Project: Document Q&A with RAG

RAG allows the LLM to answer questions based on your private documents, not just its training data.

• A simple web app (Gradio) where users upload a PDF or enter text.
• Chunk the document, create embeddings, and store in a vector database (Chroma).
• For a question, retrieve the most relevant chunks and ask the LLM to answer.

• Python installed.
• OpenAI API key.
• Basic Python knowledge.

pip install langchain chromadb gradio openai tiktoken pypdf

import gradio as gr
from langchain.document_loaders import PyPDFLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.embeddings import OpenAIEmbeddings
from langchain.vectorstores import Chroma
from langchain.chains import RetrievalQA
from langchain.chat_models import ChatOpenAI
import os

os.environ["OPENAI_API_KEY"] = "your-api-key"

def process_document(file):
    # Load PDF
    loader = PyPDFLoader(file.name)
    documents = loader.load()
    # Split into chunks
    text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
    chunks = text_splitter.split_documents(documents)
    # Create embeddings and vector store
    embeddings = OpenAIEmbeddings()
    vectorstore = Chroma.from_documents(chunks, embeddings)
    return vectorstore

def answer_question(file, question):
    if file is None:
        return "Please upload a document first."
    vectorstore = process_document(file)
    # Create QA chain
    llm = ChatOpenAI(model="gpt-3.5-turbo")
    qa_chain = RetrievalQA.from_chain_type(llm, retriever=vectorstore.as_retriever())
    answer = qa_chain.run(question)
    return answer

# Gradio interface
iface = gr.Interface(
    fn=answer_question,
    inputs=[gr.File(label="Upload PDF"), gr.Textbox(label="Your Question")],
    outputs="text",
    title="Document Q&A with RAG",
    description="Upload a PDF and ask any question about its content."
)

iface.launch()

python doc_qna.py

• The PDF is split into overlapping chunks (1000 characters each).
• Each chunk is converted into an embedding (vector) using OpenAI’s embedding model.
• Embeddings are stored in a Chroma vector database.
• When you ask a question, it is also embedded, and the most similar chunks are retrieved.
• The LLM receives the question + the retrieved chunks and generates an answer grounded in the document.