ChatOpenAI is a LangChain chat-model integration that sends structured prompts to OpenAI conversational language models and returns generated responses. In a retrieval-augmented generation pipeline, it answers questions using context retrieved from MongoDB Atlas Vector Search. section 14.5.4 section 14.5.2
A typical ChatOpenAI pipeline connects these stages:
- RunnableParallel prepares the question and retrieval context. section 14.5.4
- The retriever obtains relevant documents from an Atlas Vector Search index. section 14.5.4
- ChatPromptTemplate combines document content, metadata, and the user’s question into a structured prompt. section 14.5.4
- ChatOpenAI generates an answer, while StrOutputParser extracts the response text for presentation. section 14.5.4
Answering a retrieved question
A LangServe chatbot can retrieve documents for the question “What indexes does MongoDB support?”, pass their content and metadata through ChatPromptTemplate, and use ChatOpenAI to produce an answer. The resulting generation includes a stop reason and the model name; StrOutputParser returns the usable text. section 14.5.4
import os
from langchain_openai import ChatOpenAI
from langchain_community.document_loaders import PyPDFLoader
from langchain_openai import OpenAIEmbeddings
from langchain_community.vectorstores import MongoDBAtlasVectorSearch
from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.pydantic_v1 import BaseModel
from langchain_core.runnables import (
RunnableLambda,
RunnableParallel,
RunnablePassthrough,
)
from langchain_text_splitters import RecursiveCharacterTextSplitter
from pymongo import MongoClient
# Set DB
if os.environ.get("MONGO_URI", None) is None:
raise Exception("Missing `MONGO_URI` environment variable.")
MONGO_URI = os.environ["MONGO_URI"]
DB_NAME = "langchain"
COLLECTION_NAME = "mongodb"
ATLAS_VECTOR_SEARCH_INDEX_NAME = "default"
client = MongoClient(MONGO_URI)
db = client[DB_NAME]
MONGODB_COLLECTION = db[COLLECTION_NAME]
# Read from MongoDB Atlas Vector Search
vectorstore = MongoDBAtlasVectorSearch.from_connection_string(
MONGO_URI,
DB_NAME + "." + COLLECTION_NAME,
OpenAIEmbeddings(disallowed_special=()),
index_name=ATLAS_VECTOR_SEARCH_INDEX_NAME,
)
retriever = vectorstore.as_retriever(search_kwargs={"k": 1})
# 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": retriever,
➥"question": RunnablePassthrough()})
| prompt
| model
| StrOutputParser()
)
# Add typing for input
class Question(BaseModel):
__root__: str
chain = chain.with_types(input_type=Question)
def _ingest(url: str) -> dict:
loader = PyPDFLoader(url)
data = loader.load()
# Split docs
text_splitter = RecursiveCharacterTextSplitter
➥(chunk_size=500, chunk_overlap=0)
docs = text_splitter.split_documents(data)
# Insert the documents in MongoDB Atlas Vector Search
_ = MongoDBAtlasVectorSearch.from_documents(
documents=docs,
embedding=OpenAIEmbeddings(disallowed_special=()),
collection=MONGODB_COLLECTION,
index_name=ATLAS_VECTOR_SEARCH_INDEX_NAME,
)
return {}
ingest = RunnableLambda(_ingest)
Inspect the chain definition showing ChatOpenAI alongside document loading, embeddings, MongoDB Atlas Vector Search, prompt formatting, runnable stages, and output parsing. section 14.5.2
