mirror of
https://github.com/hwchase17/langchain
synced 2024-11-08 07:10:35 +00:00
138f97af23
Without this additional `#`, one needs to add it manually after uncommenting the section.
74 lines
2.2 KiB
Python
74 lines
2.2 KiB
Python
import os
|
|
|
|
from langchain.chat_models import ChatOpenAI
|
|
from langchain.embeddings import OpenAIEmbeddings
|
|
from langchain.prompts import ChatPromptTemplate
|
|
from langchain.retrievers import ContextualCompressionRetriever
|
|
from langchain.retrievers.document_compressors import CohereRerank
|
|
from langchain.vectorstores import Pinecone
|
|
from langchain_core.output_parsers import StrOutputParser
|
|
from langchain_core.pydantic_v1 import BaseModel
|
|
from langchain_core.runnables import RunnableParallel, RunnablePassthrough
|
|
|
|
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.")
|
|
|
|
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()
|
|
|
|
vectorstore = Pinecone.from_existing_index(PINECONE_INDEX_NAME, OpenAIEmbeddings())
|
|
|
|
# Get k=10 docs
|
|
retriever = vectorstore.as_retriever(search_kwargs={"k": 10})
|
|
|
|
# Re-rank
|
|
compressor = CohereRerank()
|
|
compression_retriever = ContextualCompressionRetriever(
|
|
base_compressor=compressor, base_retriever=retriever
|
|
)
|
|
|
|
# RAG prompt
|
|
template = """Answer the question based only on the following context:
|
|
{context}
|
|
Question: {question}
|
|
"""
|
|
prompt = ChatPromptTemplate.from_template(template)
|
|
|
|
# RAG
|
|
model = ChatOpenAI()
|
|
chain = (
|
|
RunnableParallel(
|
|
{"context": compression_retriever, "question": RunnablePassthrough()}
|
|
)
|
|
| prompt
|
|
| model
|
|
| StrOutputParser()
|
|
)
|
|
|
|
|
|
# Add typing for input
|
|
class Question(BaseModel):
|
|
__root__: str
|
|
|
|
|
|
chain = chain.with_types(input_type=Question)
|