Skip to main content
For the complete documentation index for agents and LLMs, see llms.txt.

AmazonBedrockRanker

Rank documents based on their similarity to the query using models hosted on Amazon Bedrock.

Amazon Bedrock is a fully managed service that makes state-of-the-art language models available for use through a unified API. To learn more, see Amazon Bedrock documentation.

Key Features​

  • Ranks documents by semantic similarity to the query.
  • Returns documents in descending order of relevance score.
  • Supports the following ranking models: cohere.rerank-v3-5:0 and amazon.rerank-v1:0.
  • Lets you limit the number of returned documents with top_k.
  • Can embed document metadata alongside document content for richer ranking.

Configuration​

To use this component, connect Haystack Platform with Amazon Bedrock first. You'll need:

  • The region name
  • Access key ID
  • Secret access key

For detailed explanation, see Use Amazon Bedrock and SageMaker Models.

:::

  1. Drag the AmazonBedrockRanker component onto the canvas from the Component Library.
  2. Click the component to open the configuration panel.
  3. On the General tab:
    1. Select the ranking model from the list.
  4. Go to the Advanced tab to configure the AWS credentials, number of top results, maximum chunks per document, metadata fields to embed, and metadata separator.

Connections​

AmazonBedrockRanker accepts a query string, a list of documents to rank, and an optional top_k value as inputs. It outputs a list of documents sorted by relevance to the query.

Connect a retriever's documents output to the documents input. Connect the pipeline's query input to the query input. Connect the documents output to PromptBuilder or another component that processes ranked documents.

  1. Drag the AmazonBedrockRanker component onto the canvas from the Component Library.
  2. Click on the component to open the configuration panel.
  3. On the General tab:
    • Select the ranking model.
    • Set top_k to limit the number of documents to return.
  4. Go to the Advanced tab to configure additional settings, such as max_chunks_per_doc, meta_fields_to_embed, and meta_data_separator.

Source Code​

To check this component's source code, open ranker.py in the Haystack Core Integrations repository.

Connections​

AmazonBedrockRanker receives a query string and a list of documents to rank. Connect a retriever's documents output to its documents input, and connect the pipeline query input to its query input.

It outputs the ranked documents in descending order of relevance. Connect its documents output to a prompt builder or directly to the pipeline output.

Usage Examples​

Basic Configuration​

  AmazonBedrockRanker:
    type: haystack_integrations.components.rankers.amazon_bedrock.ranker.AmazonBedrockRanker
    init_parameters:
      model: cohere.rerank-v3-5:0
      top_k: 10
      aws_access_key_id:
        type: env_var
        env_vars:
        - AWS_ACCESS_KEY_ID
        strict: false
      aws_secret_access_key:
        type: env_var
        env_vars:
        - AWS_SECRET_ACCESS_KEY
        strict: false
      aws_session_token:
        type: env_var
        env_vars:
        - AWS_SESSION_TOKEN
        strict: false
      aws_region_name:
        type: env_var
        env_vars:
        - AWS_DEFAULT_REGION
        strict: false
      aws_profile_name:
        type: env_var
        env_vars:
        - AWS_PROFILE
        strict: false
      meta_data_separator: \n

Using the Component in a Pipeline​

This is an example of a document search pipeline that uses AmazonBedrockRanker with the cohere ranking model:

components:
  bm25_retriever: # Selects the most similar documents from the document store
    type: haystack_integrations.components.retrievers.opensearch.bm25_retriever.OpenSearchBM25Retriever
    init_parameters:
      document_store:
        type: haystack_integrations.document_stores.opensearch.document_store.OpenSearchDocumentStore
        init_parameters:
          hosts:
          index: Standard-Index-English
          max_chunk_bytes: 104857600
          embedding_dim: 768
          return_embedding: false
          method:
          mappings:
          settings:
          create_index: true
          http_auth:
          use_ssl:
          verify_certs:
          timeout:
      top_k: 20 # The number of results to return
      fuzziness: 0

  query_embedder:
    type: deepset_cloud_custom_nodes.embedders.nvidia.text_embedder.DeepsetNvidiaTextEmbedder
    init_parameters:
      normalize_embeddings: true
      model: intfloat/e5-base-v2

  embedding_retriever: # Selects the most similar documents from the document store
    type: haystack_integrations.components.retrievers.opensearch.embedding_retriever.OpenSearchEmbeddingRetriever
    init_parameters:
      document_store:
        type: haystack_integrations.document_stores.opensearch.document_store.OpenSearchDocumentStore
        init_parameters:
          hosts:
          index: Standard-Index-English
          max_chunk_bytes: 104857600
          embedding_dim: 768
          return_embedding: false
          method:
          mappings:
          settings:
          create_index: true
          http_auth:
          use_ssl:
          verify_certs:
          timeout:
      top_k: 20 # The number of results to return

  document_joiner:
    type: haystack.components.joiners.document_joiner.DocumentJoiner
    init_parameters:
      join_mode: concatenate

  AmazonBedrockRanker:
    type: haystack_integrations.components.rankers.amazon_bedrock.ranker.AmazonBedrockRanker
    init_parameters:
      model: cohere.rerank-v3-5:0
      top_k: 10
      aws_access_key_id:
        type: env_var
        env_vars:
        - AWS_ACCESS_KEY_ID
        strict: false
      aws_secret_access_key:
        type: env_var
        env_vars:
        - AWS_SECRET_ACCESS_KEY
        strict: false
      aws_session_token:
        type: env_var
        env_vars:
        - AWS_SESSION_TOKEN
        strict: false
      aws_region_name:
        type: env_var
        env_vars:
        - AWS_DEFAULT_REGION
        strict: false
      aws_profile_name:
        type: env_var
        env_vars:
        - AWS_PROFILE
        strict: false
      max_chunks_per_doc:
      meta_fields_to_embed:
      meta_data_separator: \n

connections:  # Defines how the components are connected
- sender: bm25_retriever.documents
  receiver: document_joiner.documents
- sender: query_embedder.embedding
  receiver: embedding_retriever.query_embedding
- sender: embedding_retriever.documents
  receiver: document_joiner.documents
- sender: document_joiner.documents
  receiver: AmazonBedrockRanker.documents

inputs:  # Define the inputs for your pipeline
  query:  # These components will receive the query as input
  - "bm25_retriever.query"
  - "query_embedder.text"
  - "AmazonBedrockRanker.query"
  filters:  # These components will receive a potential query filter as input
  - "bm25_retriever.filters"
  - "embedding_retriever.filters"

outputs:  # Defines the output of your pipeline
  documents: "AmazonBedrockRanker.documents" # The output of the pipeline is the retrieved documents

max_runs_per_component: 100

metadata: {}

Parameters​

Inputs​

ParameterTypeDefaultDescription
querystrThe query used for ranking documents by their similarity to the query.
documentsList[Document]The documents to be ranked.
top_kOptional[int]NoneThe maximum number of documents you want the Ranker to return.

Outputs​

ParameterTypeDescription
documentsList[Document]Documents most similar to the query in descending order of similarity.

Init Parameters​

These are the parameters you can configure in Pipeline Builder:

ParameterTypeDefaultDescription
modelstrcohere.rerank-v3-5:0The ranking model to use.
top_kint10The maximum number of documents to return.
aws_access_key_idOptional[Secret]Secret.from_env_var(["AWS_ACCESS_KEY_ID"], strict=False)AWS access key ID.
aws_secret_access_keyOptional[Secret]Secret.from_env_var(["AWS_SECRET_ACCESS_KEY"], strict=False)AWS secret access key.
aws_session_tokenOptional[Secret]Secret.from_env_var([AWS_SESSION_TOKEN], strict=False)AWS session token.
aws_region_nameOptional[Secret]Secret.from_env_var(["AWS_DEFAULT_REGION"], strict=False)AWS region name.
aws_profile_nameOptional[Secret]Secret.from_env_var(["AWS_PROFILE"], strict=False)AWS profile name.
max_chunks_per_docOptional[int]NoneIf your document exceeds 512 tokens, this setting determines the maximum number of chunks a document can be split into. If set to None, uses the default of 10 chunks. This parameter is not used currently but it's included for future compatibility.
meta_fields_to_embedOptional[List[str]]NoneA list of metadata fields to embed in the document content.
meta_data_separatorstr\nThe separator used to concatenate the metadata fields to the document content.

Run Method Parameters​

These are the parameters you can configure for the component's run() method. This means you can pass these parameters at query time through the API, in Playground, or when running a job. For details, see Modify Pipeline Parameters at Query Time.

ParameterTypeDefaultDescription
querystrThe user query for ranking the documents.
documentsList[Document]The documents to rank.
top_kOptional[int]NoneThe maximum number of documents you want the Ranker to return.