mirror of https://github.com/hwchase17/langchain
AWS Bedrock RAG template (#12450)
Lance Martin
7 months ago
committed by
GitHub
parent
5d40e36c75
commit
5c564e62e1
@ -0,0 +1,21 @@
|
||||
MIT License
|
||||
|
||||
Copyright (c) 2023 LangChain, Inc.
|
||||
|
||||
Permission is hereby granted, free of charge, to any person obtaining a copy
|
||||
of this software and associated documentation files (the "Software"), to deal
|
||||
in the Software without restriction, including without limitation the rights
|
||||
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
||||
copies of the Software, and to permit persons to whom the Software is
|
||||
furnished to do so, subject to the following conditions:
|
||||
|
||||
The above copyright notice and this permission notice shall be included in all
|
||||
copies or substantial portions of the Software.
|
||||
|
||||
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
||||
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
||||
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
||||
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
||||
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
||||
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
||||
SOFTWARE.
|
||||
@ -0,0 +1,29 @@
|
||||
# RAG AWS Bedrock
|
||||
|
||||
AWS Bedrock is a managed serve that offers a set of foundation models.
|
||||
|
||||
Here we will use `Anthropic Claude` for text generation and `Amazon Titan` for text embedding.
|
||||
|
||||
We will use Pinecode as our vectorstore.
|
||||
|
||||
(See [this notebook](https://github.com/aws-samples/amazon-bedrock-workshop/blob/main/03_QuestionAnswering/01_qa_w_rag_claude.ipynb) for additional context on the RAG pipeline.)
|
||||
|
||||
(See [this notebook](https://github.com/aws-samples/amazon-bedrock-workshop/blob/58f238a183e7e629c9ae11dd970393af4e64ec44/00_Intro/bedrock_boto3_setup.ipynb#Prerequisites) for additional context on setup.)
|
||||
|
||||
## Pinecone
|
||||
|
||||
This connects to a hosted Pinecone vectorstore.
|
||||
|
||||
Be sure that you have set a few env variables in `chain.py`:
|
||||
|
||||
* `PINECONE_API_KEY`
|
||||
* `PINECONE_ENV`
|
||||
* `index_name`
|
||||
|
||||
## LLM and Embeddings
|
||||
|
||||
Be sure to set AWS enviorment variables:
|
||||
|
||||
* `AWS_DEFAULT_REGION`
|
||||
# `AWS_PROFILE`
|
||||
* `BEDROCK_ASSUME_ROLE`
|
||||
File diff suppressed because it is too large
Load Diff
@ -0,0 +1,24 @@
|
||||
[tool.poetry]
|
||||
name = "rag-aws-bedrock"
|
||||
version = "0.1.0"
|
||||
description = ""
|
||||
authors = ["Lance Martin <lance@langchain.dev>"]
|
||||
readme = "README.md"
|
||||
|
||||
[tool.poetry.dependencies]
|
||||
python = ">=3.8.1,<4.0"
|
||||
langchain = ">=0.0.313, <0.1"
|
||||
openai = ">=0.28.1"
|
||||
tiktoken = ">=0.5.1"
|
||||
pinecone-client = ">=2.2.4"
|
||||
boto3 = ">=1.28.57"
|
||||
awscli = ">=1.29.57"
|
||||
botocore = ">=1.31.57"
|
||||
|
||||
[tool.langserve]
|
||||
export_module = "rag_aws_bedrock"
|
||||
export_attr = "chain"
|
||||
|
||||
[build-system]
|
||||
requires = ["poetry-core"]
|
||||
build-backend = "poetry.core.masonry.api"
|
||||
@ -0,0 +1,50 @@
|
||||
{
|
||||
"cells": [
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "681a5d1e",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## Connect to template\n",
|
||||
"\n",
|
||||
"In `server.py`, set -\n",
|
||||
"```\n",
|
||||
"add_routes(app, chain_ext, path=\"/rag_aws_bedrock\")\n",
|
||||
"```"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"id": "d774be2a",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"from langserve.client import RemoteRunnable\n",
|
||||
"rag_app_pinecone = RemoteRunnable('http://0.0.0.0:8001/rag_aws_bedrock')\n",
|
||||
"rag_app_pinecone.invoke(\"What are the different types of agent memory\")"
|
||||
]
|
||||
}
|
||||
],
|
||||
"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": "3.9.16"
|
||||
}
|
||||
},
|
||||
"nbformat": 4,
|
||||
"nbformat_minor": 5
|
||||
}
|
||||
@ -0,0 +1,3 @@
|
||||
from rag_aws_bedrock.chain import chain
|
||||
|
||||
__all__ = ["chain"]
|
||||
@ -0,0 +1,73 @@
|
||||
import os
|
||||
|
||||
from langchain.embeddings import BedrockEmbeddings
|
||||
from langchain.llms.bedrock import Bedrock
|
||||
from langchain.prompts import ChatPromptTemplate
|
||||
from langchain.schema.output_parser import StrOutputParser
|
||||
from langchain.schema.runnable import RunnableParallel, RunnablePassthrough
|
||||
from langchain.vectorstores import Pinecone
|
||||
from utils import bedrock
|
||||
|
||||
if os.environ.get("PINECONE_API_KEY", None) is None:
|
||||
raise Exception("Missing `PINECONE_API_KEY` environment variable.")
