Project Title: ChatGroq RAG Chatbot

Overview

ChatGroq is a Retrieval-Augmented Generation (RAG) chatbot leveraging Groq’s Mixtral model, LangChain, and FAISS to provide accurate, context-aware responses. This project was developed using Streamlit for an interactive UI and is designed to scrape and retrieve relevant information from LangSmith documentation.

Tech Stack

• LLM Groq's Mixtral-8x7b-32768
• Document Loader: LangChain WebBaseLoader
• Embedding Model: Ollama Embeddings
• Vector Store
• Text Splitter: RecursiveCharacterTextSplitter
• Framework

Implementation Details:-

import streamlit as st
import os
from langchain_groq import ChatGroq
from langchain_community.document_loaders import WebBaseLoader
from langchain.embeddings import OllamaEmbeddings
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.chains.combine_documents import create_stuff_documents_chain
from langchain_core.prompts import ChatPromptTemplate
from langchain_chains import create_retrieval_chain
from langchain_community.vectorstores import FAISS
import time

 Set the API key
os.environ["GROQ_API_KEY"] = "your_groq_api_key_here"

if "vector" not in st.session_state:
    st.session_state.embeddings = OllamaEmbeddings()
    st.session_state.loader = WebBaseLoader("https://docs.smith.langchain.com/")
    st.session_state.docs = st.session_state.loader.load()

    st.session_state.text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
    st.session_state.final_documents = st.session_state.text_splitter.split_documents(st.session_state.docs[:50])
    st.session_state.vectors = FAISS.from_documents(st.session_state.final_documents, st.session_state.embeddings)

st.title("ChatGroq Demo")
llm = ChatGroq(groq_api_key=os.environ.get("GROQ_API_KEY"), model_name="mixtral-8x7b-32768")

prompt = ChatPromptTemplate.from_template(
"""
Answer the questions based on the provided context only.
Please provide the most accurate response based on the question
<context>
{context}
<context>
Questions:{input}
"""
)
document_chain = create_stuff_documents_chain(llm, prompt)
retriever = st.session_state.vectors.as_retriever()
retrieval_chain = create_retrieval_chain(retriever, document_chain)

prompt = st.text_input("Input your prompt here")

if prompt:
    start = time.process_time()
    response = retrieval_chain.invoke({"input": prompt})
    print("Response time:", time.process_time() - start)
    st.write(response['answer'])

    with st.expander("Document Similarity Search"):
        for i, doc in enumerate(response["context"]):
            st.write(doc.page_content)
            st.write("--------------------------------")

Key Features

• Retrieval-Augmented Generation
  accuracy by retrieving relevant documents before generating responses.
• Efficient Embeddings & Vector Search: Uses FAISS for fast similarity searches.
• Live Web Scraping
  loads LangSmith documentation for context.
• User-Friendly Interface: Built with Streamlit for easy interaction.
• Performance Tracking: Measures response time for efficiency evaluation.

Potential Improvements

• Multi-source Retrieval: Expanding to multiple documentation sources.
• Enhanced UI: Better visualization for retrieved documents.
• Fine-tuned LLMs: Testing different models for improved performance.

Impact

This project showcases a powerful application of Agentic AI by combining LLMs with retrieval mechanisms, making it suitable for various real-world applications such as customer support, research assistants, and knowledge management systems.