VectorDB Question Answering with Sources

This notebook goes over how to do question-answering with sources over a vector database. It does this by using the VectorDBQAWithSourcesChain, which does the lookup of the documents from a vector database.

from langchain.embeddings.openai import OpenAIEmbeddings
from langchain.embeddings.cohere import CohereEmbeddings
from langchain.text_splitter import CharacterTextSplitter
from langchain.vectorstores.elastic_vector_search import ElasticVectorSearch
from langchain.vectorstores import Chroma

with open('../../state_of_the_union.txt') as f:
    state_of_the_union = f.read()
text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
texts = text_splitter.split_text(state_of_the_union)

embeddings = OpenAIEmbeddings()

docsearch = Chroma.from_texts(texts, embeddings, metadatas=[{"source": f"{i}-pl"} for i in range(len(texts))])

Running Chroma using direct local API.
Using DuckDB in-memory for database. Data will be transient.
Exiting: Cleaning up .chroma directory

from langchain.chains import VectorDBQAWithSourcesChain

from langchain import OpenAI

chain = VectorDBQAWithSourcesChain.from_chain_type(OpenAI(temperature=0), chain_type="stuff", vectorstore=docsearch)

chain({"question": "What did the president say about Justice Breyer"}, return_only_outputs=True)

{'answer': ' The president thanked Justice Breyer for his service and mentioned his legacy of excellence.\n',
 'sources': '30-pl'}

Chain Type

You can easily specify different chain types to load and use in the VectorDBQAWithSourcesChain chain. For a more detailed walkthrough of these types, please see this notebook.

There are two ways to load different chain types. First, you can specify the chain type argument in the from_chain_type method. This allows you to pass in the name of the chain type you want to use. For example, in the below we change the chain type to map_reduce.

chain = VectorDBQAWithSourcesChain.from_chain_type(OpenAI(temperature=0), chain_type="map_reduce", vectorstore=docsearch)

chain({"question": "What did the president say about Justice Breyer"}, return_only_outputs=True)

{'answer': ' The president honored Justice Stephen Breyer for his service.\n',
 'sources': '30-pl'}

The above way allows you to really simply change the chain_type, but it does provide a ton of flexibility over parameters to that chain type. If you want to control those parameters, you can load the chain directly (as you did in this notebook) and then pass that directly to the the VectorDBQA chain with the combine_documents_chain parameter. For example:

from langchain.chains.qa_with_sources import load_qa_with_sources_chain
qa_chain = load_qa_with_sources_chain(OpenAI(temperature=0), chain_type="stuff")
qa = VectorDBQAWithSourcesChain(combine_documents_chain=qa_chain, vectorstore=docsearch)

qa({"question": "What did the president say about Justice Breyer"}, return_only_outputs=True)

{'answer': ' The president honored Justice Stephen Breyer for his service.\n',
 'sources': '30-pl'}