Merge branch 'master' into tomas-parse

Conflicts:
	src/main/java/hex/glm/GLM2.java
	src/main/java/water/DTask.java
	src/main/java/water/H2O.java
	src/main/java/water/Job.java
	src/main/java/water/RPC.java
	src/main/java/water/api/GLMPredict.java
	src/main/java/water/fvec/ParseDataset2.java
	src/main/java/water/parser/ParseDataset.java
	src/test/java/water/fvec/DatasetCompare.java

Makefile

@@ -98,7 +98,7 @@ build:
 	$(MAKE) build_h2o PROJECT_VERSION=$(PROJECT_VERSION)
 
 	@echo
-	@echo "PHASE: Building hadoop driver..."
+	@echo "PHASE: Building Hadoop driver..."
 	@echo
 	$(MAKE) -C hadoop build PROJECT_VERSION=$(PROJECT_VERSION) 1> target/logs/hadoop_build.log
 
@@ -119,6 +119,11 @@ build:
 	@echo
 	$(MAKE) build_package PROJECT_VERSION=$(PROJECT_VERSION) 1> target/logs/package_build.log
 
+	@echo
+	@echo "PHASE: Building ZooKeeper jar..."
+	@echo
+	$(MAKE) -C h2o-zookeeper PROJECT_VERSION=$(PROJECT_VERSION) 1> target/logs/zookeeper_build.log
+
 BUILD_BRANCH=$(shell git branch | grep '*' | sed 's/* //')
 BUILD_HASH=$(shell git log -1 --format="%H")
 BUILD_DESCRIBE=$(shell git describe --always --dirty)
@@ -304,6 +309,7 @@ clean:
 	$(MAKE) -C client clean
 	$(MAKE) -C h2o-scala clean
 	$(MAKE) -C hadoop clean
+	$(MAKE) -C h2o-zookeeper clean
 	$(MAKE) -C R clean
 	$(MAKE) -C launcher clean
 	$(MAKE) -C installer clean

R/H2O_Load.R

@@ -1,9 +1,9 @@
 # Change this global variable to match your own system's path
-ROOT.PATH <- "/Users/spencer/master/h2o/R/h2o-package/R/"
+ROOT.PATH <- "/Users/anqi_fu/Documents/workspace/h2o/R/h2o-package/R/"
 src <-
 function() {
   warning("MAY NOT WORK ON YOUR SYSTEM -- **TRY TO CHANGE `ROOT.PATH`!**")
-  to_src <- c("Wrapper.R", "Internal.R", "Classes.R", "ParseImport.R", "Algorithms.R")
+  to_src <- c("Wrapper.R", "Internal.R", "Classes.R", "ParseImport.R", "models.R", "Algorithms.R")
   require(rjson); require(RCurl)
   invisible(lapply(to_src,function(x){source(paste(ROOT.PATH, x, sep = ""))}))
 }

R/tests/testdir_demos/runit_demo_model_by_group_extended_large.R → R/examples/runit_demo_model_by_group_extended_large.R

@@ -11,7 +11,7 @@
 #' What this demo covers:
 #'  Algorithms:
 #'   h2o.randomForest         --The distributed random forest algorithm
-#'   h2o.SpeeDRF              --A faster random forest, but not as accurate
+#'   h2o.randomForest              --A faster random forest, but not as accurate
 #'   h2o.deeplearning         --Distributed, parallel deep learning
 #'   h2o.gbm                  --Gradient Boosted Machines (a tree-based algo)
 #'   h2o.glm                  --GLM again
@@ -39,7 +39,7 @@ library(plyr)
 
 # Read in the data
 # Path is relative to the location that I started h2o (i.e. which dir did I java -jar in?)
-flights <- h2o.importFile(h, "../../../smalldata/airlines/allyears2k_headers.zip", "flights.hex")
+flights <- h2o.importFile(h, normalizePath("../../../smalldata/airlines/allyears2k_headers.zip"), "flights.hex")
 
 #################################################################################
 #
@@ -95,7 +95,7 @@ function(origin, dataset) {
   dataset <- dataset[dataset$Origin == origin,]
   print("Beginning Random Forest with 50 trees, 20 depth, and 10-fold Cross Validation\n")
   t0 <- Sys.time()
-  model <- h2o.randomForest(x = c(FlightDate, ScheduledTimes, FlightInfo), y = Delayed, data = dataset, ntree = 50, depth = 20, nfolds = 10) 
+  model <- h2o.randomForest(x = c(FlightDate, ScheduledTimes, FlightInfo), y = Delayed, data = dataset, ntree = 50, depth = 20, nfolds = 10, type = "BigData") 
   elapsed_seconds <- as.numeric(Sys.time() - t0) 
   modelkey <- model@key
   result <- list(list(model, origin, elapsed_seconds))
@@ -108,7 +108,7 @@ function(origin, dataset) {
   dataset <- dataset[dataset$Origin == origin,]
   print("Beginning Speedy Random Forest with 50 trees, 20 depth, and 10-fold Cross Validation\n")
   t0 <- Sys.time()
-  model <- h2o.SpeeDRF(x = c(FlightDate, ScheduledTimes, FlightInfo), y = Delayed, data = dataset, ntree = 50, depth = 20, nfolds = 10) 
+  model <- h2o.randomForest(x = c(FlightDate, ScheduledTimes, FlightInfo), y = Delayed, data = dataset, ntree = 50, depth = 20, nfolds = 10, type = "fast")
   elapsed_seconds <- as.numeric(Sys.time() - t0) 
   modelkey <- model@key
   result <- list(list(model, origin, elapsed_seconds))
@@ -148,7 +148,7 @@ function(fitMethod, origins, dataset) {
 }
 
