S3 Vector, Boto3 Error in Lambda

This error occurs because your Lambda function is using a version of boto3 that doesn't include support for the S3Vectors service, which is relatively new.

To resolve this issue, you need to ensure your Lambda function is using a recent version of boto3 that includes S3Vectors support. Here are some steps to fix it:

1. Update the boto3 library in your Lambda function. You can do this by including the latest boto3 version in your deployment package or by using Lambda Layers with an updated boto3.

2. Make sure you're using the correct service name. The service name is 's3vectors' (all lowercase), which you're already using correctly.

3. Verify that you're using a region where S3Vectors is available. S3Vectors is available in regions like US East (N. Virginia), US East (Ohio), US West (Oregon), Europe (Frankfurt), and Asia Pacific (Sydney). Your code is using 'us-west-2' which is US West (Oregon), so that should be fine.

4. Ensure your Lambda execution role has the necessary permissions to interact with S3Vectors.

Here's an example of how to properly initialize and use the S3Vectors client:

import boto3

# Create S3Vectors client
s3vectors = boto3.client('s3vectors', region_name='us-west-2')

# Example: Query vectors
response = s3vectors.query_vectors(
vectorBucketName="your-vector-bucket",
indexName="your-vector-index",
queryVector={"float32": your_embedding},
topK=3,
returnDistance=True,
returnMetadata=True
)

If you're working with embeddings, you might also need to use Amazon Bedrock to generate them before querying S3Vectors, as shown in many examples.
Sources
S3Vectors - Boto3 1.39.5 documentation
Querying vectors - Amazon Simple Storage Service
Introducing Amazon S3 Vectors: First cloud storage with native vector support at scale (preview) | AWS News Blog