├── .gitignore
├── README.md
├── app.py
├── requirements.txt
├── src
    ├── extract_metadata.py
    ├── feature_extraction.py
    ├── playlist_metadata.py
    ├── predict_mood.py
    └── train_model.py
└── templates
    └── index.html


/.gitignore:
--------------------------------------------------------------------------------
 1 | src/__pycache__
 2 | 
 3 | data/labeled/labels_with_artists.csv
 4 | data/features.csv
 5 | data/music_mood_model.pkl
 6 | 
 7 | static
 8 | data/labeled/happy
 9 | data/labeled/romantic
10 | data/labeled/sad
11 | venv


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
1 | Music Mood Playlist Filter
2 | Project Overview
3 | The Music Mood Playlist Filter is an innovative tool that allows users to filter and curate music playlists based on advanced auditory features such as pitch and artist. This solution extends beyond traditional playlist filtering by enabling users to refine their music experience based on the specific pitch of songs, offering a unique way to explore and enjoy music.


--------------------------------------------------------------------------------
/app.py:
--------------------------------------------------------------------------------
  1 | from flask import Flask, render_template, request, redirect, url_for
  2 | import os
  3 | import pandas as pd
  4 | import urllib.parse
  5 | import subprocess
  6 | 
  7 | 
  8 | app = Flask(__name__)
  9 | app.config['UPLOAD_FOLDER'] = 'static/uploads'
 10 | app.config['MAX_CONTENT_LENGTH'] = 16 * 1024 * 1024  # 16 MB limit
 11 | 
 12 | if not os.path.exists(app.config['UPLOAD_FOLDER']):
 13 |     os.makedirs(app.config['UPLOAD_FOLDER'])
 14 | 
 15 | # Run the metadata extraction script
 16 | def run_metadata_script():
 17 |     script_path = 'I:/Music-classifier/src/playlist_metadata.py'  # Adjust this path
 18 |     subprocess.run(['venv/Scripts/python', script_path], check=True)
 19 | 
 20 | # Load artist names from the CSV, ensuring unique and individual artist names
 21 | def get_artists():
 22 |     df = pd.read_csv('./static/labels_with_playlist_artists.csv')
 23 |     unique_artists = sorted(set(artist.strip() for artists in df['artist'] for artist in artists.split(',')))
 24 |     return unique_artists
 25 | 
 26 | @app.route('/')
 27 | def index():
 28 |     files = os.listdir(app.config['UPLOAD_FOLDER'])
 29 |     artists = get_artists()  # Get the list of unique artists
 30 |     return render_template('index.html', files=files, artists=artists)
 31 | 
 32 | @app.route('/upload', methods=['POST'])
 33 | def upload_file():
 34 |     if 'file' not in request.files:
 35 |         return redirect(request.url)
 36 |     file = request.files['file']
 37 |     if file.filename == '':
 38 |         return redirect(request.url)
 39 |     if file:
 40 |         filename = file.filename
 41 |         file_path = os.path.join(app.config['UPLOAD_FOLDER'], filename)
 42 |         file.save(file_path)
 43 |         
 44 |         # Run the script after saving the file
 45 |         run_metadata_script()
 46 |         refresh_song_data()
 47 | 
 48 |         return redirect(url_for('index'))
 49 | 
 50 | @app.route('/remove/<filename>', methods=['POST'])
 51 | def remove_file(filename):
 52 |     decoded_filename = urllib.parse.unquote(filename)
 53 |     file_path = os.path.join(app.config['UPLOAD_FOLDER'], decoded_filename)
 54 |     if os.path.exists(file_path):
 55 |         os.remove(file_path)
 56 |         
 57 |         # Run the script after removing the file
 58 |         run_metadata_script()
 59 |         refresh_song_data()
 60 | 
 61 |     return redirect(url_for('index'))
 62 | 
 63 | song_data = pd.read_csv('I:/Music-classifier/static/labels_with_playlist_artists.csv')
 64 | 
 65 | def refresh_song_data():
 66 |     global song_data
 67 |     song_data = pd.read_csv('I:/Music-classifier/static/labels_with_playlist_artists.csv')
 68 | 
 69 | @app.route('/apply_filters', methods=['POST'])
 70 | def apply_filters():
 71 |     # Get selected moods and singers from the form
 72 |     selected_moods = request.form.getlist('emotion')
 73 |     selected_singers = request.form.getlist('singer')
 74 |     
 75 |     # Filter the DataFrame based on the selected options
 76 |     filtered_data = song_data.copy()
 77 |     
 78 |     # Apply both filters if selections are made
 79 |     if selected_moods and selected_singers:
 80 |         filtered_data = filtered_data[
 81 |             (filtered_data['mood'].isin(selected_moods)) &
 82 |             (filtered_data['artist'].apply(lambda artists: any(singer in artists for singer in selected_singers)))
 83 |         ]
 84 |     elif selected_moods:  # If only moods are selected
 85 |         filtered_data = filtered_data[filtered_data['mood'].isin(selected_moods)]
 86 |     elif selected_singers:  # If only singers are selected
 87 |         filtered_data = filtered_data[filtered_data['artist'].apply(lambda artists: any(singer in artists for singer in selected_singers))]
 88 | 
 89 | 
 90 |     
 91 |     # Extract only the filenames of the filtered songs
 92 |     filtered_files = filtered_data['filename'].tolist()
 93 |     
 94 |     # Reuse the index view but with filtered files
 95 |     artists = get_artists()  # Get the list of unique artists again for the filters
 96 |     return render_template('index.html', files=filtered_files, artists=artists)
 97 | 
 98 | if __name__ == '__main__':
 99 |     app.run(debug=True)
100 | 


