README.md

Model API with AWS Lambda

This folder contains the code and supporting files needed to create a serverless model API. It includes the following files and folders:

• api/app.py - Code for the application's Lambda function.
• api/Dockerfile - The Dockerfile to build the container image for local testing.
• api/requirements.txt - The pip requirements installed in the test container.

Start API locally in a Docker container

Requirements:

• Docker

Create and start the API as a local docker container using docker-compose:

docker compose build lambda
docker compose -f docker-compose.yml -f docker-compose.override.yml up -d lambda

Testing with curl:

curl -POST "http://localhost:9000/2015-03-31/functions/function/invocations" -d '{"text": "A test sentence."}'

Testing with python and requests:

import json

import requests

url = "http://localhost:9000/2015-03-31/functions/function/invocations"
headers = {"Content-Type": "application/json"}
payload = json.dumps({"text": 'A test sentence.'})
response = requests.post(url, data=payload, headers=headers)

Testing running the code in the services/jupyter/src/notebooks/query_model_api.ipynb notebook.

Deploy API to AWS Lambda

Requirements:

• Docker
• An AWS Account
• AWS CLI installed and configured
• a .env file in the root of this project (based on .env.sample). Set the AWS_ACCOUNT_ID and AWS_REGION variables in this file.

1. Create a docker image and push to the ECR repository: make deploy_to_aws_ecr
2. Create a IAM role for the Lambda function: make create_lambda_role
3. Create the Lambda function: make create_lambda_function

Contributing to the Lambda API service

If you want to contribute to this service you should:

1. Update one or more of the following files (based on how you want to change the API):
   • api/app.py: the logic of the API, the lambda_handler function is the entrypoint for the AWS Lambda function
   • api/requirements.txt: the python requirements of the app
   • api/Dockerfile: in case you want to change external dependency or the build of the image
2. Create and test API locally as described in 'Create API locally in a Docker container'