import streamlit as st
import pandas as pd


st.title("Movie Recommendation System")

st.write("This app recommends similar movies based on user ratings.")


# Verileri okuyoruz
ratings = pd.read_csv("src/ratings.csv")
movies = pd.read_csv("src/movies.csv")


# Verileri birleştiriyoruz
df = pd.merge(ratings, movies, on="movieId")


# Film istatistikleri
movie_stats = df.groupby("title")["rating"].agg(["mean", "count"])


# Kullanıcı-film puan matrisi
movie_matrix = df.pivot_table(
    index="userId",
    columns="title",
    values="rating"
)


def recommend_movies(movie_name):

    movie_user_ratings = movie_matrix[movie_name]

    similar_movies = movie_matrix.corrwith(
        movie_user_ratings
    )

    corr_df = pd.DataFrame(
        similar_movies,
        columns=["Correlation"]
    )

    corr_df.dropna(inplace=True)

    corr_df = corr_df.join(
        movie_stats["count"]
    )

    recommendations = corr_df[
        corr_df["count"] > 100
    ]

    recommendations = recommendations.sort_values(
        "Correlation",
        ascending=False
    )

    recommendations = recommendations[
        recommendations.index != movie_name
    ]

    return recommendations.head(10)



movie_list = sorted(
    movie_stats[movie_stats["count"] > 100].index
)

selected_movie = st.selectbox(
    "Select a movie",
    movie_list
)


if st.button("Recommend"):

    result = recommend_movies(selected_movie)

    st.subheader("Recommended Movies")

    if result.empty:
        st.warning("Bu film için yeterli öneri bulunamadı.")
    else:
        for movie in result.index:
            st.write(movie)