import streamlit as st
import numpy as np
from tensorflow.keras.models import load_model
from tensorflow.keras.preprocessing.sequence import pad_sequences
import joblib
import pandas as pd
from sklearn.neighbors import NearestNeighbors
from sklearn.preprocessing import StandardScaler

# Load the LSTM model for emotion prediction
emotion_model = load_model('lstm_model.h5')

# Load the KNN model
knn_model = joblib.load('knn_model.joblib')

# Load the tokenizer
tokenizer = joblib.load('tokenizer.pkl')

# Load the dataset
df = pd.read_csv('df1.csv')

# Preprocess for KNN
audio_feature_columns = ['danceability', 'energy', 'key', 'loudness', 'mode', 'speechiness',
       'acousticness', 'instrumentalness', 'liveness', 'valence', 'tempo',
       'duration_ms', 'time_signature']

audio_features = df[audio_feature_columns]
mood_cats = df[['mood_cats']]
mood_cats_df = pd.DataFrame(mood_cats)

# Normalize audio features for KNN
scaler_knn = StandardScaler()
audio_features_scaled_knn = scaler_knn.fit_transform(audio_features)
audio_features_df_knn = pd.DataFrame(audio_features_scaled_knn, columns=audio_feature_columns)
combined_features_knn = pd.concat([mood_cats_df, audio_features_df_knn], axis=1)

# Function for KNN-based recommendation
def recommend_knn(query_index, n_recommendations=5):
    distances, indices = knn_model.kneighbors(combined_features_knn.iloc[query_index].values.reshape(1, -1), n_neighbors=n_recommendations)
    recommended_songs = df.iloc[indices.flatten()].copy()
    # Convert distances to scores
    recommended_songs['score'] = 1 / (1 + distances.flatten())  # Inverse of distance
    return recommended_songs

# Set up the title of the app
st.title('KNN Recommender App')

# Get song index from user input
song_index_to_recommend = st.number_input('Enter song index:', min_value=0, max_value=len(df)-1, value=0)

# Get lyrics for emotion prediction
lyrics = df.iloc[song_index_to_recommend]['lyrics']

# Process the lyrics
sequence = tokenizer.texts_to_sequences([lyrics])
padded_sequence = pad_sequences(sequence, maxlen=50)  # Adjust the maxlen to match the expected input size
emotion = emotion_model.predict(padded_sequence).flatten()

# Combine emotion and audio features for recommendation
combined_features = np.concatenate([emotion, audio_features_scaled_knn[song_index_to_recommend]])

# Get KNN-based recommendations
knn_recs = recommend_knn(song_index_to_recommend)

# Display the predicted emotion and recommendations
st.write(f"Predicted Emotion: {emotion}")

if not knn_recs.empty:
    st.write("KNN Recommendations:")
    for index, row in knn_recs.iterrows():
        st.write(f"Song Index: {index}, Title: {row['title']}, Artist: {row['artist']}, Score: {row['score']}")
else:
    st.write("No KNN Recommendations found."