--------------------------------------------------------------------------------
/requirements.txt:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Midhun31/Music-classifier-Machine-Learning/89427a87862583a9a7ad6b402ed0122e94d68459/requirements.txt


--------------------------------------------------------------------------------
/src/extract_metadata.py:
--------------------------------------------------------------------------------
 1 | import os
 2 | from mutagen.id3 import ID3
 3 | 
 4 | def extract_artist(file_path):
 5 |     try:
 6 |         audio = ID3(file_path)
 7 |         artist_tag = audio.get('TPE1')  # 'TPE1' is the ID3 tag for artist
 8 |         artist = artist_tag.text[0] if artist_tag else 'Unknown Artist'
 9 |         if ' -' in artist:
10 |             artist = artist.split(' -')[0]
11 |         return artist
12 |     except Exception as e:
13 |         return f"Error: {e}"
14 | 
15 | # Directory containing labeled audio files
16 | base_dir = 'data/labeled/'
17 | 
18 | # Initialize lists to store file names, their corresponding labels, and artists
19 | file_names = []
20 | labels = []
21 | artists = []
22 | 
23 | # Iterate over each mood directory
24 | for label in os.listdir(base_dir):
25 |     mood_dir = os.path.join(base_dir, label)
26 |     if os.path.isdir(mood_dir):
27 |         # Iterate over each audio file in the mood directory
28 |         for file_name in os.listdir(mood_dir):
29 |             if file_name.endswith('.mp3'):  # Adjust if using other formats
30 |                 file_path = os.path.join(mood_dir, file_name)
31 |                 file_names.append(file_name)
32 |                 labels.append(label)
33 | 
34 |                 # Extract artist's name
35 |                 artist = extract_artist(file_path)
36 |                 artists.append(artist)
37 | 
38 | # Create DataFrame and save to CSV
39 | import pandas as pd
40 | df = pd.DataFrame({
41 |     'filename': file_names,
42 |     'mood': labels,
43 |     'artist': artists
44 | })
45 | csv_file = os.path.join(base_dir, 'labels_with_artists.csv')
46 | df.to_csv(csv_file, index=False)
47 | 
48 | print(f"CSV file created at: {csv_file}")
49 | 


