This repository contains sample code that uses the libpostal library to parse addresses into parts. The code is deployed as a Lambda function and is packaged as a Docker container. The image size is over 2GB in size because the libpostal library contains NLP models that are more than 2GB.
export AWS_REGION=us-east-2
# Get AWS account ID for later use
ACCOUNT_ID=`aws sts get-caller-identity |jq -r .Account`
# Create ECR repository
aws ecr create-repository --repository-name libpostal-demo
# Package the Docker image
docker build -t libpostal-demo:latest .
# Get ECR Docker login
aws ecr get-login-password --region ${AWS_REGION} | docker login --username AWS --password-stdin ${ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com
# Tag image
docker tag libpostal-demo:latest ${ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com/libpostal-demo:latest
# Push image
docker push ${ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com/libpostal-demo:latest
# Create the CloudFormation stack
aws cloudformation deploy \
--template template.yaml \
--stack-name libpostal-example \
--parameter-overrides \
ImageName=${ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com/libpostal-demo \
ImageTag=latest \
--capabilities CAPABILITY_IAM
# Get the Lambda function name created by CloudFormation
LAMBDA_FUNCTION=`aws cloudformation describe-stacks --stack-name libpostal-example --query "Stacks[0].Outputs[?OutputKey=='FunctionName'].OutputValue" --output text`
# Sample input address
PAYLOAD='{ "address" : "123 33rd St, New York City, NY 10001"}'
PAYLOAD_BASE64=`echo $PAYLOAD |base64`
# Invoke the function and redirect the output to a file named 'output'
aws lambda invoke --function-name $LAMBDA_FUNCTION --payload $PAYLOAD_BASE64 output
# View the output
cat output
Output:
{
"house_number": "123",
"road": "33rd st",
"city": "new york city",
"state": "ny",
"postcode": "10001"
}