def main():
df = fill_song_pd()
lyrics = "" #going to be one huge string
db_name = './markov/' + genre
mc = MarkovChain(db_name)
#creating new markov dataset if it doesn't exist
if not os.path.isfile(db_name):
print("creating new data set based on the " + str(genre) + " genre...")
for index, row in df.iterrows():
if row['genre'] == genre_dict[genre]:
lyrics += row["lyrics"] + " "
mc.generateDatabase(lyrics)
mc.dumpdb()
for i in range(int(lines) + 1):
print(mc.generateString())
def generate_database():
"""
Generates the database that the Markov Chain will use to make its word-by-
word predictions. It will attempt to create this file in the same directory
as where the script is currently located.
"""
currpath = os.path.dirname(__file__)
path_to_data = os.path.join(currpath, 'in.txt')
chain = MarkovChain()
with open(path_to_data) as f:
chain.generateDatabase(f.read())
chain.dumpdb()
print(chain.generateString())
def main():
args = parser.parse_args()
dirname=os.path.split(__file__)[0]
filename=os.path.join(dirname,"phil.txt")
title_filename=os.path.join(dirname,"phil_titles.txt")
dbname1 = "database.pkl"
dbname2 = "database_title.pkl"
new_db = not os.path.exists(dbname1)
body_maker = MarkovChain(dbname1)
title_maker = MarkovChain(dbname2)
if new_db:
title_maker.generateDatabase(open(title_filename).read())
title_maker.dumpdb()
body_maker.generateDatabase(open(filename).read())
body_maker.dumpdb()
name = title_maker.generateString()
body = ' '.join([body_maker.generateString()+'.' for i in xrange(3)])
if args.repo:
if args.token:
token = args.token
else:
token_filename = os.path.join(dirname, "token.txt")
if not os.path.exists(token_filename):
sys.stderr.write("Please either specify --token=XXX on the command line or put a github API token in token.txt\n")
sys.stderr.write("You can generate a token here: https://github.com/settings/tokens\n")
sys.exit(1)
token = open(token_filename).read().strip()
import github
gh=github.Github(token)
user=gh.get_user()
repo=user.get_repo(args.repo)
issue = repo.create_issue(title=name, body=body)
print issue.html_url
else:
print
print name
print "-"*len(name)
print body
artist_name.lower().encode('utf-8')).hexdigest()
mc = MarkovChain(db_name_hashed)
# Checking if the database already exists, if so uses the cache instead another API call
if not os.path.isfile(db_name_hashed):
print(
"No data cached. Please be patient while we search the lyrics of %s."
% artist_name)
# Adding lyrics to a single gigant string
lyrics = ''
# Parsing each lyric from this artist.
# [http://api.wikia.com/wiki/LyricWiki_API]
artist = requests.get(API_URI, params=params).json()
for album in artist['albums']:
for song in album['songs']:
params = {'artist': artist_name, 'song': song}
print("Parsing \"{}\" from Wikia.".format(song))
response = requests.get(API_URI,
params=params).json()["lyrics"]
lyrics += response.replace('[...]', '') + ' '
# Generating the database
mc.generateDatabase(lyrics)
mc.dumpdb()
# Printing a string
for i in range(0, int(number_of_phrases)):
print(mc.generateString())
if not os.path.isfile(DB_FILE):
# Handle common user errors
if not os.path.isfile(SOURCE_FILE):
if os.path.isfile(DB_FILE + '.7z'):
sys.exit("NOTICE: Please extract the archive containing the Markov database before use.")
sys.exit("NOTICE: You can't regenerate the Markov database without the source text.");
# Moving this in here avoids an annoying warning message if either of the
# above two sys.exit() calls would be triggered
mc = MarkovChain(DB_FILE)
# Generate the database
with open(SOURCE_FILE, 'r') as f:
mc.generateDatabase(f.read(), sentenceSep='[.!?"\n]', n=2)
mc.dumpdb()
else:
mc = MarkovChain(DB_FILE)
def generate_string(max_length):
# Generate the string
# We could be a bit smarter about this, but it works fairly well
gen_string = ''
short_counter = 0
while len(gen_string) < max_length:
new_str = mc.generateString().strip()
new_str = re.sub(r' , ?', ', ', new_str)
# Too short or too long to be meaningful
if len(new_str) < 4 or len(new_str) > 100:
continue
if "//" in t:
continue
if "cw: " in t:
continue
# Prune short tags
if ARGS.prune and len(t) <= 3:
continue
# Tags which are just numbers should not be in the corpus
try:
int(t.strip())
continue
except ValueError:
pass
if ARGS.nohash:
CORPUS += t + " "
else:
CORPUS += '#' + t + " "
CORPUS += "\n"
if ARGS.debug:
print(CORPUS)
exit(1)
print("Generating database...")
BOT = MarkovChain(TARGET_FILE)
BOT.generateDatabase(CORPUS)
print("Dumping database to {}".format(TARGET_FILE))
BOT.dumpdb()