--------------------------------------------------------------------------------
/src/feature_extraction.py:
--------------------------------------------------------------------------------
 1 | import os
 2 | import librosa
 3 | import pandas as pd
 4 | import numpy as np
 5 | 
 6 | # Directory containing labeled audio files
 7 | base_dir = './data/labeled/'  # Adjust this path based on your actual directory structure
 8 | 
 9 | # Initialize lists to store features and labels
10 | features = []
11 | labels = []
12 | 
13 | # Iterate over each mood directory
14 | for label in os.listdir(base_dir):
15 |     mood_dir = os.path.join(base_dir, label)
16 |     if os.path.isdir(mood_dir):
17 |         # Iterate over each audio file in the mood directory
18 |         for file_name in os.listdir(mood_dir):
19 |             if file_name.endswith('.wav') or file_name.endswith('.mp3'):
20 |                 file_path = os.path.join(mood_dir, file_name)
21 |                 try:
22 |                     # Load the audio file
23 |                     y, sr = librosa.load(file_path, sr=None)
24 |                     
25 |                     # Extract features
26 |                     mfccs = np.mean(librosa.feature.mfcc(y=y, sr=sr, n_mfcc=13).T, axis=0)
27 |                     chroma = np.mean(librosa.feature.chroma_stft(y=y, sr=sr).T, axis=0)
28 |                     spectral_contrast = np.mean(librosa.feature.spectral_contrast(y=y, sr=sr).T, axis=0)
29 |                     zero_crossing_rate = np.mean(librosa.feature.zero_crossing_rate(y=y).T, axis=0)
30 |                     rms = np.mean(librosa.feature.rms(y=y).T, axis=0)
31 |                     spectral_rolloff = np.mean(librosa.feature.spectral_rolloff(y=y, sr=sr).T, axis=0)
32 |                     
33 |                     # Combine all features
34 |                     feature_vector = np.concatenate([
35 |                         mfccs,
36 |                         chroma,
37 |                         spectral_contrast,
38 |                         zero_crossing_rate,
39 |                         rms,
40 |                         spectral_rolloff
41 |                     ])
42 |                     
43 |                     # Append features and label
44 |                     features.append(feature_vector)
45 |                     labels.append(label)
46 |                 except Exception as e:
47 |                     print(f"Error processing {file_path}: {e}")
48 | 
49 | # Create a DataFrame from the features and labels
50 | columns = [
51 |     'mfcc_' + str(i) for i in range(13)
52 | ] + [
53 |     'chroma_' + str(i) for i in range(12)
54 | ] + [
55 |     'spectral_contrast_' + str(i) for i in range(7)
56 | ] + [
57 |     'zero_crossing_rate'
58 | ] + [
59 |     'rms'
60 | ] + [
61 |     'spectral_rolloff'
62 | ]
63 | 
64 | df = pd.DataFrame(features, columns=columns)
65 | df['mood'] = labels
66 | 
67 | # Save the DataFrame to a CSV file
68 | features_csv_file = './data/features.csv'
69 | df.to_csv(features_csv_file, index=False)
70 | 
71 | print(f"Features CSV file created at: {features_csv_file}")
72 | 


--------------------------------------------------------------------------------
/src/playlist_metadata.py:
--------------------------------------------------------------------------------
 1 | import os
 2 | from mutagen.id3 import ID3
 3 | import pandas as pd
 4 | from predict_mood import predict_mood
 5 | 
 6 | def extract_artist(file_path):
 7 |     try:
 8 |         audio = ID3(file_path)
 9 |         artist_tag = audio.get('TPE1')  # 'TPE1' is the ID3 tag for artist
10 |         artist = artist_tag.text[0] if artist_tag else 'Unknown Artist'
11 |         if ' -' in artist:
12 |             artist = artist.split(' -')[0]
13 |         return artist
14 |     except Exception as e:
15 |         return f"Error: {e}"
16 | 
17 | # Directory containing the audio files
18 | base_dir = 'I:/Music-classifier/static/uploads/'
19 | 
20 | # Initialize lists to store file names and their corresponding artists
21 | file_names = []
22 | moods=[]
23 | artists = []
24 | 
25 | # Iterate over each audio file in the base directory
26 | for file_name in os.listdir(base_dir):
27 |     if file_name.endswith('.mp3'):  # Adjust if using other formats
28 |         file_path = os.path.join(base_dir, file_name)
29 |         file_names.append(file_name)
30 | 
31 |         # Extract artist's name
32 |         artist = extract_artist(file_path)
33 |         artists.append(artist)
34 | 
35 |         # Extract mood
36 |         mood = predict_mood(file_path)
37 |         moods.append(mood)
38 | 
39 | # Create DataFrame and save to CSV (without the mood column)
40 | df = pd.DataFrame({
41 |     'filename': file_names,
42 |     'mood':moods,
43 |     'artist': artists
44 | })
45 | csv_file = os.path.join('I:/Music-classifier/static/', 'labels_with_playlist_artists.csv')
46 | df.to_csv(csv_file, index=False)
47 | 
48 | print(f"CSV file created at: {csv_file}")
49 | 


