class TemplateBank:
def __init__(self, title_bank):
self.markov = MarkovChain(3)
for item in title_bank.values():
self.markov.add(item['title'].replace('—', '-'))
def _random_template(self):
title = self.markov.generate()
replacements = {}
tokens = []
doc = nlp(title)
i = 0
for token in doc:
# Consider named entities as single token.
if token.ent_type_ in ('PERSON', 'FAC', 'GPE', 'LOC'):
if token.ent_iob == 1:
tokens[-2] += tokens[-1] + token.text
tokens[-1] = token.whitespace_
else:
tokens.append(token.text)
tokens.append(token.whitespace_)
replacements[i] = '[[PERSON]]' if token.ent_type_ == 'PERSON' else '[[LOC]]'
i += 2
continue
tokens.append(token.text)
tokens.append(token.whitespace_)
if token.tag_ in ("NN", "NNP"):
replacements[i] = "[[NOUN]]"
elif token.tag_ in ("NNS", "NNPS"):
replacements[i] = "[[NOUNS]]"
elif token.pos_ == "ADJ":
replacements[i] = "[[ADJ]]"
i += 2
if len(replacements) < 2:
return None
logger.debug('generated title: ' + ''.join(tokens))
# Create a template by replacing two random tokens with POS tags
for i, replacement in random.sample(replacements.items(), 2):
tokens[i] = replacement
logger.debug('generated template: ' + ''.join(tokens))
return tokens
def random_template(self):
"""Get random template from the bank."""
for i in range(0, 25):
template = self._random_template()
if template is not None:
return template
raise RecursionError("Title generation was unable to find fitting template.")
class MidiParser:
def __init__(self, filename):
self.filename = filename
self.ticks = None
self.tempo = get_tempo(filename)
self.chain = MarkovChain()
self.parse()
def parse(self):
"""
reads midi and makes markov insertions
"""
midi = mido.MidiFile(self.filename)
self.ticks = midi.ticks_per_beat
prev = []
curr = []
for track in midi.tracks:
for msg in track:
if msg.type is 'note_on':
if msg.time is 0:
curr.append(msg.note)
else:
self.make_transition(prev, curr, msg.time)
prev = curr
curr = []
def make_transition(self, prev, curr, time):
"""
inserts all state transition possibilities into markov
"""
print(prev, curr, time)
for i in prev:
for j in curr:
self.chain.add(i, j, self.millisecond(time))
def millisecond(self, ticks):
"""
tick to millisecond converter
"""
try:
ms = ((ticks / self.ticks) * self.tempo) / 1000
return int(ms - (ms % 250) + 250)
except TypeError:
raise TypeError(
"Could not read a tempo and ticks_per_beat from midi")
def get_markov(self):
return self.chain
