def dev():
import core.markovchain as mc
mcW = mc.MarkovChain()
sm = sentence.SentenceMaker(mcW)
ss = sentences_from_thesaurus(["cat", "anger", "sorry", "cat"], mcW)
for s in ss:
sent = sm.polish_sentence(s)
print(" " + sm.to_string(sent))
def seq_to_para(seq, mc):
"""Takes a sequence of seed phrases & returns one sentence for each, as a list of lists of words"""
sm = sentence.SentenceMaker(mc)
para = []
for seed in seq:
seed_tokens = seed.split()
tokens = sm.generate_sentence_tokens(seed_tokens)
para.append(tokens) # ss.to_string(tokens) + " "
return para
def make_speech_bits(self, seeds):
n = len(seeds)
speaker_id = self.speaker_sequence(0, n)
speech_acts_seq = self.speech_sequence(n)
bits = []
ss = sentence.SentenceMaker(self.mc)
for i in range(n):
sent_toks = ss.generate_sentence_tokens([seeds[i]], self.target_len[speaker_id[i]])
sent_toks = ss.polish_sentence(sent_toks)
bits.append({'speaker_name': self.speakers[speaker_id[i]]["name"],
'speech_act': speech_acts_seq[speaker_id[i]],
'seq_id': speaker_id[i],
'speech': sent_toks,
'paragraph': True})
return(bits)
def load_text():
# mcW = mc.MarkovChain()
# mcW.delete_all_in_redis_careful()
doc = bs.get_clean_text(107)
cache = store.files(store.Storage_type.local_dev)
cache.write_text(doc['text'], "107.txt")
with open("database/resources/texts/107.txt") as file_in:
text = file_in.read()
tokens = tokenize.tokenize(text[2000:3000])
print(tokens['tokens'][0:50])
print(tokens['entities'])
# mcW.train_words(tokens['tokens'])
# mcW.append_ner(tokens['entities'])
generator = sentence.SentenceMaker(mc)
s = generator.generate_sentence_tokens(["the", "man"])
print(" ".join(s))
def dev():
mcW = mc.MarkovChain()
w2vec = w2v.WordVectors()
seeds = w2vec.path("swim", 20)
print(seeds)
generator = sentence.SentenceMaker(mcW)
for sd in seeds:
sen = generator.generate_sentence_tokens([sd])
sen = generator.polish_sentence(sen)
print(" " + sentence.SentenceMaker.to_string(sen))
print('')
nm = names.NameMaker()
speakers = [nm.random_person() for i in range(1, 4)]
dm = dialogue.dialogue_maker([n['name'] for n in speakers],
[n['pronoun'] for n in speakers], mcW, seeds)
dlg = dm.make_dialogue()
for s in dlg:
print(" " + sentence.SentenceMaker.to_string(s))
story_grammar.make_story(generator)
import argparse
import core.sentence as sentence
import core.markovchain as mc
import core.paragraphs as paras
import nlp.story_grammar as story_grammar
import knowledge.names as names
import core.dialogue as dialogue
mcW = mc.MarkovChain()
generator = sentence.SentenceMaker(mcW)
def make_seeds(seeds, max=5):
if seeds is None:
seeds = generator.generate_sentence_tokens(["the"]) # Make a random sentence of seeds
seeds = seeds[0:max]
return seeds
def make_sentences(n, seeds=None):
seeds = make_seeds(seeds)
# if seeds is None:
# seeds = generator.generate_sentence_tokens(["the"]) # Make a random sentence of seeds
if isinstance(seeds, str):
seeds = seeds.split()
p = paras.seq_to_para(seeds[0:n], mcW)
for sent in p:
print(generator.to_string(generator.polish_sentence(sent)))
def make_story():