--------------------------------------------------------------------------------
/src/predict_mood.py:
--------------------------------------------------------------------------------
 1 | import librosa
 2 | import numpy as np
 3 | import joblib
 4 | 
 5 | # Load the trained model
 6 | model = joblib.load('I:/Music-classifier/data/music_mood_model.pkl')
 7 | 
 8 | # Function to extract features from a new audio file
 9 | def extract_features(file_path):
10 |     try:
11 |         # Load the audio file
12 |         y, sr = librosa.load(file_path, sr=None)
13 |         
14 |         # Extract features
15 |         mfccs = np.mean(librosa.feature.mfcc(y=y, sr=sr, n_mfcc=13).T, axis=0)
16 |         chroma = np.mean(librosa.feature.chroma_stft(y=y, sr=sr).T, axis=0)
17 |         spectral_contrast = np.mean(librosa.feature.spectral_contrast(y=y, sr=sr).T, axis=0)
18 |         zero_crossing_rate = np.mean(librosa.feature.zero_crossing_rate(y=y).T, axis=0)
19 |         rms = np.mean(librosa.feature.rms(y=y).T, axis=0)
20 |         spectral_rolloff = np.mean(librosa.feature.spectral_rolloff(y=y, sr=sr).T, axis=0)
21 |         
22 |         # Combine all features into a single feature vector
23 |         feature_vector = np.concatenate([
24 |             mfccs,
25 |             chroma,
26 |             spectral_contrast,
27 |             zero_crossing_rate,
28 |             rms,
29 |             spectral_rolloff
30 |         ])
31 |         
32 |         return feature_vector
33 |     except Exception as e:
34 |         print(f"Error extracting features from {file_path}: {e}")
35 |         return None
36 | 
37 | # Function to predict the mood of a new song
38 | def predict_mood(file_path):
39 |     # Extract features from the new song
40 |     features = extract_features(file_path)
41 |     
42 |     if features is not None:
43 |         # Reshape the features to match the model's expected input format
44 |         features = features.reshape(1, -1)
45 |         
46 |         # Predict the mood using the model
47 |         prediction = model.predict(features)
48 |         
49 |         # Print the predicted mood
50 |         return prediction[0]
51 |     else:
52 |         return "Could not predict mood due to feature extraction error."
53 | 
54 | # Provide the path to your new song here
55 | new_song_path = 'I:/Music-classifier/static/uploads/Aaluma-Doluma.mp3'
56 | mood=predict_mood(new_song_path)
57 | print(mood)


--------------------------------------------------------------------------------
/src/train_model.py:
--------------------------------------------------------------------------------
 1 | import pandas as pd
 2 | from sklearn.model_selection import train_test_split
 3 | from sklearn.ensemble import RandomForestClassifier
 4 | from sklearn.metrics import accuracy_score, classification_report
 5 | import joblib
 6 | 
 7 | # Load the features
 8 | data = pd.read_csv('I:/Music-classifier/data/features.csv')
 9 | 
10 | # Separate features and labels
11 | X = data.drop(columns=['mood'])
12 | y = data['mood']
13 | 
14 | # Split the data
15 | X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
16 | 
17 | # Train the model
18 | model = RandomForestClassifier(n_estimators=100, random_state=42)
19 | model.fit(X_train, y_train)
20 | 
21 | # Evaluate the model
22 | y_pred = model.predict(X_test)
23 | print(f'Accuracy: {accuracy_score(y_test, y_pred)}')
24 | print(classification_report(y_test, y_pred))
25 | 
26 | # Save the model
27 | joblib.dump(model, 'I:/Music-classifier/data/music_mood_model.pkl')
28 | 


