREnv[["libname"]]),
                                    broadcastArr,
                                    callJMethod(prev_jrdd, "classTag"))
             }

R/pkg/R/SQLContext.R

@@ -457,7 +457,7 @@ dropTempTable <- function(sqlContext, tableName) {
 read.df <- function(sqlContext, path = NULL, source = NULL, schema = NULL, ...) {
   options <- varargsToEnv(...)
   if (!is.null(path)) {
-    options[['path']] <- path
+    options[["path"]] <- path
   }
   if (is.null(source)) {
     sqlContext <- get(".sparkRSQLsc", envir = .sparkREnv)
@@ -506,7 +506,7 @@ loadDF <- function(sqlContext, path = NULL, source = NULL, schema = NULL, ...) {
 createExternalTable <- function(sqlContext, tableName, path = NULL, source = NULL, ...) {
   options <- varargsToEnv(...)
   if (!is.null(path)) {
-    options[['path']] <- path
+    options[["path"]] <- path
   }
   sdf <- callJMethod(sqlContext, "createExternalTable", tableName, source, options)
   dataFrame(sdf)

R/pkg/R/client.R

@@ -36,9 +36,9 @@ connectBackend <- function(hostname, port, timeout = 6000) {
 
 determineSparkSubmitBin <- function() {
   if (.Platform$OS.type == "unix") {
-    sparkSubmitBinName = "spark-submit"
+    sparkSubmitBinName <- "spark-submit"
   } else {
-    sparkSubmitBinName = "spark-submit.cmd"
+    sparkSubmitBinName <- "spark-submit.cmd"
   }
   sparkSubmitBinName
 }

R/pkg/R/column.R

@@ -187,6 +187,23 @@ setMethod("substr", signature(x = "Column"),
             column(jc)
           })
 
+#' between
+#'
+#' Test if the column is between the lower bound and upper bound, inclusive.
+#'
+#' @rdname column
+#'
+#' @param bounds lower and upper bounds
+setMethod("between", signature(x = "Column"),
+          function(x, bounds) {
+            if (is.vector(bounds) && length(bounds) == 2) {
+              jc <- callJMethod(x@jc, "between", bounds[1], bounds[2])
+              column(jc)
+            } else {
+              stop("bounds should be a vector of lower and upper bounds")
+            }
+          })
+
 #' Casts the column to a different data type.
 #'
 #' @rdname column

R/pkg/R/deserialize.R

@@ -23,6 +23,7 @@
 # Int -> integer
 # String -> character
 # Boolean -> logical
+# Float -> double
 # Double -> double
 # Long -> double
 # Array[Byte] -> raw

R/pkg/R/generics.R

@@ -59,6 +59,10 @@ setGeneric("count", function(x) { standardGeneric("count") })
 # @export
 setGeneric("countByValue", function(x) { standardGeneric("countByValue") })
 
+# @rdname statfunctions
+# @export
+setGeneric("crosstab", function(x, col1, col2) { standardGeneric("crosstab") })
+
 # @rdname distinct
 # @export
 setGeneric("distinct", function(x, numPartitions = 1) { standardGeneric("distinct") })
@@ -567,6 +571,10 @@ setGeneric("asc", function(x) { standardGeneric("asc") })
 #' @export
 setGeneric("avg", function(x, ...) { standardGeneric("avg") })
 
+#' @rdname column
+#' @export
+setGeneric("between", function(x, bounds) { standardGeneric("between") })
+
 #' @rdname column
 #' @export
 setGeneric("cast", function(x, dataType) { standardGeneric("cast") })
@@ -657,3 +665,7 @@ setGeneric("toRadians", function(x) { standardGeneric("toRadians") })
 #' @rdname column
 #' @export
 setGeneric("upper", function(x) { standardGeneric("upper") })
+
+#' @rdname glm
+#' @export
+setGeneric("glm")

R/pkg/R/group.R

@@ -87,7 +87,7 @@ setMethod("count",
 setMethod("agg",
           signature(x = "GroupedData"),
           function(x, ...) {
-            cols = list(...)
+            cols <- list(...)
             stopifnot(length(cols) > 0)
             if (is.character(cols[[1]])) {
               cols <- varargsToEnv(...)
@@ -97,7 +97,7 @@ setMethod("agg",
               if (!is.null(ns)) {
                 for (n in ns) {
                   if (n != "") {
-                    cols[[n]] = alias(cols[[n]], n)
+                    cols[[n]] <- alias(cols[[n]], n)
                   }
                 }
               }

