self.genres = genres
self.styles = styles
# Updates a song with discogs artists, labels, genres, and styles information
def update(self, artists, labels, genres, styles):
self.atists = artists
self.labels = labels
self.genres = genres
self.styles = styles
if scrape_chart_data:
# Grab charts from 1/1/2000 to 1/1/2012 as training data
start_date = datetime.date(2000, 1, 1)
#end_date = datetime.date(2000, 1, 14)
end_date = datetime.date(2012,1,1)
charts = chart_scraper.get_charts('hot-100', start_date, end_date)
print('Charts returned')
# Separate out all unique songs by combining all songs with same title and artist
ce_to_songs = dict()
for chart in charts:
for entry in chart:
entry_string = entry.title + ' - ' + entry.artist
# Deal with Featuring
index = entry.artist.find('Featuring')
artist = entry.artist
if index != -1:
artist = entry.artist[:index-1]
ce_to_songs[entry_string] = Song(entry.title, [artist], entry.weeks, entry.peakPos)
pickle.dump(ce_to_songs, open('chart-data/chart_songs_train.pickle', 'wb'))
# Get and save raw charts from billboard
import pickle
import datetime
import chart_scraper
'''
hot-100: top 100 songs
billboard-200: top 200 albums
r-b-hip-hop-songs: top 25
pop-songs: top 20
country-songs: top 25
rock-songs: top 25
dance-electronic-songs: top 25
latin-songs: top 25
christian-songs: top 25
'''
# Note: some charts do not have data for entire range of dates
chart_names = ['hot-100', 'billboard-200', 'r-b-hip-hop-songs', 'pop-songs', 'latin-songs', \
'country-songs', 'rock-songs', 'dance-electronic-songs', 'christian-songs']
for chart_name in chart_names:
# Grab charts from 1/1/2000 to 11/1/2015
start_date = datetime.date(2000, 1, 1)
end_date = datetime.date(2015, 11, 1)
charts = chart_scraper.get_charts(chart_name, start_date, end_date)
print('Charts returned for ' + chart_name)
pickle.dump(charts, open('chart-data/' + chart_name + '_charts.pickle', 'wb'))
