|
|
import os |
|
|
import streamlit as st |
|
|
from langchain.embeddings import HuggingFaceEmbeddings |
|
|
from langchain.chains import RetrievalQA |
|
|
from langchain_community.vectorstores import FAISS |
|
|
from langchain_core.prompts import PromptTemplate |
|
|
from langchain_huggingface import HuggingFaceEndpoint |
|
|
from dotenv import load_dotenv, find_dotenv |
|
|
|
|
|
|
|
|
|
|
|
load_dotenv(find_dotenv()) |
|
|
|
|
|
|
|
|
DB_FAISS_PATH = "vectorstore/db_faiss" |
|
|
|
|
|
@st.cache_resource |
|
|
def get_vectorstore(): |
|
|
"""Loads the FAISS vector store with embeddings.""" |
|
|
try: |
|
|
embedding_model = HuggingFaceEmbeddings(model_name='sentence-transformers/all-MiniLM-L6-v2') |
|
|
return FAISS.load_local(DB_FAISS_PATH, embedding_model, allow_dangerous_deserialization=True) |
|
|
except Exception as e: |
|
|
st.error(f"β οΈ Error loading vector store: {str(e)}") |
|
|
return None |
|
|
|
|
|
@st.cache_resource |
|
|
def load_llm(): |
|
|
"""Loads the Hugging Face LLM model for text generation.""" |
|
|
HUGGINGFACE_REPO_ID = "mistralai/Mistral-7B-Instruct-v0.3" |
|
|
HF_TOKEN = os.getenv("HF_TOKEN") |
|
|
|
|
|
if not HF_TOKEN: |
|
|
st.error("β οΈ Hugging Face API token is missing. Please check your environment variables.") |
|
|
return None |
|
|
|
|
|
try: |
|
|
return HuggingFaceEndpoint( |
|
|
repo_id=HUGGINGFACE_REPO_ID, |
|
|
task="text-generation", |
|
|
temperature=0.3, |
|
|
model_kwargs={"token": HF_TOKEN, "max_length": 256} |
|
|
) |
|
|
except Exception as e: |
|
|
st.error(f"β οΈ Error loading LLM: {str(e)}") |
|
|
return None |
|
|
|
|
|
def set_custom_prompt(): |
|
|
"""Defines the chatbot's behavior with a custom prompt template.""" |
|
|
return PromptTemplate( |
|
|
template=""" |
|
|
You are an SEO chatbot with advanced knowledge. Answer based **strictly** on the provided documents. |
|
|
|
|
|
If the answer is in the context, provide a **clear, professional, and concise** response with sources. |
|
|
If the question is **outside the given context**, politely decline: |
|
|
|
|
|
**"I'm sorry, but I can only provide answers based on the available documents."** |
|
|
|
|
|
**Context:** {context} |
|
|
**Question:** {question} |
|
|
|
|
|
**Answer:** |
|
|
""", |
|
|
input_variables=["context", "question"] |
|
|
) |
|
|
|
|
|
def generate_response(prompt, vectorstore, llm): |
|
|
"""Retrieves relevant documents and generates a response from the LLM.""" |
|
|
if not vectorstore or not llm: |
|
|
return "β Unable to process your request due to initialization issues." |
|
|
|
|
|
try: |
|
|
qa_chain = RetrievalQA.from_chain_type( |
|
|
llm=llm, |
|
|
chain_type="stuff", |
|
|
retriever=vectorstore.as_retriever(search_kwargs={'k': 3}), |
|
|
return_source_documents=True, |
|
|
chain_type_kwargs={'prompt': set_custom_prompt()} |
|
|
) |
|
|
|
|
|
response_data = qa_chain.invoke({'query': prompt}) |
|
|
result = response_data.get("result", "") |
|
|
source_documents = response_data.get("source_documents", []) |
|
|
|
|
|
if not result or not source_documents: |
|
|
return "β Sorry, but I can only provide answers based on the available documents." |
|
|
|
|
|
formatted_sources = "\n\nπ **Sources:**" + "".join( |
|
|
[f"\n- {doc.metadata.get('source', 'Unknown')} (Page: {doc.metadata.get('page', 'N/A')})" for doc in source_documents] |
|
|
) |
|
|
return f"{result}{formatted_sources}" |
|
|
|
|
|
except Exception as e: |
|
|
return f"β οΈ **Error:** {str(e)}" |
|
|
|
|
|
def main(): |
|
|
"""Runs the Streamlit chatbot application.""" |
|
|
st.title("π§ Brainmines SEO Chatbot - Your AI Assistant for SEO Queries π") |
|
|
|
|
|
|
|
|
vectorstore = get_vectorstore() |
|
|
llm = load_llm() |
|
|
|
|
|
if not vectorstore or not llm: |
|
|
st.error("β οΈ Failed to initialize vector store or LLM. Please check configurations.") |
|
|
return |
|
|
|
|
|
|
|
|
if "messages" not in st.session_state: |
|
|
st.session_state.messages = [ |
|
|
{"role": "assistant", "content": "Hello! π I'm here to assist you with SEO-related queries. π"}, |
|
|
] |
|
|
|
|
|
|
|
|
for message in st.session_state.messages: |
|
|
st.chat_message(message["role"]).markdown(message["content"]) |
|
|
|
|
|
prompt = st.chat_input("π¬ Enter your SEO question here") |
|
|
|
|
|
if prompt: |
|
|
st.chat_message("user").markdown(prompt) |
|
|
st.session_state.messages.append({"role": "user", "content": prompt}) |
|
|
|
|
|
with st.spinner("Thinking... π€"): |
|
|
response = generate_response(prompt, vectorstore, llm) |
|
|
|
|
|
st.chat_message("assistant").markdown(response) |
|
|
st.session_state.messages.append({"role": "assistant", "content": response}) |
|
|
|
|
|
if __name__ == "__main__": |
|
|
main() |
|
|
|