R/pkg/R/mllib.R

@@ -0,0 +1,73 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements.  See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You under the Apache License, Version 2.0
+# (the "License"); you may not use this file except in compliance with
+# the License.  You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+# mllib.R: Provides methods for MLlib integration
+
+#' @title S4 class that represents a PipelineModel
+#' @param model A Java object reference to the backing Scala PipelineModel
+#' @export
+setClass("PipelineModel", representation(model = "jobj"))
+
+#' Fits a generalized linear model
+#'
+#' Fits a generalized linear model, similarly to R's glm(). Also see the glmnet package.
+#'
+#' @param formula A symbolic description of the model to be fitted. Currently only a few formula
+#'                operators are supported, including '~' and '+'.
+#' @param data DataFrame for training
+#' @param family Error distribution. "gaussian" -> linear regression, "binomial" -> logistic reg.
+#' @param lambda Regularization parameter
+#' @param alpha Elastic-net mixing parameter (see glmnet's documentation for details)
+#' @return a fitted MLlib model
+#' @rdname glm
+#' @export
+#' @examples
+#'\dontrun{
+#' sc <- sparkR.init()
+#' sqlContext <- sparkRSQL.init(sc)
+#' data(iris)
+#' df <- createDataFrame(sqlContext, iris)
+#' model <- glm(Sepal_Length ~ Sepal_Width, df)
+#'}
+setMethod("glm", signature(formula = "formula", family = "ANY", data = "DataFrame"),
+          function(formula, family = c("gaussian", "binomial"), data, lambda = 0, alpha = 0) {
+            family <- match.arg(family)
+            model <- callJStatic("org.apache.spark.ml.api.r.SparkRWrappers",
+                                 "fitRModelFormula", deparse(formula), data@sdf, family, lambda,
+                                 alpha)
+            return(new("PipelineModel", model = model))
+          })
+
+#' Make predictions from a model
+#'
+#' Makes predictions from a model produced by glm(), similarly to R's predict().
+#'
+#' @param model A fitted MLlib model
+#' @param newData DataFrame for testing
+#' @return DataFrame containing predicted values
+#' @rdname glm
+#' @export
+#' @examples
+#'\dontrun{
+#' model <- glm(y ~ x, trainingData)
+#' predicted <- predict(model, testData)
+#' showDF(predicted)
+#'}
+setMethod("predict", signature(object = "PipelineModel"),
+          function(object, newData) {
+            return(dataFrame(callJMethod(object@model, "transform", newData@sdf)))
+          })

R/pkg/R/pairRDD.R

@@ -215,7 +215,6 @@ setMethod("partitionBy",
                                        serializedHashFuncBytes,
                                        getSerializedMode(x),
                                        packageNamesArr,
-                                       as.character(.sparkREnv$libname),
                                        broadcastArr,
                                        callJMethod(jrdd, "classTag"))
 

R/pkg/R/schema.R

@@ -69,11 +69,14 @@ structType.structField <- function(x, ...) {
 #' @param ... further arguments passed to or from other methods
 print.structType <- function(x, ...) {
   cat("StructType\n",
-      sapply(x$fields(), function(field) { paste("|-", "name = \"", field$name(),
-                                           "\", type = \"", field$dataType.toString(),
-                                           "\", nullable = ", field$nullable(), "\n",
-                                           sep = "") })
-      , sep = "")
+      sapply(x$fields(),
+             function(field) {
+               paste("|-", "name = \"", field$name(),
+                     "\", type = \"", field$dataType.toString(),
+                     "\", nullable = ", field$nullable(), "\n",
+                     sep = "")
+             }),
+      sep = "")
 }
 
 #' structField
@@ -123,6 +126,7 @@ structField.character <- function(x, type, nullable = TRUE) {
   }
   options <- c("byte",
                "integer",
+               "float",
                "double",
                "numeric",
                "character",

R/pkg/R/serialize.R

@@ -140,8 +140,8 @@ writeType <- function(con, class) {
                  jobj = "j",
                  environment = "e",
                  Date = "D",
-                 POSIXlt = 't',
-                 POSIXct = 't',
+                 POSIXlt = "t",
+                 POSIXct = "t",
                  stop(paste("Unsupported type for serialization", class)))
   writeBin(charToRaw(type), con)
 }