Labelbox / labelbox-aws-stream

This repo will demonstrate how to connect a AWS bucket with a Labelbox Dataset using AWS Lambda Functions.

Data Streaming with AWS using AWS Lambda Functions

This repo will provide with 3 deployable Lambda Functions:

• stream_data_rows - This will create a Labelbox data row every time a new asset is uploaded to the AWS bucket
• update_metadata - This will update Labelbox Metadata for existing data rows any time an object's metadata is modified in AWS
• delete_data_rows - This will delete a data row every time an asset is deleted in the AWS bucket

Prerequisites:

1. Set up Delegated Access to Labelbox, and ensure that the default setting is to the target bucket.
   • See how here: https://docs.labelbox.com/docs/import-aws-s3-data
2. Only files supported by Labelbox will be uploaded
   • See latest supported data types here: https://docs.labelbox.com/docs/supported-data-types
3. The AWS account performing these steps has access to the AWS bucket and permissions to create Lambda functions
4. The Labelbox account performing these steps has Admin-level permissions in a Labelbox organization

How to Use:

1. Authenticate your AWS CLI in your Terminal:

3. Create / Redeploy your AWS function from GitHub:

4. Clone Repo and Define Variables:

git clone https://github.com/Labelbox/labelbox-aws-stream.git
cd labelbox-aws-stream
AWS_BUCKET_NAME="my-bucket"
LABELBOX_API_KEY="my-api-key"
LABELBOX_INTEGRATION_NAME="my-integration-name"

Deploy stream_data_rows:

Deploy update_metadata:

Deploy delete_data_rows: