This package contains the LangChain integrations with AWS.
pip install -U langchain-awsAll integrations in this package assume that you have the credentials setup to connect with AWS services.
ChatBedrock class exposes chat models from Bedrock.
from langchain_aws import ChatBedrock
llm = ChatBedrock()
llm.invoke("Sing a ballad of LangChain.")BedrockEmbeddings class exposes embeddings from Bedrock.
from langchain_aws import BedrockEmbeddings
embeddings = BedrockEmbeddings()
embeddings.embed_query("What is the meaning of life?")BedrockLLM class exposes LLMs from Bedrock.
from langchain_aws import BedrockLLM
llm = BedrockLLM()
llm.invoke("The meaning of life is")AmazonKendraRetriever class provides a retriever to connect with Amazon Kendra.
from langchain_aws import AmazonKendraRetriever
retriever = AmazonKendraRetriever(
index_id="561be2b6d-9804c7e7-f6a0fbb8-5ccd350"
)
retriever.get_relevant_documents(query="What is the meaning of life?")AmazonKnowledgeBasesRetriever class provides a retriever to connect with Amazon Knowledge Bases.
from langchain_aws import AmazonKnowledgeBasesRetriever
retriever = AmazonKnowledgeBasesRetriever(
knowledge_base_id="IAPJ4QPUEU",
retrieval_config={"vectorSearchConfiguration": {"numberOfResults": 4}},
)
retriever.get_relevant_documents(query="What is the meaning of life?")InMemoryVectorStore class provides a vectorstore to connect with Amazon MemoryDB.
from langchain_aws.vectorstores.inmemorydb import InMemoryVectorStore
vds = InMemoryVectorStore.from_documents(
chunks,
embeddings,
redis_url="rediss://cluster_endpoint:6379/ssl=True ssl_cert_reqs=none",
vector_schema=vector_schema,
index_name=INDEX_NAME,
)Here we go over different options for using the vector store as a retriever.
There are three different search methods we can use to do retrieval. By default, it will use semantic similarity.
retriever=vds.as_retriever()