langchain/docs/versioned_docs/version-0.2.x/integrations/retrievers/bedrock.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "b6636c27-35da-4ba7-8313-eca21660cab3",
   "metadata": {},
   "source": [
    "# Bedrock (Knowledge Bases)\n",
    "\n",
    "> [Knowledge bases for Amazon Bedrock](https://aws.amazon.com/bedrock/knowledge-bases/) is an Amazon Web Services (AWS) offering which lets you quickly build RAG applications by using your private data to customize FM response.\n",
    "\n",
    "> Implementing `RAG` requires organizations to perform several cumbersome steps to convert data into embeddings (vectors), store the embeddings in a specialized vector database, and build custom integrations into the database to search and retrieve text relevant to the user’s query. This can be time-consuming and inefficient.\n",
    "\n",
    "> With `Knowledge Bases for Amazon Bedrock`, simply point to the location of your data in `Amazon S3`, and `Knowledge Bases for Amazon Bedrock` takes care of the entire ingestion workflow into your vector database. If you do not have an existing vector database, Amazon Bedrock creates an Amazon OpenSearch Serverless vector store for you. For retrievals, use the Langchain - Amazon Bedrock integration via the Retrieve API to retrieve relevant results for a user query from knowledge bases.\n",
    "\n",
    "> Knowledge base can be configured through [AWS Console](https://aws.amazon.com/console/) or by using [AWS SDKs](https://aws.amazon.com/developer/tools/)."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b34c8cbe-c6e5-4398-adf1-4925204bcaed",
   "metadata": {},
   "source": [
    "## Using the Knowledge Bases Retriever"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "26c97d36-911c-4fe0-a478-546192728f30",
   "metadata": {},
   "outputs": [],
   "source": [
    "%pip install --upgrade --quiet  boto3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "30337664-8844-4dfe-97db-077abb51af68",
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain_community.retrievers import AmazonKnowledgeBasesRetriever\n",
    "\n",
    "retriever = AmazonKnowledgeBasesRetriever(\n",
    "    knowledge_base_id=\"PUIJP4EQUA\",\n",
    "    retrieval_config={\"vectorSearchConfiguration\": {\"numberOfResults\": 4}},\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f9fefa50-f0fb-40e3-b4e4-67c5b232a090",
   "metadata": {},
   "outputs": [],
   "source": [
    "query = \"What did the president say about Ketanji Brown?\"\n",
    "\n",
    "retriever.get_relevant_documents(query=query)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7de9b61b-597b-4aba-95fb-49d11e84510e",
   "metadata": {},
   "source": [
    "### Using in a QA Chain"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0fd71709-aaed-42b5-a990-e3067bfa7143",
   "metadata": {},
   "outputs": [],
   "source": [
    "from botocore.client import Config\n",
    "from langchain.chains import RetrievalQA\n",
    "from langchain_community.llms import Bedrock\n",
    "\n",
    "model_kwargs_claude = {\"temperature\": 0, \"top_k\": 10, \"max_tokens_to_sample\": 3000}\n",
    "\n",
    "llm = Bedrock(model_id=\"anthropic.claude-v2\", model_kwargs=model_kwargs_claude)\n",
    "\n",
    "qa = RetrievalQA.from_chain_type(\n",
    "    llm=llm, retriever=retriever, return_source_documents=True\n",
    ")\n",
    "\n",
    "qa(query)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}