import streamlit as st
import pickle
import pandas as pd
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics.pairwise import cosine_similarity

# Load the model and necessary files
@st.cache_resource
def load_model():
    with open("model2.pkl", "rb") as f:
        tfidf_matrix, vectorizer, similarity_matrix, df = pickle.load(f)
    return tfidf_matrix, vectorizer, similarity_matrix, df

# Load the model
tfidf_matrix, vectorizer, similarity_matrix, df = load_model()

# Streamlit user interface
st.title("Book Recommender System")

# Input from the user
title = st.text_input("Enter a book title:")

# Function to get recommendations
def get_recommendations(title, df, similarity_matrix):
    if title in df['book_name'].values:
        book_index = df[df['book_name'] == title].index[0]
        similarity_scores = list(enumerate(similarity_matrix[book_index]))
        similarity_scores = sorted(similarity_scores, key=lambda x: x[1], reverse=True)[1:6]

        recommendations = []
        for i in similarity_scores:
            recommended_book = df.iloc[i[0]]['book_name']
            if recommended_book != title and recommended_book not in recommendations:
                recommendations.append(recommended_book)
        return recommendations
    else:
        return ["Book not found in dataset."]

# Show recommendations when the user inputs a title
if title:
    recommendations = get_recommendations(title, df, similarity_matrix)
    st.write(f"Recommendations for '{title}':")
    for rec in recommendations:
        st.write(f"- {rec}")