--------------------------------------------------------------------------------
/templates/index.html:
--------------------------------------------------------------------------------
  1 | <!DOCTYPE html>
  2 | <html lang="en">
  3 | 
  4 | <head>
  5 |     <meta charset="UTF-8">
  6 |     <meta name="viewport" content="width=device-width, initial-scale=1.0">
  7 |     <title>My Playlist</title>
  8 |     <style>
  9 |         body {
 10 |             font-family: Arial, sans-serif;
 11 |             margin: 20px;
 12 |         }
 13 | 
 14 |         .container {
 15 |             max-width: 800px;
 16 |             margin: auto;
 17 |         }
 18 | 
 19 |         h1 {
 20 |             text-align: center;
 21 |         }
 22 | 
 23 |         .upload-section {
 24 |             margin-bottom: 20px;
 25 |             text-align: center;
 26 |         }
 27 | 
 28 |         .upload-section button {
 29 |             padding: 10px 20px;
 30 |             font-size: 16px;
 31 |         }
 32 | 
 33 |         .uploaded-files {
 34 |             margin-top: 20px;
 35 |         }
 36 | 
 37 |         .file-item {
 38 |             display: flex;
 39 |             align-items: center;
 40 |             justify-content: space-between;
 41 |             margin-bottom: 15px;
 42 |             padding: 10px;
 43 |             border: 1px solid #ccc;
 44 |             border-radius: 5px;
 45 |         }
 46 | 
 47 |         .serial-number {
 48 |             width: 40px;
 49 |             text-align: center;
 50 |             font-weight: bold;
 51 |         }
 52 | 
 53 |         .file-name {
 54 |             flex-grow: 1;
 55 |             margin-right: 20px;
 56 |             font-size: 16px;
 57 |         }
 58 | 
 59 |         .file-controls {
 60 |             display: flex;
 61 |             align-items: center;
 62 |         }
 63 | 
 64 |         .file-controls audio {
 65 |             margin-right: 20px;
 66 |         }
 67 | 
 68 |         .file-controls button {
 69 |             background-color: #ff4d4d;
 70 |             color: white;
 71 |             border: none;
 72 |             padding: 5px 10px;
 73 |             cursor: pointer;
 74 |             border-radius: 5px;
 75 |         }
 76 | 
 77 |         .file-controls button:hover {
 78 |             background-color: #ff1a1a;
 79 |         }
 80 | 
 81 |         /* Filter Section Styles */
 82 |         .filter-section {
 83 |             position: absolute;
 84 |             top: 20px;
 85 |             right: 20px;
 86 |             background-color: #f9f9f9;
 87 |             border: 1px solid #ccc;
 88 |             padding: 10px;
 89 |             border-radius: 5px;
 90 |             width: 200px;
 91 |             box-shadow: 0px 4px 8px rgba(0, 0, 0, 0.1);
 92 |         }
 93 | 
 94 |         .filter-section h3 {
 95 |             margin: 0 0 10px;
 96 |             font-size: 18px;
 97 |         }
 98 | 
 99 |         .filter-section .filter-group {
100 |             margin-bottom: 15px;
101 |         }
102 | 
103 |         .filter-section label {
104 |             display: block;
105 |             margin: 5px 0;
106 |         }
107 | 
108 |         .filter-section input[type="checkbox"] {
109 |             margin-right: 8px;
110 |         }
111 | 
112 |         .filter-section .select-all {
113 |             font-weight: bold;
114 |             margin-bottom: 5px;
115 |         }
116 |     </style>
117 |     <script>
118 |         function toggleSelectAll(groupName, selectAllCheckbox) {
119 |             const checkboxes = document.querySelectorAll(`input[name="${groupName}"]`);
120 |             checkboxes.forEach(checkbox => checkbox.checked = selectAllCheckbox.checked);
121 |         }
122 |     </script>
123 | </head>
124 | 
125 | <body>
126 |     <div class="container">
127 |         <h1>My Playlist</h1>
128 | 
129 |         <div class="upload-section">
130 |             <form action="/upload" method="post" enctype="multipart/form-data">
131 |                 <input type="file" name="file" accept="audio/*">
132 |                 <button type="submit">Add</button>
133 |                 <button id="playAllBtn" type="button">Play All</button>
134 |             </form>
135 |         </div>
136 | 
137 |         <!-- Filter Section -->
138 | <!-- Filter Section -->
139 | <div class="filter-section" style="margin-top: 20px; padding: 10px; border: 1px solid #ccc; border-radius: 5px;">
140 |     <h3 style="margin-bottom: 10px;">Filter By</h3>
141 | 
142 |     <!-- Filter Form -->
