Image Pipeline Demo for AWS Lambda

This project contains AWS Lambda function definitions intended to create a image processing pipeline with AWS Step Functions.

• image-pipeline-trigger: Intended to be triggered by Object Create events in an S3 bucket, it will generate a unique ID for the event, and trigger the Step Function with that ID, and the bucket name and object key.
• image-resize: Will create an image thumbnail (using 50% scaling) and drop it into a different S3 bucket.
• image-analysis: Uses AWS Rekognition to generate tags for the image.
• image-db-update: Stores the location of the thumbnail as well as the tags in a DynamoDB table.
• image-api: Read-only API that accesses DynamoDB table.

The S3 Buckets, DynamoDB table, Step Function, and Lambdas are all provisioned using the SAM template sam-template.yml.

Testing with SAM Local

SAM Local is a CLI tool that allows you to invoke Lambdas locally. Here are some tests for these functions.

NOTE: You will need to change the contents of test-env.json and <lambda_dir>/test-event.json to point to data you own, or these commands won't work!

image-analysis:

sam local invoke ImageAnalysis -e image-analysis/test-event.json

image-api:

sam local start-api --env-vars test-env.json

This starts a local copy of the AWS API Gateway. It will show you what endpoints you can test.

image-db-update:

sam local invoke ImageDbUpdate -e image-db-update/test-event.json --env-vars ./test-env.json

image-pipeline-trigger:

sam local generate-event s3 | sam local invoke ImagePipelineTrigger --env-vars ./test-env.json

image-resize:

sam local invoke ImageResize -e image-resize/test-event.json --env-vars ./test-env.json