feat(sagemaker models): use Generative AI CDK constructs for SageMaker model deployment

[krokoko]

Description of changes:

• Upgrade CDK to minimum version required by the new dependency
• Add Generative AI CDK constructs as a dependency
• Update lib/models/index.ts to consume the constructs

How was this tested ?

Used npm create to enable all models available from the list and run a request for each one of them through the UI, see images below:

Notes:

I wasn't able at first to test the IDEFICS models as I was facing an error when running cdk deploy with those models enabled:

Resource handler returned message: "CloudWatch Logs role ARN must be set in account settings to enable logging (Service: ApiGateway, Status Code: 400, Request ID: bbfb4458-
d5c2-4184-9a8a-0ea549b52d47)" (RequestToken: 56d1d88b-b812-701d-c11b-98765a7f5c14, HandlerErrorCode: InvalidRequest)

I Had to create an IAM role with apigateway.amazonaws.com as its trusted entity, attach the managed policy AmazonAPIGatewayPushToCloudWatchLogs to the IAM role, and set the IAM role ARN on the cloudWatchRoleArn property of my account. Haven't seen that step mentioned in the documentation.

Questions:

@bigadsoleiman @spugachev

1. I didn't remove here the code related to the original constructs, since the custom inference script feature is not available yet in the Generative AI CDK constructs. Is that fine with you ?
2. Should we also remove the code in scripts/get-model-packages ? Seems this is not necessary anymore
3. Do you want me to update the documentation as well (sagemaker-model/README.md) to reflect the changes above ?
4. Anything else I may have missed ?

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[bigadsoleiman]

Thanks @krokoko

1. Yes
2. Yes
3. Yes please but then move it to https://github.com/aws-samples/aws-genai-llm-chatbot/tree/main/docs you can add a dedicated page for model constructs if you'd like.
4. No