143 |     <form action="{{ url_for('apply_filters') }}" method="post" id="filter-form">
144 |         <!-- Emotions Filter -->
145 |         <div class="filter-group" style="margin-bottom: 15px;">
146 |             <strong>Emotions</strong>
147 |             <label class="select-all" style="display: block; margin-bottom: 5px;">
148 |                 <input type="checkbox" onclick="toggleSelectAll('emotion', this)"> Select All
149 |             </label>
150 |             <label style="display: block; margin-left: 10px;">
151 |                 <input type="checkbox" name="emotion" value="happy"> Happy
152 |             </label>
153 |             <label style="display: block; margin-left: 10px;">
154 |                 <input type="checkbox" name="emotion" value="romantic"> Romantic
155 |             </label>
156 |             <label style="display: block; margin-left: 10px;">
157 |                 <input type="checkbox" name="emotion" value="sad"> Sad
158 |             </label>
159 |         </div>
160 | 
161 |         <!-- Singer Filter -->
162 |         <div class="filter-group">
163 |             <strong>Singer</strong>
164 |             <label class="select-all" style="display: block; margin-bottom: 5px;">
165 |                 <input type="checkbox" onclick="toggleSelectAll('singer', this)"> Select All
166 |             </label>
167 |             <!-- Loop to display unique singers -->
168 |             {% for singer in artists %}
169 |             <label style="display: block; margin-left: 10px;">
170 |                 <input type="checkbox" name="singer" value="{{ singer }}">{{ singer }}
171 |             </label>
172 |             {% endfor %}
173 |         </div>
174 | 
175 |         <!-- Apply Filters Button -->
176 |         <button type="submit" style="margin-top: 15px; padding: 8px 15px; background-color: #4CAF50; color: white; border: none; border-radius: 5px; cursor: pointer;">
177 |             Apply Filters
178 |         </button>
179 |     </form>
180 | </div>
181 | 
182 | 
183 |         <div class="uploaded-files" id="playlist">
184 |             {% for file in files %}
185 |             <div class="file-item">
186 |                 <div class="serial-number">{{ loop.index }}.</div>
187 |                 <div class="file-name">{{ file }}</div>
188 |                 <div class="file-controls">
189 |                     <audio controls>
190 |                         <source src="{{ url_for('static', filename='uploads/' ~ file | urlencode) }}" type="audio/mpeg">
191 |                         Your browser does not support the audio element.
192 |                     </audio>
193 |                     <form action="{{ url_for('remove_file', filename=file | urlencode) }}" method="post" style="display: inline;">
194 |                         <button type="submit">Remove</button>
195 |                     </form>
196 |                 </div>
197 |             </div>
198 |             {% endfor %}
199 |         </div>
200 |     </div>
201 | 
202 |     <script>
203 |         document.getElementById('playAllBtn').addEventListener('click', function() {
204 |             const songs = document.querySelectorAll('#playlist audio');
205 |             let currentIndex = 0;
206 | 
207 |             function playNextSong() {
208 |                 if (currentIndex < songs.length) {
209 |                     songs[currentIndex].play();
210 |                     songs[currentIndex].addEventListener('ended', function() {
211 |                         currentIndex++;
212 |                         playNextSong(); // Play the next song when the current one ends
213 |                     });
214 |                 }
215 |             }
216 | 
217 |             // Start playing the first song
218 |             playNextSong();
219 |         });
220 |         const audioPlayers = document.querySelectorAll('#playlist audio');
221 | 
222 |         // Add event listeners to each audio element
223 |         audioPlayers.forEach(audio => {
224 |             audio.addEventListener('play', () => {
225 |                 // Pause all other audio elements except the one currently playing
226 |                 audioPlayers.forEach(otherAudio => {
227 |                     if (otherAudio !== audio) {
228 |                         otherAudio.pause();
229 |                     }
230 |                 });
231 |             });
232 |         });
233 |     </script>
234 | </body>
235 | 
236 | </html>
237 | 


--------------------------------------------------------------------------------