app.py

import os
import pandas as pd
import streamlit as st
import logging
from transformers import pipeline

# Configure logging
logging.basicConfig(level=logging.DEBUG)


# Retrieve Hugging Face API token from environment variables
HUGGINGFACEHUB_API_TOKEN = os.getenv("HUGGINGFACEHUB_API_TOKEN")
if not HUGGINGFACEHUB_API_TOKEN:
    logging.error("Hugging Face API token is missing or invalid.")
    st.error("Hugging Face API token is missing or invalid.")
    st.stop()

# Now you can use HUGGINGFACEHUB_API_TOKEN in your code, for example:
from transformers import pipeline

# Load the model using the Hugging Face API token
MODEL_NAME = "mistralai/Mistral-Large-Instruct-2407"
llm_pipeline = pipeline("text-generation", model=MODEL_NAME, use_auth_token=HUGGINGFACEHUB_API_TOKEN)

# Load datasets
def load_datasets():
    """Load datasets from CSV files."""
    try:
        electronics_df = pd.read_csv('electronics.csv')
        fashion_df = pd.read_csv('fashion.csv')
        return electronics_df, fashion_df
    except Exception as e:
        logging.error(f"Error loading datasets: {e}")
        st.error("Error loading datasets.")
        st.stop()

electronics_df, fashion_df = load_datasets()

# Keywords for routing queries
electronics_keywords = [
    'electronics', 'device', 'gadget', 'battery', 'performance', 
    'phone', 'mobile', 'laptop', 'tv', 'bluetooth', 'speakers', 
    'washing machine', 'headphones', 'camera', 'tablet', 'charger', 
    'smartwatch', 'refrigerator'
]
fashion_keywords = [
    'fashion', 'clothing', 'size', 'fit', 'material', 'shirt', 
    'pants', 'coats', 'shoes', 'girls dress', 'sarees', 'skirts', 
    'jackets', 'sweaters', 'suits', 'accessories', 't-shirts'
]

def determine_category(query):
    """Determine the category based on the query."""
    query_lower = query.lower()
    if any(keyword in query_lower for keyword in electronics_keywords):
        logging.debug(f"Query '{query}' categorized as 'electronics'.")
        return 'electronics'
    elif any(keyword in query_lower for keyword in fashion_keywords):
        logging.debug(f"Query '{query}' categorized as 'fashion'.")
        return 'fashion'
    else:
        logging.debug(f"Query '{query}' categorized as 'general'.")
        return 'general'

def format_electronics_response(row):
    """Format response using data from an Electronics DataFrame row."""
    response = (
        f"**Product Name:** {row['ProductName']}\n\n"
        f"**Description:**\n{row['Description']}\n\n"
        f"**Price:** ${row['Price']}\n"
        f"**Brand:** {row['Brand']}\n"
        f"**Model:** {row['Model']}\n"
        f"**Department:** {row['Department']}\n\n"
        f"**Reviews:**\n{row['Reviews']}\n\n"
        f"**Ratings:** {row['Ratings']} / 5\n"
    )
    return response

def format_fashion_response(row):
    """Format response using data from a Fashion DataFrame row."""
    response = (
        f"**Product Name:** {row['Name']}\n\n"
        f"**Description:**\n{row['Description']}\n\n"
        f"**Price:** ${row['Price']}\n"
        f"**Brand:** {row['Brand']}\n"
        f"**Model:** {row['Model']}\n\n"
        f"**Rating:** {row['Rating']} / 5\n"
    )
    return response

def extract_filters(query):
    """Extract filters from the user's query."""
    filters = {}
    query_lower = query.lower()

    if 'best' in query_lower:
        if 'rating' in query_lower:
            filters['Rating'] = category_df['Rating'].max()
    
    if 'phones' in query_lower:
        filters['Category'] = 'phone'
    elif 'laptops' in query_lower:
        filters['Category'] = 'laptop'

    return filters

def apply_filters(df, filters):
    """Apply filters to a DataFrame based on the provided filter dictionary."""
    for key, value in filters.items():
        if key in df.columns:
            if isinstance(value, str):
                df = df[df[key].str.contains(value, case=False, na=False)]
            elif isinstance(value, (int, float)):
                df = df[df[key] == value]
    return df

def fetch_response_from_df(query, category_df, format_response_func):
    """Fetch response from the dataset based on the category and filters."""
    filters = extract_filters(query)
    
    filtered_df = apply_filters(category_df, filters)
    
    if filtered_df.empty:
        return "Sorry, I couldn't find an answer in our records."

    responses = []
    for _, row in filtered_df.iterrows():
        responses.append(format_response_func(row))
    
    if responses:
        return "\n\n".join(responses)
    
    return "Sorry, I couldn't find an answer in our records."

def electronics_response(query):
    """Get response from the electronics dataset."""
    return fetch_response_from_df(query, electronics_df, format_electronics_response)

def fashion_response(query):
    """Get response from the fashion dataset."""
    return fetch_response_from_df(query, fashion_df, format_fashion_response)

def get_response(user_input):
    """Determine the category and fetch the appropriate response."""
    if 'hi' in user_input.lower() or 'hello' in user_input.lower():
        return "Hi, welcome to the customer support chatbot. How can I help you?"

    category = determine_category(user_input)
    
    if category == 'electronics':
        response = electronics_response(user_input)
    elif category == 'fashion':
        response = fashion_response(user_input)
    else:
        response = "Sorry, I couldn't find an answer in our records."

    return response

# Streamlit Interface
def main():
    st.title("Customer Support Chatbot")

    # Custom CSS for chat bubbles
    st.markdown("""
    <style>
    .chat-container {
        max-width: 800px;
        margin: 0 auto;
        padding: 20px;
    }
    .chat-bubble {
        border-radius: 15px;
        padding: 10px;
        margin: 5px 0;
        max-width: 70%;
        display: inline-block;
        word-wrap: break-word;
    }
    .user-bubble {
        background-color: #DCF8C6;
        float: right;
        text-align: left;
    }
    .assistant-bubble {
        background-color: #FFFFFF;
        float: left;
        text-align: left;
    }
    .chat-history {
        max-height: 500px;
        overflow-y: auto;
    }
    </style>
    """, unsafe_allow_html=True)

    if 'chat_history' not in st.session_state:
        st.session_state.chat_history = []

    st.markdown('<div class="chat-container">', unsafe_allow_html=True)
    
    st.markdown('<div class="chat-history">', unsafe_allow_html=True)
    for message in st.session_state.chat_history:
        if message['role'] == 'user':
            st.markdown(f'<div class="chat-bubble user-bubble">{message["content"]}</div>', unsafe_allow_html=True)
        else:
            st.markdown(f'<div class="chat-bubble assistant-bubble">{message["content"]}</div>', unsafe_allow_html=True)
    st.markdown('</div>', unsafe_allow_html=True)
    
    user_input = st.text_input("Type your message here:")

    if st.button("Send"):
        if user_input:
            response_message = get_response(user_input)
            st.session_state.chat_history.append({"role": "user", "content": user_input})
            st.session_state.chat_history.append({"role": "assistant", "content": response_message})
            
            # Clear the input after sending
            st.session_state.user_input = ''
    
    st.markdown('</div>', unsafe_allow_html=True)

if __name__ == "__main__":
    main()