Merge branch 'main' into feat/automate_model_gen

.projenrc.ts

@@ -40,7 +40,7 @@ const project = new awscdk.AwsCdkConstructLibrary({
   projenVersion: '~0.78.8',
   constructsVersion: '10.3.0',
   defaultReleaseBranch: 'main',
-  jsiiVersion: '~5.1.0',
+  jsiiVersion: '~5.3.0',
   name: '@' + PUBLICATION_NAMESPACE + '/' + PROJECT_NAME,
   projenrcTs: true,
   repositoryUrl: 'https://github.com/' + GITHUB_USER + '/' + PROJECT_NAME,

src/patterns/gen-ai/aws-langchain-common-layer/README.md

@@ -15,6 +15,7 @@
 | **Language**     | **Package**        |
 |:-------------|-----------------|
 |![Typescript Logo](https://docs.aws.amazon.com/cdk/api/latest/img/typescript32.png) Typescript|`@cdklabs/generative-ai-cdk-constructs`|
+|![Python Logo](https://docs.aws.amazon.com/cdk/api/latest/img/python32.png) Python|`cdklabs.generative_ai_cdk_constructs`|
 
 ## Table of contents
 
@@ -79,6 +80,48 @@ const lambdaCommonLayer = new LangchainCommonLayer(this, 'lambdagenaicommonlayer
 //Then pass the layers above to your lambda function constructor
 ```
 
+Python
+``` python
+from constructs import Construct
+from aws_cdk import Aws, aws_lambda as lambda_
+from cdklabs.generative_ai_cdk_constructs import (
+    LangchainCommonDepsLayer,
+    LangchainCommonLayer,
+)
+
+lambda_architecture = lambda_.Architecture.ARM_64
+lambda_runtime = lambda_.Runtime.PYTHON_3_10
+
+# This is one way of getting a lambda powertools layer
+powertools_arn = (
+    f'arn:aws:lambda:{Aws.REGION}:017000801446:'
+    'layer:AWSLambdaPowertoolsPythonV2-Arm64:42'
+)
+
+if lambda_architecture == lambda_.Architecture.X86_64:
+    powertools_arn = (
+        f'arn:aws:lambda:{Aws.REGION}:017000801446:'
+        'layer:AWSLambdaPowertoolsPythonV2:42'
+    )
+
+lambda_deps_layer = LangchainCommonDepsLayer(
+    self,
+    'lambdagenaidepslayer',
+    runtime=lambda_runtime,
+    architecture=lambda_architecture,
+    auto_upgrade=True,
+)
+
+lambda_common_layer = LangchainCommonLayer(
+    self,
+    'lambdagenaicommonlayer',
+    compatible_runtimes=[lambda_runtime],
+    compatible_architectures=[lambda_architecture],
+)
+
+# Then pass the layers above to your lambda function constructor
+```
+
 Here is an example of a Lambda handler which uses the layers above:
 
 ```

src/patterns/gen-ai/aws-model-deployment-sagemaker/README_custom_sagemaker_endpoint.md

@@ -16,6 +16,7 @@
 | **Language**     | **Package**        |
 |:-------------|-----------------|
 |![Typescript Logo](https://docs.aws.amazon.com/cdk/api/latest/img/typescript32.png) Typescript|`@cdklabs/generative-ai-cdk-constructs`|
+|![Python Logo](https://docs.aws.amazon.com/cdk/api/latest/img/python32.png) Python|`cdklabs.generative_ai_cdk_constructs`|
 
 ## Table of contents
 
@@ -45,6 +46,7 @@ env: {
 
 Here is a minimal deployable pattern definition:
 
+Typescript
 ```typescript
 
 import { Construct } from 'constructs';
@@ -64,6 +66,35 @@ new CustomSageMakerEndpoint(this, 'customModel', {
   });
 ```
 
