Build RAG purposes with MongoDB Atlas, now obtainable in Knowledge Bases for Amazon Bedrock


Voiced by Polly

Foundational fashions (FMs) are skilled on massive volumes of information and use billions of parameters. However, in an effort to reply clients’ questions associated to domain-specific non-public knowledge, they should reference an authoritative information base outdoors of the mannequin’s coaching knowledge sources. This is often achieved utilizing a way referred to as Retrieval Augmented Generation (RAG). By fetching knowledge from the group’s inner or proprietary sources, RAG extends the capabilities of FMs to particular domains, while not having to retrain the mannequin. It is a cheap strategy to bettering mannequin output so it stays related, correct, and helpful in varied contexts.

Knowledge Bases for Amazon Bedrock is a completely managed functionality that helps you implement your complete RAG workflow from ingestion to retrieval and immediate augmentation with out having to construct customized integrations to knowledge sources and handle knowledge flows.

Today, we’re saying the provision of MongoDB Atlas as a vector retailer in Knowledge Bases for Amazon Bedrock. With MongoDB Atlas vector retailer integration, you may construct RAG options to securely join your group’s non-public knowledge sources to FMs in Amazon Bedrock. This integration provides to the record of vector shops supported by Knowledge Bases for Amazon Bedrock, together with Amazon Aurora PostgreSQL-Compatible Edition, vector engine for Amazon OpenSearch Serverless, Pinecone, and Redis Enterprise Cloud.

Build RAG purposes with MongoDB Atlas and Knowledge Bases for Amazon Bedrock
Vector Search in MongoDB Atlas is powered by the vectorSearch index sort. In the index definition, you could specify the sector that incorporates the vector knowledge because the vector sort. Before utilizing MongoDB Atlas vector search in your utility, you will have to create an index, ingest supply knowledge, create vector embeddings and retailer them in a MongoDB Atlas assortment. To carry out queries, you will have to transform the enter textual content right into a vector embedding, after which use an aggregation pipeline stage to carry out vector search queries in opposition to fields listed because the vector sort in a vectorSearch sort index.

Thanks to the MongoDB Atlas integration with Knowledge Bases for Amazon Bedrock, many of the heavy lifting is taken care of. Once the vector search index and information base are configured, you may incorporate RAG into your purposes. Behind the scenes, Amazon Bedrock will convert your enter (immediate) into embeddings, question the information base, increase the FM immediate with the search outcomes as contextual data and return the generated response.

Let me stroll you thru the method of organising MongoDB Atlas as a vector retailer in Knowledge Bases for Amazon Bedrock.

Configure MongoDB Atlas
Start by making a MongoDB Atlas cluster on AWS. Choose an M10 devoted cluster tier. Once the cluster is provisioned, create a database and collection. Next, create a database consumer and grant it the Read and write to any database position. Select Password because the Authentication Method. Finally, configure community entry to switch the IP Access List – add IP deal with to permit entry from wherever.

Use the next index definition to create the Vector Search index:

  "fields": [
      "numDimensions": 1536,
      "similarity": "cosine",
      "type": "vector"
      "type": "filter"
      "type": "filter"

Configure the information base
Create an AWS Secrets Manager secret to securely retailer the MongoDB Atlas database consumer credentials. Choose Other because the Secret sort. Create an Amazon Simple Storage Service (Amazon S3) storage bucket and add the Amazon Bedrock documentation consumer information PDF. Later, you’ll use the information base to ask questions on Amazon Bedrock.

You also can use one other doc of your alternative as a result of Knowledge Base helps a number of file codecs (together with textual content, HTML, and CSV).

Navigate to the Amazon Bedrock console and confer with the Amzaon Bedrock User Guide to configure the information base. In the Select embeddings mannequin and configure vector retailer, select Titan Embeddings G1 – Text because the embedding mannequin. From the record of databases, select MongoDB Atlas.

Enter the essential data for the MongoDB Atlas cluster (Hostname, Database title, and many others.) in addition to the ARN of the AWS Secrets Manager secret you had created earlier. In the Metadata subject mapping attributes, enter the vector retailer particular particulars. They ought to match the vector search index definition you used earlier.

