aws-lambda-r-runtime

Copy from https://github.com/bakdata/aws-lambda-r-runtime

This package makes it easy to run AWS Lambda Functions written in R.

Build custom R lambda layer

In order to install additional R packages than on the default layer, you can create a lambda layer yourself containing the libraries you want. You must use the the compiled package files. The easiest way is to install the package with install.packages() and copy the resulting folder in $R_LIBS. Using only the package sources does not suffice. The file structure must be R/library/<MY_LIBRARY>. See build_recommended.sh for an example. If your package requires system libraries, place them in R/lib/.

STEP 1. Create AWS EC2 instance

Start an EC2 instance which uses the Lambda AMI.

Or command line (then you need your Keypair already in place (step 2)).

aws ec2 run-instances --image-id ami-657bd20a --count 1 --instance-type t2.medium --key-name <MyKeyPair>

STEP 2. Create and load key MyKeyPair

Step 1 link gives you step by step instructions and asks you to create key pair if you don't have one you want to use already.

STEP 3. Connect to instance

You can copy this line of code from the AWS console by clicking your instance and connect.

ssh -i "mykey-ec2.pem" ec2-user@XXXXXX.us-west-2.compute.amazonaws.com

STEP 4. Run (or copy, modify and run) build_r.sh script

Now run the build_r.sh script. You must pass the R version as a parameter to the script, e.g., 3.5.1. The script produces a zip containing a functional R installation in /opt/R/, e.g., /opt/R/R-3.5.1.zip. Use this R distribution in the following.

STEP 5. Create S3 bucket

Eg. AWS quickstart guides

STEP 6. Save R zip to S3 for further use

Create a new user with AWS S3 permission and load secrets. Configure s3 keys on EC2.

aws configure set aws_access_key_id XXXXXXX
aws configure set aws_secret_access_key XXXXX

An then copy using command line (replace $VERSION with your version).

aws s3 cp /opt/R/R-$VERSION.zip \
  s3://bucket_name/layers/R-3.5.1.zip

STEP 7. Build lambda layer (finally!)

With a compiled R distribution, you can build the runtime layer. See the #remove some libraries to save space recommended and modify to save space, since Lambda layer cannot exceed 50Mt.

If you plan to publish the runtime, you need to have a recent version of aws cli (>=1.16). Copy the R distribution to the repository containing necessary scripts (build_runtime.sh, runtime.R) and run the build_runtime_and_publish.sh script. This creates a lambda layer named r-runtime in your AWS account. You can see the result in console.

Build lambda function with R

You need a function you want to work in the Lambda. Eg. return value for a trained model.

STEP 1. Create R script

Create a script that has function in it. Eg. example folder lm_cars.r or

require(stats)
linearMod <- lm(dist ~ speed, data=cars)  # build linear regression model on full data

make_prediction <- function(x) {

  new_df <- data.frame(speed = x)
  prediction<- predict(linearMod, new_df)
  return(prediction)
}

Or if you wish to load your model from S3. Here pre-trained ruimtehol model.

library(aws.s3)
library(ruimtehol)

# get model file and save file locally
readRenviron('.Renviron')

modelfile <- save_object("s3://YOU_FOLFDER/textspace.ruimtehol",
                key = Sys.getenv("AWS_ACCESS_KEY_ID"),
                secret = Sys.getenv("AWS_SECRET_ACCESS_KEY"),
                region = Sys.getenv("AWS_DEFAULT_REGION"),
                file = '~/textspace.ruimtehol'
                )
model <- starspace_load_model(modelfile)

make_prediction <- function(x) {
    prediction <- predict(model, x, k = 3)
    return(prediction)}

STEP 2. Create Lambda

To run the example, we need to create a IAM role executing our lambda. This role should have the following properties:

• Trusted entity – Lambda.
• Permissions – AWSLambdaAllpermissions. Copy you role eg. arn:aws:iam::XXXXXXXX:role/r-runtime and paste to the script.

Furthermore you need a current version of the AWS CLI.

Then create a lambda function which uses the R runtime layer with create_lambda.sh or example:

#cd example/
chmod 755 lm_cars.r #which script you want to put in the Lambda
zip function.zip script.r #zip the function
aws lambda create-function --function-name r-example \ #give the Lambda function name
    --zip-file fileb://function.zip --handler lm_cars.make_prediction \ #tell which script and function
    --runtime provided --timeout 60 \
    --layers arn:aws:lambda:eu-central-1:131329294410:layer:r-runtime-3_5_1:1 \ #basic layer
      #arn:aws:lambda:eu-central-1:XXXXXXX:layer:r-runtime:1 \ #own layer with some other packages
    --role arn:aws:iam::XXXXXXXX:role/r-runtime --region eu-central-1 #paste your role info here

Note, using eg. 3-4 layers requires too much so keep it simple!

STEP 3. Test Lambda

Either in console setting a test case manually or with command line. Setting the test case, you can see the possible errors on console. Function returns output of the model. The lambda function returns whatever is returned by the R function as a JSON object with result as a root element.

Invoke the function:

aws lambda invoke --function-name r-example \
    --payload '{"x":4}' --region eu-central-1 response.txt
cat response.txt ##see the response

The expected result should look similar to this:

{
  "result": -1.8495
}

Provided layers

Layers are only accessible in the AWS region they were published. See provided layer from https://github.com/bakdata/aws-lambda-r-runtime .

Limitations

AWS Lambda is limited to running with 3GB RAM and must finish within 15 minutes. It is therefore not feasible to execute long running R scripts with this runtime. Furthermore, only the /tmp/ directory is writeable on AWS Lambda. This must be considered when writing to the local disk.