LangChain API

Search Options for Deep Lake Vector Stores in LangChain

This tutorial requires installation of:

!pip3 install deeplake langchain openai tiktoken

Vector Search in Python

Let's load the same vector store used in the Quickstart and run embeddings search based on a user prompt using the LangChain API.

from langchain.vectorstores import DeepLake
from langchain.chains import RetrievalQA
from langchain.llms import OpenAIChat
from langchain.embeddings.openai import OpenAIEmbeddings
import os

os.environ['OPENAI_API_KEY'] = <OPENAI_API_KEY>

vector_store_path = 'hub://activeloop/paul_graham_essay'

embedding_function = OpenAIEmbeddings(model='text-embedding-ada-002')

# Re-load the vector store
db = DeepLake(dataset_path = vector_store_path, embedding_function=embedding_function, read_only=True)

qa = RetrievalQA.from_chain_type(llm=OpenAIChat(model='gpt-3.5-turbo'), chain_type='stuff', retriever=db.as_retriever())

Vector Similarity Search

Let's run a similarity search on Paul Graham's essay based on a query we want to answer. The query is embedded and a similarity search is performed against the stored embeddings, with execution taking place on the client.

prompt = 'What are the first programs he tried writing?'

query_docs = db.similarity_search(query = prompt)

If we print the first document using query_docs[0].page_content, it appears to be relevant to the query:

What I Worked On

February 2021

Before college the two main things I worked on, outside of school, were writing and programming. I didn't write essays. I wrote what beginning writers were supposed to write then, and probably still are: short stories. My stories were awful. They had hardly any plot, just characters with strong feelings, which I imagined made them deep.

The first programs I tried writing were on the IBM 1401 that our school district used for what was then called "data processing." This was in 9th grade, so I was 13 or 14. The school district's 1401 happened to be in the basement of our junior high school, and my friend Rich Draves and I got permission to use it. It was like a mini Bond villain's lair down there, with all these alien-looking machines — CPU, disk drives, printer, card reader — sitting up on a raised floor under bright fluorescent lights.

Vector Search in an LLM Context

We can directly use LangChain to run a Q&A using an LLM and answer the question about Paul Graham's essay. Internally, this API performs an embedding search to find the most relevant data to feeds them into the LLM context.

qa = RetrievalQA.from_chain_type(llm = OpenAIChat(model = 'gpt-3.5-turbo'), 
                                 chain_type = 'stuff', 
                                 retriever = db.as_retriever())

qa.run(prompt)

'The first programs he tried writing were on the IBM 1401 that his school district used for "data processing" in 9th grade.'

Vector Search Using the Compute Engine on the Client Side in LangChain

The workflow above can also be performed on the client using Deep Lake's Compute Engine, which offers faster execution for larger Vector Stores and complex queries.

The similarity search, queries are executed on Compute Engine by and by specifying exec_option = "compute_engine" in the parameters for similarity_search.

query_docs = db.similarity_search(query = prompt, exec_option = "compute_engine")

For chains, we should specify the exec_option as "compute_engine" in the search_kwargs.

retriever = db.as_retriever()
retriever.search_kwargs["exec_option"] = "compute_engine"

qa = RetrievalQA.from_chain_type(llm = OpenAIChat(model = 'gpt-3.5-turbo'), 
                                 chain_type = 'stuff', 
                                 retriever = retriever)
                                 
qa.run(prompt) 

Vector Search Using the Managed Tensor Database in LangChain

Vector search using the Tensor Database + LangChain API will be available soon.