Initiate the information base creation. Once full, synchronise the information supply (S3 bucket knowledge) with the MongoDB Atlas vector search index.

Once the synchronization is full, navigate to MongoDB Atlas to verify that the information has been ingested into the gathering you created.

Notice the next attributes in every of the MongoDB Atlas paperwork:

  • AMAZON_BEDROCK_TEXT_CHUNK – Contains the uncooked textual content for every knowledge chunk.
  • AMAZON_BEDROCK_CHUNK_VECTOR – Contains the vector embedding for the information chunk.
  • AMAZON_BEDROCK_METADATA – Contains extra knowledge for supply attribution and wealthy question capabilities.

Test the information base
It’s time to ask questions on Amazon Bedrock by querying the information base. You might want to select a basis mannequin. I picked Claude v2 on this case and used “What is Amazon Bedrock” as my enter (question).

If you’re utilizing a distinct supply doc, alter the questions accordingly.

You also can change the muse mannequin. For instance, I switched to Claude 3 Sonnet. Notice the distinction within the output and choose Show supply particulars to see the chunks cited for every footnote.

Integrate information base with purposes
To construct RAG purposes on high of Knowledge Bases for Amazon Bedrock, you need to use the RetrieveAndGenerate API which lets you question the information base and get a response.

Here is an instance utilizing the AWS SDK for Python (Boto3):

import boto3

bedrock_agent_runtime = boto3.shopper(
    service_name = "bedrock-agent-runtime"

def retrieveAndGenerate(enter, kbId):
    return bedrock_agent_runtime.retrieve_and_generate(
            'textual content': enter
            'sort': 'KNOWLEDGE_BASE',
            'informationBaseConfiguration': {
                'informationBaseId': kbId,
                'modelArn': 'arn:aws:bedrock:us-east-1::foundation-model/anthropic.claude-3-sonnet-20240229-v1:0'

response = retrieveAndGenerate("What is Amazon Bedrock?", "BFT0P4NR1U")["output"]["text"]

If you wish to additional customise your RAG options, think about using the Retrieve API, which returns the semantic search responses that you need to use for the remaining a part of the RAG workflow.

import boto3

bedrock_agent_runtime = boto3.shopper(
    service_name = "bedrock-agent-runtime"

def retrieve(question, kbId, numberOfResults=5):
    return bedrock_agent_runtime.retrieve(
        retrievalQuery= {
            'textual content': question
        retrievalConfiguration= {
            'vectorSearchConfiguration': {
                'numberOfResults': numberOfResults

response = retrieve("What is Amazon Bedrock?", "BGU0Q4NU0U")["retrievalResults"]

Things to know

  • MongoDB Atlas cluster tier – This integration requires requires an Atlas cluster tier of at the least M10.
  • AWS PrivateLink – For the needs of this demo, MongoDB Atlas database IP Access List was configured to permit entry from wherever. For manufacturing deployments, AWS PrivateLink is the beneficial strategy to have Amazon Bedrock set up a safe connection to your MongoDB Atlas cluster. Refer to the Amazon Bedrock User information (below MongoDB Atlas) for particulars.
  • Vector embedding dimension – The dimension dimension of the vector index and the embedding mannequin must be the identical. For instance, in the event you plan to make use of Cohere Embed (which has a dimension dimension of 1024) because the embedding mannequin for the information base, be sure to configure the vector search index accordingly.
  • Metadata filters – You can add metadata in your supply information to retrieve a well-defined subset of the semantically related chunks primarily based on utilized metadata filters. Refer to the documentation to study extra about how you can use metadata filters.

Now obtainable
MongoDB Atlas vector retailer in Knowledge Bases for Amazon Bedrock is accessible within the US East (N. Virginia) and US West (Oregon) Regions. Be positive to verify the full Region record for future updates.

Learn extra

Try out the MongoDB Atlas integration with Knowledge Bases for Amazon Bedrock! Send suggestions to AWS re:Post for Amazon Bedrock or by means of your regular AWS contacts and have interaction with the generative AI builder neighborhood at



Please enter your comment!
Please enter your name here