Generate genre-based melodies using Magenta

In [2]:

    

import os
import shutil
import spotipy
import pickle
import pandas as pd
import numpy as np
%matplotlib notebook
import matplotlib.pyplot as plt
import seaborn as sns

from sklearn.manifold import TSNE
from sklearn.decomposition import PCA
from collections import Counter


if not os.path.exists('clean_midi'):
    # Download the 'Clean MIDI' dataset from http://colinraffel.com/projects/lmd/
    from six.moves import urllib
    import StringIO
    import gzip
    import tarfile
    FILE_URL = 'http://hog.ee.columbia.edu/craffel/lmd/clean_midi.tar.gz'
    response = urllib.request.urlopen(FILE_URL)
    print("INFO: Downloaded {}".format(FILE_URL))
    compressedFile = StringIO.StringIO()
    compressedFile.write(response.read())
    compressedFile.seek(0)
    decompressedFile = gzip.GzipFile(fileobj=compressedFile, mode='rb')
    OUTFILE_PATH = 'clean_midi.tar'
    with open(OUTFILE_PATH, 'wb') as outfile:
        outfile.write(decompressedFile.read())
    tar = tarfile.open(OUTFILE_PATH)
    tar.extractall()
    tar.close()
    print("INFO: Extracted data")
else:
    print("INFO: Found `clean_midi` directory")


    

In [ ]:

    

if not os.path.exists("genres.p"):
    # Use Spotify's API to genre lookup. Login first and get your OAuth token: 
    # https://developer.spotify.com/web-api/search-item/
    # NOTE: Replace `AUTH` value with your token.
    AUTH = "ENTER-MY-AUTH-KEY"

    # Get artists from folder names
    artists = [item for item in os.listdir('clean_midi') if not item.startswith('.')]

    sp = spotipy.Spotify(auth=AUTH)
    genres = {}
    for i,artist in enumerate(artists):
        try:
            results = sp.search(q=artist, type='artist',limit=1)
            items = results['artists']['items']
            genre_list = items[0]['genres'] if len(items) else items['genres']
            genres[artist] = genre_list
            if i < 5:
                print("INFO: Preview {}/5".format(i + 1),
                      artist, genre_list[:5])
        except Exception as e:
            print("INFO: ", artist, "not included: ", e)

    # Save to pickle file
    pickle.dump(genres,open("genres.p","wb"))
else:
    # Load genres meta-data
    genres = pickle.load(open("genres.p","rb"))


    

In [13]:

    

# Get the most common genres
flattened_list = [item for sublist in list(genres.values()) for item in sublist]
c = Counter(flattened_list)
c.most_common()[:20]


    

    
Out[13]:





[('mellow gold', 385),
 ('soft rock', 373),
 ('rock', 355),
 ('classic rock', 289),
 ('album rock', 275),
 ('folk rock', 262),
 ('new wave pop', 247),
 ('adult standards', 235),
 ('brill building pop', 225),
 ('bubblegum pop', 208),
 ('christmas', 191),
 ('disco', 190),
 ('motown', 187),
 ('classic funk rock', 182),
 ('rock-and-roll', 175),
 ('singer-songwriter', 172),
 ('hard rock', 169),
 ('pop rock', 169),
 ('quiet storm', 169),
 ('soul', 167)]

In [337]:

    

# Convert labels to vectors
categories = set(sorted(list(flattened_list)))
df = pd.DataFrame(columns=categories)

for author,genre_list in genres.items():
    row = pd.Series(np.zeros(len(categories)),name=author)
    for ind,genre in enumerate(categories):
        if genre in genre_list:
            row[ind] = 1
    d = pd.DataFrame(row).T
    d.columns = categories
    df = pd.concat([df,d])
df = df.reindex_axis(sorted(df.columns), axis=1)

# Assign label for each author corresponding with meta-genre (eg, rock, classical)
def getStyle(genre_substring):
    """Get data where features contain `genre_substring`."""
    style_index = np.asarray([genre_substring in x for x in df.columns])
    style_array = df.iloc[:,style_index].any(axis=1)
    return style_array

