IPFS + Amazon Rekognition Image Moderation External Adapter

This external adapter downloads image bytes from IPFS given the content hash and requests image moderation labels from the Amazon Rekognition cloud-based computer vision platform. The moderation labels are pinned to IPFS using Pinata and a bytes32 hex string is returned from which the base58 encoded IPFS content hash can be reconstructed. Currently bytes32 is the maximum response size for the result of a Chainlink job.

Environment variables

You will need an AWS account to be able to make requests to Amazon Rekognition. See Amazon's documentation for help on setting up an account.

Variable		Description	Example
AWS_ACCESS_KEY_ID	Required	The ID of your AWS access key	ABCDEFGABCDEFG
AWS_SECRET_ACCESS_KEY	Required	Your AWS secret access key	AbCdEfGaBcDeFgAbCdEfGaBcDeFg
AWS_REGION	Required	The AWS region you would like to use	eu-west-2
PINATA_API_KEY	Required	Your Pinata API key	0298f5aa704402e9a63a
PINATA_SECRET_API_KEY	Required	Your Pinata secret API key	91f9d07d04da2970a43e0ecf44f9fe64504a064e0537c031668c91399bcc4392

Example request

Input

• hash: The image's IPFS content hash

Output

{
  "jobRunID": "1",
  "data": {
    "ModerationLabels": [
      {
        "Confidence": 99.92990112304688,
        "Name": "Middle Finger",
        "ParentName": "Rude Gestures"
      },
      {
        "Confidence": 99.92990112304688,
        "Name": "Rude Gestures",
        "ParentName": ""
      }
    ],
    "ModerationModelVersion": "4.0",
    "result": "0xafacbc7a25ed1d9a9c4af37bc972fb2b599f5cc6496fe84ffd5417ff34e3bb62"
  },
  "result": "0xafacbc7a25ed1d9a9c4af37bc972fb2b599f5cc6496fe84ffd5417ff34e3bb62"
}

Install Locally

Install dependencies:

npm i

Test

Run the local tests:

npm run test

Natively run the application (defaults to port 8080):

Run

npm start

Call the external adapter/API server

curl -X POST -H "content-type:application/json" "http://localhost:8080/" --data '{ "id": 0, "data": { "hash": "QmdT7hKV1EfuaXSAYa65KUZWJnxF96yRPZNS9WeG8gUsR2" } }'

Docker

If you wish to use Docker to run the adapter, you can build the image by running the following command:

docker build . -t external-adapter

Then run it with:

docker run -p 8080:8080 -it external-adapter:latest

Serverless hosts

After installing locally:

Create the zip

zip -r external-adapter.zip .

Install to AWS Lambda

• In Lambda Functions, create function
• On the Create function page:
  • Give the function a name
  • Use Node.js 12.x for the runtime
  • Choose an existing role or create a new one
  • Click Create Function
• Under Function code, select "Upload a .zip file" from the Code entry type drop-down
• Click Upload and select the external-adapter.zip file
• Handler:
  • index.handler for REST API Gateways
  • index.handlerv2 for HTTP API Gateways
• Add the environment variable (repeat for all environment variables):
  • Key: API_KEY
  • Value: Your_API_key
• Save

To Set Up an API Gateway (HTTP API)

If using a HTTP API Gateway, Lambda's built-in Test will fail, but you will be able to externally call the function successfully.

• Click Add Trigger
• Select API Gateway in Trigger configuration
• Under API, click Create an API
• Choose HTTP API
• Select the security for the API
• Click Add

To Set Up an API Gateway (REST API)

If using a REST API Gateway, you will need to disable the Lambda proxy integration for Lambda-based adapter to function.

• Click Add Trigger
• Select API Gateway in Trigger configuration
• Under API, click Create an API
• Choose REST API
• Select the security for the API
• Click Add
• Click the API Gateway trigger
• Click the name of the trigger (this is a link, a new window opens)
• Click Integration Request
• Uncheck Use Lamba Proxy integration
• Click OK on the two dialogs
• Return to your function
• Remove the API Gateway and Save
• Click Add Trigger and use the same API Gateway
• Select the deployment stage and security
• Click Add

Install to GCP

• In Functions, create a new function, choose to ZIP upload
• Click Browse and select the external-adapter.zip file
• Select a Storage Bucket to keep the zip in
• Function to execute: gcpservice
• Click More, Add variable (repeat for all environment variables)
  • NAME: API_KEY
  • VALUE: Your_API_key