 #iterate over the fit fcns as well
-model.fit.fcns <- c(lr.fit, rf.fit, srf.fit, gbm.fit, dl.fit)
+model.fit.fcns <- c(lr.fit, srf.fit) #rf.fit, srf.fit, gbm.fit, dl.fit)
 # See the Notes section below to get insight into the following one-liner
 models.by.airport.origin <- unlist(recursive = F, lapply(model.fit.fcns, all.fit, frequent.origin.codes, flights))
 

R/tests/testdir_demos/runit_demo_model_score_many_predictors_large.R → R/examples/runit_demo_model_score_many_predictors_large.R

@@ -38,7 +38,7 @@ library(plyr)
 
 
 # Read in the data
-flights <- h2o.importFile(h, "../../../smalldata/airlines/allyears2k_headers.zip", "flights.hex")
+flights <- h2o.importFile(h, normalizePath("../../../smalldata/airlines/allyears2k_headers.zip"), "flights.hex")
 
 #################################################################################
 #
@@ -168,7 +168,7 @@ rf.fit<-
 function(response, dataset, testdata) {
   print("Beginning Random Forest with 10 trees, 20 depth, and 2-fold Cross Validation\n")
   t0 <- Sys.time()
-  model <- h2o.randomForest(x = c(FlightDate, ScheduledTimes, FlightInfo), y = response, data = dataset, ntree = 10, depth = 20, nfolds = 2, balance.classes = T) 
+  model <- h2o.randomForest(x = c(FlightDate, ScheduledTimes, FlightInfo), y = response, data = dataset, ntree = 10, depth = 20, nfolds = 2, balance.classes = T, type = "BigData") 
   elapsed_seconds <- as.numeric(Sys.time() - t0) 
   modelkey <- model@key
 
@@ -185,7 +185,7 @@ srf.fit<-
 function(response, dataset, testdata) {
   print("Beginning Speedy Random Forest with 10 trees, 20 depth, and 2-fold Cross Validation\n")
   t0 <- Sys.time()
-  model <- h2o.SpeeDRF(x = c(FlightDate, ScheduledTimes, FlightInfo), y = response, data = dataset, ntree = 10, depth = 20, nfolds = 2, balance.classes = T)
+  model <- h2o.randomForest(x = c(FlightDate, ScheduledTimes, FlightInfo), y = response, data = dataset, ntree = 10, depth = 20, nfolds = 2, balance.classes = T, type = "fast")
   elapsed_seconds <- as.numeric(Sys.time() - t0) 
   modelkey <- model@key
 
@@ -238,7 +238,7 @@ function(fitMethod, responses, dataset, testdata) {
 }
 
 #iterate over the fit fcns as well as the tgts
-model.fit.fcns <- c(lr.fit, rf.fit, srf.fit, gbm.fit, dl.fit)
+model.fit.fcns <- c(lr.fit, rf.fit, srf.fit, gbm.fit)#, dl.fit)
 
 # This will loop over all of the models and score for each of the responses in tgts
 models.by.tgt <- unlist(recursive = F, lapply(model.fit.fcns, all.fit, tgts, train, test))

R/h2o-DESCRIPTION.template

@@ -4,11 +4,11 @@ Title: H2O R Interface
 Version: SUBST_PROJECT_VERSION
 Date: 2014-05-15
 Author: Anqi Fu, Tom Kraljevic and Petr Maj, with contributions from the 0xdata team
-Maintainer: Anqi Fu <anqi@0xdata.com>
+Maintainer: Ariel Rao <ariel@0xdata.com>
 Description: This is a package for running H2O via its REST API from within R.
 License: Apache License (== 2.0)
 Depends: R (>= 2.13.0), RCurl, rjson, statmod, tools, methods, utils
-Collate: Wrapper.R Internal.R Classes.R ParseImport.R Algorithms.R
+Collate: Wrapper.R Internal.R Classes.R ParseImport.R models.R Algorithms.R
 NeedsCompilation: no
 SystemRequirements: Java (>= 1.6)
 URL: http://www.0xdata.com

R/h2o-package.template

@@ -27,7 +27,7 @@ Note that no actual data is stored in the R workspace; and no actual work is car
 \author{
 Anqi Fu, Tom Kraljevic and Petr Maj, with contributions from the 0xdata team
 
-Maintainer: Anqi Fu <anqi@0xdata.com>
+Maintainer: Ariel Rao <ariel@0xdata.com>
 }
 \references{
 \itemize{