Skip to content

Latest commit

 

History

History
118 lines (82 loc) · 3 KB

r4ml-install.md

File metadata and controls

118 lines (82 loc) · 3 KB
Raw

#R4ML Logo

How to install

1.0 Install Dependencies

1.1 Install Core R

Install R as per https://cloud.r-project.org/

2.0 Install SparkR

2.1 Install core SparkR

You can follow one of the following steps

  • 2.1.1: download prebuilt binary
  • 2.1.2: build your own or if you already have spark source code
2.1.1 Direct SparkR download

Download directly from spark resources. http://spark.apache.org/downloads.html

Issue the following shell commands

# assumming that the downloaded file is at /tmp/spark-2.1.0-bin-hadoop2.7.tgz
mkdir -p ~/codait
pushd ~/codait
mv /tmp/spark-2.1.0-bin-hadoop2.7.tgz .

# extract the spark binary 
tar -xvzf spark-2.1.0-bin-hadoop2.7.tgz

# rename/remove the appropriately
mv spark-2.1.0-bin-hadoop2.7 spark; \rm -rf spark-2.1.0-bin-hadoop2.7.tgz

popd
2.1.2 Download Source and build it

Execute the following shell script

mkdir -p ~/codait
pushd ~/codait

# 2.1 maintenance branch with stability fixes on top of Spark 2.1.0
git clone git://github.com/apache/spark.git -b branch-2.1

# build the sparkr
pushd spark; mvn -DskipTests -Pnetlib-lgpl -Psparkr clean package; popd

popd
2.2 Set up the environment path in the .Renviron

A) first get the existing system library path by running on the terminal

 Rscript -e "paste(.libPaths(), collapse=':')"

B) Add the following lines in ~/.Renviron (note: replace <OUTPUT_FROM_PREVIOUS_CODE> appropriately

# change <VAR> to the appropriate value
R_LIBS=<OUTPUT_FROM_PREVIOUS_CODE>:~/codait/spark/R/lib:$R_LIBS

C) Add the SPARK_HOME env variable in ~/.Renviron

# SPARK_HOME=<SPARK_INSTALL_DIR>/spark
SPARK_HOME=~/codait/spark

3.0 Install R4ML's dependencies (R packages)

Install all the as per the DESCRIPTION file of R4ML using the install.packages R cmd Here is the R shell command to achieve the above

# start the R shell in the OS prompt
bash>R
# now you have the R shell, run the following R cmds
R>

# must have the following packages
R>r4ml_mandatory_pkgs <- c("uuid", "R6", "survival")
R>install.packages(r4ml_mandatory_pkgs)

# the following packages are needed for dev, testing and documentations
R>r4ml_dev_pkgs <- c("testthat", "knitr", "devtools", "roxygen2")
R>install.packages(r4ml_dev_pkgs)

R>quit("no")

4.0 Install R4ML

A) clone github and install R4ML.

mkdir -p ~/codait
pushd ~/codait
git clone https://github.com/CODAIT/R4ML R4ML
pushd R4ML
bin/install-all.sh
popd
popd

B) Add the following path to the R_LIBS in the $HOME/.Renviron file

# note: replace <VAR> with the appropriate value for your environment
R_LIBS=<PREVIOUS_PATH_FOR_R_LIBRARY>:<PATH_FOR_SPARKR_LIB>:~/codait/R4ML/lib:$R_LIBS