# Create array of color/labels
color_array = np.zeros(df.shape[0])
genre_labels = ['other','rock','metal','pop','mellow','country', 'rap','classical']
for i,g in enumerate(genre_labels):
    if g == 'other':
        pass
    else:
        color_array[np.where(getStyle(g))] = i


    

In [175]:

    

# 2-dimensions
model = TSNE(random_state=0)
np.set_printoptions(suppress=True)
X_tsne = model.fit_transform(df.values)
X_pca = PCA().fit_transform(df.values)

# 3-dimensions
model3d = TSNE(n_components=3, random_state=0)
np.set_printoptions(suppress=True)
X_tsne3d = model3d.fit_transform(df.values)
X_pca3d = PCA(n_components=3).fit_transform(df.values)


    

In [345]:

    

%pylab inline
CMAP_NAME = 'Set1'
cmap = matplotlib.cm.get_cmap(CMAP_NAME,lut=max(color_array)+1)
figure(figsize=(10, 5))
suptitle('Artist-genre visualization (2-D)')
axes_tsne = []
axes_pca = []

# TSNE
subplot(121)
title('TSNE')
for l in set(color_array):
    ax = plt.scatter(X_tsne[color_array==l][:, 0], X_tsne[color_array==l][:, 1],c=cmap(l/max(color_array)), s=5)
    axes_tsne.append(ax)
legend(handles=axes_tsne, labels=genre_labels,frameon=True,markerscale=2)

# PCA
subplot(122)
title('PCA')
for l in set(color_array):
    ax = plt.scatter(X_pca[color_array==l][:, 0], X_pca[color_array==l][:, 1],c=cmap(l/max(color_array)), s=5)
    axes_pca.append(ax)

legend(handles=axes_pca, labels=genre_labels,frameon=True,markerscale=2)
plt.show()


    

    




Populating the interactive namespace from numpy and matplotlib






    

In [348]:

    

from mpl_toolkits.mplot3d import Axes3D
fig = plt.figure(figsize=(10, 5))
suptitle('Artist-genre visualization (3-D)')
ax = fig.add_subplot(121, projection='3d')
title('TSNE')
axes_tsne = []
axes_pca = []
scatter_proxies = []
for l in set(color_array):
    ax.scatter(X_tsne3d[color_array==l][:, 0], X_tsne3d[color_array==l][:, 1],X_tsne3d[color_array==l][:, 2],c=cmap(l/max(color_array)), s=5,zdir='x')
    axes_tsne.append(fig.gca())
    proxy = matplotlib.lines.Line2D([0],[0], linestyle="none", c=cmap(l/max(color_array)), marker = 'o')
    scatter_proxies.append(proxy)                                

legend(handles=scatter_proxies, labels=genre_labels,numpoints=1,frameon=True,markerscale=0.8)
ax2 = fig.add_subplot(122, projection='3d')
title('PCA')
for l in set(color_array):        
    ax2.scatter(X_pca3d[color_array==l][:, 0], X_pca3d[color_array==l][:, 1], X_pca3d[color_array==l][:, 2],c=cmap(l/max(color_array)),s=5)
    axes_pca.append(fig.gca())
    
legend(handles=scatter_proxies, labels=genre_labels,numpoints=1,frameon=True,markerscale=0.8)
plt.show()


    

In [5]:

    

MIDI_DIR = os.path.join(os.getcwd(),'clean_midi')

def get_artists(genre):
    """Get artists with label `genre`."""
    artists = [artist for artist, gs in genres.items() if genre in gs]
    return artists

# Get artist with genres 'soft rock' and 'disco'
genre_data = {}
metal = get_artists('metal')
classical = get_artists('classical')

genre_data['metal'] = metal
genre_data['classical'] = classical

# Copy artists to a genre-specific folder
for genre, artists in genre_data.items():
    try:
        for artist in artists:
            _genre = genre.replace(' ','_').replace('&','n')
            shutil.copytree(os.path.join(MIDI_DIR,artist),os.path.join(os.getcwd(),'subsets',_genre,artist))
    except Exception as e:
        print(e)