In sampoll/rLambda:

The AWS Lambda function is registered for use by our group account. By default, AWS will look for credentials in the file ~/.aws/credentials. This default can be overridden by setting environment variables AWS_ACCESS_KEY_ID and
AWS_SECRET_ACCESS_KEY.

The package uses the local directory for temporary file storage. It is recommended that the user make an empty temporary directory to run in.

# system("mkdir -p /tmp/testdir")
# setwd("/tmp/testdir")
# closeAllConnections()    # a wise precaution

Initialize python with the path to the desired python executable.

library(rLambda)
rLambdaInitialize(path="/anaconda3/bin/python")

rLambda runs the no-parameter function run() in the current environment, and assign the return value to result.

This means the user should:

• Define a run function that takes no parameters.
• Avoid defining objects in the environment that run does not use.
• Design run to return an object and/or to assign objects to its enclosing environment.
• Avoid using the name result for an object that should not be overwritten.

rm(list=ls())            # clean up
df <- data.frame(x = 1:10, y = 1:10 + rnorm(n=10))
run <- function()  {
  lm(y ~ x, df)
}
rLambdaExecute()

If no error occurs, there should be a result object in the local environment. However, this is not working at the moment, for reasons not yet understood. So we can load and use the response with:

# load("newdata.Rdata")
# summary(result)

When the result has been used/saved, it may be a good idea to clean up, to prevent leaving things in the execution directory that may interfere with subsequent executions.

# setwd("~")
# system("rm -Rf /tmp/tempdir")

sampoll/rLambda documentation built on May 13, 2019, 1:09 p.m.