+Python
+```python
+from constructs import Construct
+from cdklabs.generative_ai_cdk_constructs import (
+    CustomSageMakerEndpoint,
+    DeepLearningContainerImage,
+    SageMakerInstanceType,
+)
+
+# Need an existing bucket containing model artifacts that this construct can access
+bucket_name = 'bucket'
+key_name = 'key'
+
+CustomSageMakerEndpoint(
+    self,
+    'customModel',
+    model_id='bgeinf2',
+    instance_type=SageMakerInstanceType.ML_INF2_XLARGE,
+    container=DeepLearningContainerImage.from_deep_learning_container_image(
+        'huggingface-pytorch-inference-neuronx',
+        '1.13.1-transformers4.34.1-neuronx-py310-sdk2.15.0-ubuntu20.04',
+    ),
+    model_data_url=f's3://{bucket_name}/{key_name}/model.tar.gz',
+    endpoint_name='testbgebase',
+    instance_count=1,
+    volume_size_in_gb=100,
+)
+```
+
 ## Initializer
 
 ```

src/patterns/gen-ai/aws-model-deployment-sagemaker/README_hugging_face.md

@@ -16,6 +16,7 @@
 | **Language**     | **Package**        |
 |:-------------|-----------------|
 |![Typescript Logo](https://docs.aws.amazon.com/cdk/api/latest/img/typescript32.png) Typescript|`@cdklabs/generative-ai-cdk-constructs`|
+|![Python Logo](https://docs.aws.amazon.com/cdk/api/latest/img/python32.png) Python|`cdklabs.generative_ai_cdk_constructs`|
 
 ## Table of contents
 
@@ -45,6 +46,7 @@ env: {
 
 Here is a minimal deployable pattern definition:
 
+Typescript
 ```typescript
 
 import { Construct } from 'constructs';
@@ -65,6 +67,29 @@ new HuggingFaceSageMakerEndpoint(this, 'Mistral', {
 });
 ```
 
+Python
+```python
+from constructs import Construct
+from cdklabs.generative_ai_cdk_constructs import (
+    HuggingFaceSageMakerEndpoint,
+    DeepLearningContainerImage,
+    SageMakerInstanceType,
+)
+
+HuggingFaceSageMakerEndpoint(
+    self,
+    'Mistral',
+    model_id='mistralai/Mistral-7B-Instruct-v0.1',
+    instance_type=SageMakerInstanceType.ML_G5_2_XLARGE,
+    container=DeepLearningContainerImage.HUGGINGFACE_PYTORCH_TGI_INFERENCE_2_0_1_TGI1_1_0_GPU_PY39_CU118_UBUNTU20_04,
+    environment={
+        'SM_NUM_GPUS': '1',
+        'MAX_INPUT_LENGTH': '2048',
+        'MAX_TOTAL_TOKENS': '4096',
+    },
+)
+```
+
 ## Initializer
 
 ```

src/patterns/gen-ai/aws-model-deployment-sagemaker/README_jumpstart.md

@@ -16,6 +16,7 @@
 | **Language**     | **Package**        |
 |:-------------|-----------------|
 |![Typescript Logo](https://docs.aws.amazon.com/cdk/api/latest/img/typescript32.png) Typescript|`@cdklabs/generative-ai-cdk-constructs`|
+|![Python Logo](https://docs.aws.amazon.com/cdk/api/latest/img/python32.png) Python|`cdklabs.generative_ai_cdk_constructs`|
 
 ## Table of contents
 
@@ -45,6 +46,7 @@ env: {
 
 Here is a minimal deployable pattern definition:
 
+Typescript
 ```typescript
 
 import { Construct } from 'constructs';
@@ -58,6 +60,23 @@ new JumpStartSageMakerEndpoint(this, 'LLAMA2', {
 });
 ```
 
+Python
+```python
+from constructs import Construct
+from cdklabs.generative_ai_cdk_constructs import (
+    JumpStartSageMakerEndpoint,
+    JumpStartModel,
+    SageMakerInstanceType,
+)
+
+JumpStartSageMakerEndpoint(
+    self,
+    'LLAMA2',
+    model=JumpStartModel.META_TEXTGENERATION_LLAMA_2_7_B_F_2_0_2,
+    instance_type=SageMakerInstanceType.ML_G5_2_XLARGE,
+)
+```
+
 ## Initializer
 
 ```

src/patterns/gen-ai/aws-qa-appsync-opensearch/README.md

@@ -15,6 +15,7 @@
 | **Language**     | **Package**        |
 |:-------------|-----------------|
 |![Typescript Logo](https://docs.aws.amazon.com/cdk/api/latest/img/typescript32.png) Typescript|`@cdklabs/generative-ai-cdk-constructs`|
+|![Python Logo](https://docs.aws.amazon.com/cdk/api/latest/img/python32.png) Python|`cdklabs.generative_ai_cdk_constructs`|
 
 ## Table of contents
 
@@ -82,6 +83,41 @@ const rag_source = new QaAppsyncOpensearch(
     )
 ```
 
