artist = song_arr[i][0]
track = song_arr[i][1]
uri = spot_client.search_song(artist, track)
if uri is None:
print(f"No song found for {artist} {track}")
else:
uri_list.append(uri)
print(f"Found {artist} {track}")
found_counter += 1
print(f"Found {found_counter} out of {len(song_arr)} songs on spotify")
# Make request to add 100 songs to playlist (max)
request_lists = []
start_i = 0
end_i = 100
for i in range(len(uri_list) // 100):
request_lists.append(uri_list[start_i:end_i])
start_i += 100
end_i += 100
end_i -= 100
end_i += len(uri_list) % 100
request_lists.append(uri_list[start_i:end_i])
for request in request_lists:
spot_client.add_songs(spot_playlist_id, request)
print("Songs added")
from spotify_client import SpotifyClient
# from looking at the KDE plots in plots.py, the darkest regions were in these bounds:
max_energy = 0.8
min_energy = 0.6
max_valence = 0.63
min_valence = 0.47
max_tempo = 133
min_tempo = 77
new_songs = []
# now we want to go through our dataframe of liked songs and get our "ideal songs":
# songs which fall into the bounds of the KDE plot highest density regions.
name = "audio_features.csv"
ideal_songs = DataAnalysis.ideal_song(pd, name, max_energy, max_valence,
max_tempo, min_energy, min_valence,
min_tempo)
print("Obtained ideal songs to generate your recommendations successfully.")
# now, using this list of ideal songs, we need to get our list of recommendations:
# we will loop through so that we can get recommendations from all of our ideal songs:
SpotifyClient.recommended_songs(sp, ideal_songs, new_songs, max_energy,
max_valence, max_tempo, min_energy,
min_valence, min_tempo)
print("Found recommendations for you successfully.")
# create the playlist to put our recommended songs into
playlist = SpotifyClient.create_playlist(sp)
# add the recommended songs to our new playlist
SpotifyClient.add_songs(sp, playlist, new_songs)