|
||||
|
||||
if os.environ.get("PINECONE_ENVIRONMENT", None) is None:
|
||||
raise Exception("Missing `PINECONE_ENVIRONMENT` environment variable.")
|
||||
|
||||
if os.environ.get("AWS_DEFAULT_REGION", None) is None:
|
||||
raise Exception("Missing `AWS_DEFAULT_REGION` environment variable.")
|
||||
|
||||
if os.environ.get("AWS_PROFILE", None) is None:
|
||||
raise Exception("Missing `AWS_PROFILE` environment variable.")
|
||||
|
||||
if os.environ.get("BEDROCK_ASSUME_ROLE", None) is None:
|
||||
raise Exception("Missing `BEDROCK_ASSUME_ROLE` environment variable.")
|
||||
|
||||
PINECONE_INDEX_NAME = os.environ.get("PINECONE_INDEX", "langchain-test")
|
||||
|
||||
### Ingest code - you may need to run this the first time
|
||||
# Load
|
||||
# from langchain.document_loaders import WebBaseLoader
|
||||
# loader = WebBaseLoader("https://lilianweng.github.io/posts/2023-06-23-agent/")
|
||||
# data = loader.load()
|
||||
|
||||
# # Split
|
||||
# from langchain.text_splitter import RecursiveCharacterTextSplitter
|
||||
# text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=0)
|
||||
# all_splits = text_splitter.split_documents(data)
|
||||
|
||||
# # Add to vectorDB
|
||||
# vectorstore = Pinecone.from_documents(
|
||||
# documents=all_splits, embedding=OpenAIEmbeddings(), index_name=PINECONE_INDEX_NAME
|
||||
# )
|
||||
# retriever = vectorstore.as_retriever()
|
||||
|
||||
# Set LLM and embeddings
|
||||
boto3_bedrock = bedrock.get_bedrock_client(
|
||||
assumed_role=os.environ.get("BEDROCK_ASSUME_ROLE", None),
|
||||
region=os.environ.get("AWS_DEFAULT_REGION", None)
|
||||
)
|
||||
model = Bedrock(model_id="anthropic.claude-v2",
|
||||
client=boto3_bedrock,
|
||||
model_kwargs={'max_tokens_to_sample':200})
|
||||
bedrock_embeddings = BedrockEmbeddings(model_id="amazon.titan-embed-text-v1",
|
||||
client=boto3_bedrock)
|
||||
|
||||
# Set vectostore
|
||||
vectorstore = Pinecone.from_existing_index(PINECONE_INDEX_NAME, bedrock_embeddings)
|
||||
retriever = vectorstore.as_retriever()
|
||||
|
||||
# RAG prompt
|
||||
template = """Answer the question based only on the following context:
|
||||
{context}
|
||||
Question: {question}
|
||||
"""
|
||||
prompt = ChatPromptTemplate.from_template(template)
|
||||
|
||||
# RAG
|
||||
chain = (
|
||||
RunnableParallel({"context": retriever, "question": RunnablePassthrough()})
|
||||
| prompt
|
||||
| model
|
||||
| StrOutputParser()
|
||||
)
|
||||
Loading…
Reference in New Issue