def __init__(self, paths):
"""
Constructor for the DialogueManager
- Loads the intent recognizer (is this about programming, or just
chit-chatting?)
- Loads the tf-idf vectorizer (the vectorizer trained on the dialogue
and StackOverflow thread questions)
Parameters
----------
paths : dict
Where the keys are names, and the values are lists of paths
relative to this directory
"""
print("Loading resources...")
# Create the chatbot
self.create_chitchat_bot()
# Intent recognition:
self.intent_recognizer = \
unpickle_file(Path(*paths['INTENT_RECOGNIZER']))
self.tfidf_vectorizer = \
unpickle_file(Path(*paths['TFIDF_VECTORIZER']))
self.ANSWER_TEMPLATE = ('I think its about {}\n'
'This thread might help you: '
'https://stackoverflow.com/questions/{}')
# Goal-oriented part:
self.tag_classifier = unpickle_file(Path(*paths['TAG_CLASSIFIER']))
self.thread_ranker = ThreadRanker(paths)
def __init__(self, ):
print('Initializing corrector model...')
self.trie = unpickle_file('model/correct_datrie.bin')
self.word_frequence = unpickle_file('data/word_frequence.bin')
self.load_lm()
self.vocabulary = unpickle_file('data/vocabulary.bin')
print('Corrector model established.')
def __init__(self, paths):
# print("Loading resources...")
# Intent recognition:
self.intent_recognizer = unpickle_file(paths['INTENT_RECOGNIZER'])
self.tfidf_vectorizer = unpickle_file(paths['TFIDF_VECTORIZER'])
self.ANSWER_TEMPLATE = 'I think its about %s\n This thread might help you: ' \
'https://stackoverflow.com/questions/%s '
# Goal-oriented part:
self.tag_classifier = unpickle_file(paths['TAG_CLASSIFIER'])
self.thread_ranker = ThreadRanker(paths)
def __init__(self, paths=utils.RESOURCE_PATH):
print("Loading resources...")
# Intent recognition:
self.intent_recognizer = utils.unpickle_file(paths['INTENT_RECOGNIZER'])
self.tfidf_vectorizer = utils.unpickle_file(paths['TFIDF_VECTORIZER'])
self.ANSWER_TEMPLATE = 'I think its about %s\nThis thread might help you: https://stackoverflow.com/questions/%s'
# Goal-oriented part:
self.tag_classifier = utils.unpickle_file(paths['TAG_CLASSIFIER'])
self.thread_ranker = ThreadRanker(paths)
self.chatbot=ChatBot('tiksbot')
trainer = ChatterBotCorpusTrainer(self.chatbot)
trainer.train("chatterbot.corpus.english")
def __init__(
self,
datapath,
):
print('Initializing retrival model...')
self.data = pd.read_csv(datapath)
self.tfidf = unpickle_file('model/tfidf.model')
self.tfidf_vec = unpickle_file('data/doc_tfidf_vec.bin')
self.inverse_idx = unpickle_file('data/inverse_idx_table.bin')
self.word_2_id = unpickle_file('data/full_word2id.bin')
self.id_2_word = {d: w for w, d in self.word_2_id.items()}
self.word_2_id_for_filter = unpickle_file('data/tfidf_word2vec.bin')
self.idf, self.avg_len = unpickle_file('data/idf_and_avglen.bin')
self.word_vec = load_gensim_vec('data/wordvec_fasttext_300.txt')
self.annoy = load_annoy_index('data/sentvec.ann')
self.ltp = nlp.Ltp(seg=True,
pos=True,
seg_lexicon_path='data/lexicon_seg.txt',
pos_lexicon_path='data/lexicon.txt')
self.text_precess = nlp.ProcessText()
self.stopwords = nlp.load_stop(
['data/chinese_stopwords.txt', 'data/哈工大停用词表.txt'])
self.nonsense_word = ['请问', '想知道']
self.corrector = Corrector()
print('Retrival model established.')
def __load_embeddings_by_tag(self, tag_name):
"""
Returns the thread_id and the thread_embeddings
Parameters
----------
tag_name : str
Name of the tag
Returns
-------
thread_ids : array-like, shape (n_ids, )
Array of the ids corresponding to the tag
thread_embeddings : np.array, shape (n_ids, embedding_dim)
The embedding to the corresponding tag
"""
embeddings_path = self.thread_embeddings_dir.joinpath(tag_name +
'.pkl')
thread_ids, thread_embeddings = unpickle_file(embeddings_path)
return thread_ids, thread_embeddings
def __load_embeddings_by_tag(self, tag_name):
embeddings_path = os.path.join(self.thread_embeddings_folder, tag_name + ".pkl")
thread_ids, thread_embeddings = unpickle_file(embeddings_path)
return thread_ids, thread_embeddings
def __init__(self, paths):
self.embeddings, self.embeddings_dim = self._load_embeddings(
paths['DIALOGUE_EMBEDDINGS'])
self.question_vectors = unpickle_file(paths['QUESTION_VECTORS'])
self.dialogues = datasets.readCornellData(paths['DIALOGUE_FOLDER'],
max_len=100)