+Python
+
+``` python
+from constructs import Construct
+from aws_cdk import (
+    aws_opensearchservice as os,
+    aws_cognito as cognito,
+)
+from cdklabs.generative_ai_cdk_constructs import QaAppsyncOpensearch
+
+# get an existing OpenSearch provisioned cluster
+os_domain = os.Domain.from_domain_attributes(
+    self, 
+    'osdomain',
+    domain_arn='arn:aws:es:us-east-1:XXXXXX:resource-id',
+    domain_endpoint='https://XXXXX.us-east-1.es.amazonaws.com',
+)
+
+# get an existing userpool 
+cognito_pool_id = 'us-east-1_XXXXX';
+user_pool_loaded = cognito.UserPool.from_user_pool_id(
+    self,
+    'myuserpool',
+    user_pool_id=cognito_pool_id,
+)
+
+rag_source = QaAppsyncOpensearch(
+    self,
+    'QaAppsyncOpensearch',
+    existing_opensearch_domain=os_domain,
+    open_search_index_name='demoindex',
+    cognito_user_pool=user_pool_loaded,
+)
+```
+
 After deploying the CDK stack, the QA process can be invoked using GraphQL APIs. The API Schema details are present here: resources/gen-ai/aws-qa-appsync-opensearch/schema.graphql.
 
 The code below provides an example of a mutation call and associated subscription to trigger a question and get response notifications. The subscription call will wait for mutation requests to send the notifications.

src/patterns/gen-ai/aws-qa-appsync-opensearch/index.ts

@@ -28,7 +28,7 @@ import { NagSuppressions } from 'cdk-nag';
 import { Construct } from 'constructs';
 import * as opensearch_helper from '../../../common/helpers/opensearch-helper';
 import * as s3_bucket_helper from '../../../common/helpers/s3-bucket-helper';
-import { version } from '../../../common/helpers/utils';
+import { version, lambdaMemorySizeLimiter } from '../../../common/helpers/utils';
 import * as vpc_helper from '../../../common/helpers/vpc-helper';
 
 /**
@@ -533,7 +533,7 @@ export class QaAppsyncOpensearch extends Construct {
         tracing: lambda_tracing,
         vpcSubnets: { subnetType: ec2.SubnetType.PRIVATE_WITH_EGRESS },
         securityGroups: [this.securityGroup],
-        memorySize: 1_769 * 4,
+        memorySize: lambdaMemorySizeLimiter(this, 1_769 * 4),
         timeout: Duration.minutes(15),
         role: question_answering_function_role,
         environment: {

src/patterns/gen-ai/aws-rag-appsync-stepfn-opensearch/README.md

@@ -15,6 +15,7 @@
 | **Language**     | **Package**        |
 |:-------------|-----------------|
 |![Typescript Logo](https://docs.aws.amazon.com/cdk/api/latest/img/typescript32.png) Typescript|`@cdklabs/generative-ai-cdk-constructs`|
+|![Python Logo](https://docs.aws.amazon.com/cdk/api/latest/img/python32.png) Python|`cdklabs.generative_ai_cdk_constructs`|
 
 ## Table of contents
 
@@ -87,6 +88,42 @@ const rag_source = new RagAppsyncStepfnOpensearch(
       }
     )
 ```
+
+Python
+``` python
+from constructs import Construct
+from aws_cdk import (
+    aws_opensearchservice as os,
+    aws_cognito as cognito,
+)
+from cdklabs.generative_ai_cdk_constructs import RagAppsyncStepfnOpensearch
+
+# get an existing OpenSearch provisioned cluster in the same VPC as of RagAppsyncStepfnOpensearch construct 
+# Security group for the existing opensearch cluster should allow traffic on 443.
+os_domain = os.Domain.from_domain_attributes(
+    self,
+    'osdomain',
+    domain_arn='arn:aws:es:us-east-1:XXXXXX:resource-id',
+    domain_endpoint='https://XXXXX.us-east-1.es.amazonaws.com',
+)
+
+# get an existing userpool 
+cognito_pool_id = 'us-east-1_XXXXX';
+user_pool_loaded = cognito.UserPool.from_user_pool_id(
+    self,
+    'myuserpool',
+    user_pool_id=cognito_pool_id,
+)
+
+rag_source = RagAppsyncStepfnOpensearch(
+    self,
+    'RagAppsyncStepfnOpensearch',
+    existing_opensearch_domain=os_domain,
+    open_search_index_name='demoindex',
+    cognito_user_pool=user_pool_loaded,
+)
+```
+
 After deploying the CDK stack, the document summarization workflow can be invoked using GraphQL APIs. The API schema details are here: resources/gen-ai/aws-rag-appsync-stepfn-opensearch/schema.graphql.
 
 The code below provides an example of a mutation call and associated subscription to trigger a pipeline call and get status notifications: