import streamlit as st
import pandas as pd
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.neighbors import NearestNeighbors
from sklearn.decomposition import TruncatedSVD
import numpy as np
# Page config
st.set_page_config(
page_title="MusicMind - Smart Music Recommendations",
page_icon="🎵",
layout="wide"
)
# Custom CSS
st.markdown("""
""", unsafe_allow_html=True)
# Load and prepare data
@st.cache_data
def load_data():
df = pd.read_csv("song_dataset.csv")
return df
df = load_data()
@st.cache_resource
def run_imps(df):
required_columns = ['user', 'song', 'play_count', 'title', 'artist_name', 'release']
if not all(col in df.columns for col in required_columns):
raise ValueError(f"Dataset must contain the following columns: {required_columns}")
df = df.drop_duplicates(subset=['song', 'title', 'artist_name', 'release'])
df['combined_features'] = (df['title'] + " " + df['artist_name'] + " " + df['release']).fillna("")
# Content-Based Filtering
tfidf = TfidfVectorizer(max_features=5000, stop_words='english')
tfidf_matrix = tfidf.fit_transform(df['combined_features'])
nn = NearestNeighbors(n_neighbors=10, metric='cosine', algorithm='auto')
nn.fit(tfidf_matrix)
# Collaborative Filtering
user_song_matrix = df.pivot_table(index='user', columns='song', values='play_count', fill_value=0)
svd = TruncatedSVD(n_components=20)
user_factors = svd.fit_transform(user_song_matrix)
song_factors = svd.components_.T
return df, tfidf, tfidf_matrix, nn, user_song_matrix, user_factors, song_factors
df, tfidf, tfidf_matrix, nn, user_song_matrix, user_factors, song_factors = run_imps(df)
# Content-based recommendation function
def content_based_recommend(song_title, top_n=5):
try:
idx = df[df['title'] == song_title].index[0]
distances, indices = nn.kneighbors(tfidf_matrix[idx], n_neighbors=top_n + 1)
song_indices = indices.flatten()[1:]
return df.iloc[song_indices][['title', 'artist_name', 'release']].drop_duplicates()
except IndexError:
return pd.DataFrame(columns=['title', 'artist_name', 'release'])
def collaborative_recommend(user_id, top_n=5):
if user_id not in user_song_matrix.index:
return pd.DataFrame(columns=['title', 'artist_name', 'release'])
user_vector = user_factors[user_song_matrix.index.get_loc(user_id)]
scores = np.dot(song_factors, user_vector)
listened_songs = user_song_matrix.loc[user_id][user_song_matrix.loc[user_id] > 0].index
scores = {song: score for song, score in zip(user_song_matrix.columns, scores) if song not in listened_songs}
recommended_songs = sorted(scores.items(), key=lambda x: x[1], reverse=True)[:top_n]
recommended_song_ids = [song for song, _ in recommended_songs]
return df[df['song'].isin(recommended_song_ids)][['title', 'artist_name', 'release']].drop_duplicates()
# Hybrid Recommendation
def hybrid_recommendv2(user_id, song_titles, top_n=5):
collab_recs = collaborative_recommend(user_id, top_n)
content_recs = pd.DataFrame()
for song_title in song_titles:
content_recs = pd.concat([content_recs, content_based_recommend(song_title, top_n)], ignore_index=True)
hybrid_recs = pd.concat([collab_recs, content_recs]).drop_duplicates().sample(frac=1).reset_index(drop=True)
return hybrid_recs.head(top_n)
# Sidebar and Main UI
with st.sidebar:
st.header("🎯 Customize Your Recommendations")
user_id = st.selectbox(
"Select User ID",
options=df['user'].unique(),
index=0
)
user_songs = df[df['user'] == user_id]['title'].unique()
song_title = st.multiselect(
"Select Songs You Like",
options=user_songs,
default=user_songs[:1] if len(user_songs) > 0 else None
)
top_n = st.slider("Number of Recommendations", min_value=1, max_value=10, value=5)
get_recs = st.button("Get Recommendations! 🎶")
if get_recs:
st.header("🎵 Your Recommendations")
recommendations = hybrid_recommendv2(user_id, song_title, top_n)
if recommendations.empty:
st.error("No recommendations found. Try selecting different songs or users.")
else:
st.balloons()
for idx, row in recommendations.iterrows():
youtube_link = f"https://www.youtube.com/results?search_query={row['title']}+{row['artist_name']}"
st.markdown(f"""
{row['title']}
Artist: {row['artist_name']}
Album: {row['release']}
Watch on YouTube
""", unsafe_allow_html=True)