Advanced examples (#986)

* Created a testing suite for the spotipy client.Spotify class

* Delete pyvenv.cfg

* Corrected error in description

* Added some advanced usage examples, some of which involve utilizing basic data analysis methods using other libraries such as Pandas

* Minor formatting fixes

---------

Co-authored-by: Stéphane Bruckert <stephane.bruckert@gmail.com>

CHANGELOG.md

@@ -16,6 +16,7 @@ Add your changes below.
 - Added test_artist_id, test_artist_url, and test_artists_mixed_ids to non_user_endpoints test.py
 - Added rate/request limit to FAQ
 - Added custom `urllib3.Retry` class for printing a warning when a rate/request limit is reached.
+- Added `personalized_playlist.py`, `track_recommendations.py`, and `audio_features_analysis.py` to `/examples`.
 
 ### Fixed
 - Audiobook integration tests

examples/audio_features_analysis.py

@@ -0,0 +1,33 @@
+import spotipy
+from spotipy.oauth2 import SpotifyClientCredentials
+
+# Import the extra necessary libraries for this example
+# These libraries are not included in the default packages
+import pandas as pd
+import matplotlib.pyplot as plt
+import seaborn as sns
+
+# Set up Spotify credentials
+client_credentials_manager = SpotifyClientCredentials()
+sp = spotipy.Spotify(client_credentials_manager=client_credentials_manager)
+
+# Fetch audio features of tracks from any playlist
+playlist_id = '37i9dQZEVXbMDoHDwVN2tF'
+results = sp.playlist_tracks(playlist_id)
+tracks = results['items']
+track_ids = [track['track']['id'] for track in tracks]
+audio_features = sp.audio_features(track_ids)
+
+# Create a DataFrame of audio features
+df = pd.DataFrame(audio_features)
+df = df[['danceability', 'energy', 'speechiness', 'acousticness',
+         'instrumentalness', 'liveness', 'valence', 'tempo']]
+
+# Generate a correlation matrix
+correlation_matrix = df.corr()
+
+# Plot the correlation matrix using seaborn
+sns.heatmap(correlation_matrix, annot=True, cmap='coolwarm')
+
+# Show the plot
+plt.show()

examples/personalized_playlist.py

@@ -0,0 +1,53 @@
+import spotipy
+from spotipy.oauth2 import SpotifyOAuth
+
+# Import the extra necessary libraries for this example
+# These libraries are not included in the default packages
+import pandas as pd
+from sklearn.cluster import KMeans
+
+# Set up Spotify credentials
+sp = spotipy.Spotify(auth_manager=SpotifyOAuth(
+    client_id="YOUR_APP_CLIENT_ID",
+    client_secret="YOUR_APP_CLIENT_SECRET",
+    redirect_uri="YOUR_APP_REDIRECT_URI",
+    scope="playlist-modify-private,user-library-read"))
+
+
+# get the user's username
+username = sp.me()['id']
+
+# Get the user's liked tracks
+saved_tracks = sp.current_user_saved_tracks(limit=50)['items']
+
+# Extract audio features for liked tracks
+track_ids = []
+audio_features = []
+for track in saved_tracks:
+    track_ids.append(track['track']['id'])
+    audio_features.append(sp.audio_features(track['track']['id'])[0])
+
+# Create a DataFrame from the audio features
+df = pd.DataFrame(audio_features)
+
+# Perform clustering on some audio features
+features = df[['danceability', 'energy', 'valence', 'acousticness']]
+kmeans = KMeans(n_clusters=5, random_state=42, n_init=10)
+df['cluster'] = kmeans.fit_predict(features)
+
+# Select a representative track from each cluster
+representative_tracks = []
+for cluster in range(5):
+    cluster_tracks = df[df['cluster'] == cluster]
+    representative_track = cluster_tracks.iloc[0]['id']
+    representative_tracks.append(representative_track)
+
+# Create a playlist with the representative tracks
+playlist = sp.user_playlist_create(
+    user=username, name='Personalized Playlist', public=False)
+sp.playlist_add_items(
+    playlist_id=playlist['id'], items=representative_tracks)
+
+# Print the URL of the created playlist
+print("Playlist created successfully. You can access it at:",
+      playlist['external_urls']['spotify'])

examples/track_recommendations.py

@@ -0,0 +1,45 @@
+import spotipy
+from spotipy.oauth2 import SpotifyOAuth
+import random
+
+# Set up Spotify credentials
+sp = spotipy.Spotify(auth_manager=SpotifyOAuth(
+    client_id="YOUR_APP_CLIENT_ID",
+    client_secret="YOUR_APP_CLIENT_SECRET",
+    redirect_uri="YOUR_APP_REDIRECT_URI",
+    scope="user-top-read,user-library-read,user-read-recently-played"))
+
+# Get the user's top tracks
+top_tracks = sp.current_user_top_tracks(limit=10, time_range='long_term')
+
+# Get the user's liked tracks
+liked_tracks = sp.current_user_saved_tracks(limit=10)
+
+# Get the user's listening history
+history = sp.current_user_recently_played(limit=10)
+
+# Extract a list of the top track IDs
+top_track_ids = [track['id'] for track in top_tracks['items']]
+
+# Extract a list of the liked track IDs
+liked_track_ids = [track['track']['id'] for track in liked_tracks['items']]
+
+# Extract a list of the history track IDs
+history_track_ids = [track['track']['id'] for track in history['items']]
+
+# Combine the three lists and shuffle them randomly
+seed_track_ids = top_track_ids + liked_track_ids + history_track_ids
+random.shuffle(seed_track_ids)
+
+# Use the IDs to get some recommendations
+# Note: the seed_tracks parameter can accept up to 5 tracks
+recommendations = sp.recommendations(
+    seed_tracks=seed_track_ids[0:5], limit=10, country='US')
+
+# Display the recommendations
+for i, track in enumerate(recommendations['tracks']):
+    print(
+        "{}. {} by {}"
+        .format(i+1, track['name'], ', '
+                .join([artist['name'] for artist in track['artists']]))
+    )