def process_sent(doc,
word2vec,
vocab,
ivocab,
word_vector_size,
to_return="word2vec",
silent=False,
encoder_decoder=None,
vocab_dict={}):
document_vector = []
if to_return == "word2vec":
document_vector = [
process_word(w,
word2vec,
vocab,
ivocab,
word_vector_size,
to_return,
silent=True) for w in doc
]
elif to_return == "skip_thought":
sentences = punkt_sentences(doc)
norm_sentences = [normalize.xml_normalize(s) for s in sentences]
document_vector = [sk.encode(encoder_decoder, norm_sentences)]
elif to_return == "one_hot":
data_gen.run_onehot(doc, vocab_dict)
return document_vector
def process_sent(doc, word2vec, vocab, ivocab, word_vector_size, to_return="word2vec", silent=False, encoder_decoder=None, vocab_dict={}):
document_vector = []
if to_return=="word2vec":
document_vector = [process_word(w, word2vec, vocab, ivocab , word_vector_size, to_return, silent=True) for w in doc]
elif to_return=="skip_thought":
sentences = punkt_sentences(doc)
norm_sentences = [normalize.xml_normalize(s) for s in sentences]
document_vector = [ sk.encode(encoder_decoder, norm_sentences)]
elif to_return=="one_hot":
data_gen.run_onehot(doc, vocab_dict)
return document_vector
def transform_doc(self, doc, corpus):
hot_doc = data_gen.run_onehot(doc, self.vocab_dict, self.doc_length, self.doc_length)
hot_corpus = data_gen.run_onehot(corpus, self.vocab_dict, self.doc_length, self.doc_length)
hot_docs = []
hot_docs.append(hot_doc)
hot_docs.append(hot_corpus)
fake_labels = []
fake_labels.append(self.hot_fake_label)
fake_labels.append(self.hot_fake_label)
#run the network for both docs as once as the input expects a list and it is likely faster
#the labels aren't used so it doesn't matter what those are
connected_layers = self.session.run(self.connected_layer, feed_dict={self.x: hot_docs, self.y: fake_labels})
return connected_layers
def prep_data(self, in_data, vocab, min_length, max_length):
documents = []
labels = []
for entry in in_data:
text = entry["body_text"]
documents.append(
data_gen.run_onehot(text, vocab, min_length, max_length))
label = entry["cluster_id"]
labels.append(label)
#encode the labels in a dictionary
unique_labels = np.unique(labels)
i = 0
unique_label_dict = {}
for u_c in unique_labels:
unique_label_dict[u_c] = i
i += 1
hot_labels = []
n_classes = len(unique_labels)
for c in labels:
cluster_vect = np.zeros(n_classes, dtype=int)
cluster_vect[unique_label_dict[c]] = 1
hot_labels.append(cluster_vect.tolist())
return documents, hot_labels, n_classes
def prep_news_data(self, vocab, min_length, max_length):
from sklearn.datasets import fetch_20newsgroups
newsgroups = fetch_20newsgroups()
documents = [
data_gen.run_onehot(normalize.xml_normalize(text), vocab,
min_length, max_length)
for text in newsgroups.data
]
labels = newsgroups.target
#encode the labels in a dictionary
unique_labels = np.unique(labels)
i = 0
unique_label_dict = {}
for u_c in unique_labels:
unique_label_dict[u_c] = i
i += 1
hot_labels = []
n_classes = len(unique_labels)
for c in labels:
cluster_vect = np.zeros(n_classes, dtype=int)
cluster_vect[unique_label_dict[c]] = 1
hot_labels.append(cluster_vect.tolist())
return documents, hot_labels, n_classes
def prep_data(self, in_data, vocab, min_length, max_length):
documents = []
labels = []
for entry in in_data:
text = entry["body_text"]
documents.append(data_gen.run_onehot(text, vocab, min_length, max_length))
label = entry["cluster_id"]
labels.append(label)
#encode the labels in a dictionary
unique_labels = np.unique(labels)
i = 0
unique_label_dict = {}
for u_c in unique_labels:
unique_label_dict[u_c] = i
i +=1
hot_labels = []
n_classes = len(unique_labels)
for c in labels:
cluster_vect = np.zeros(n_classes, dtype=int)
cluster_vect[unique_label_dict[c]]=1
hot_labels.append(cluster_vect.tolist())
return documents, hot_labels, n_classes
def test_onehot():
"""Test one-hot document generation"""
doc = ["hello", "you", "wanton", "civet", ",", "you"]
vocab = {"hello": 0, "you": 1, "civet": 2, ",": 3}
doc_onehot = data_gen.run_onehot(doc, vocab)
doc_onehot_minlength = data_gen.run_onehot(doc, vocab, min_length=10)
doc_onehot_maxlength = data_gen.run_onehot(doc, vocab, max_length=2)
doc_onehot_encoded = np.array([[1., 0., 0., 0., 0.], [0., 1., 0., 0., 1.],
[0., 0., 1., 0., 0.], [0., 0., 0., 1., 0.]],
dtype=np.float32)
# encoding is correct
assert (doc_onehot == doc_onehot_encoded).all()
# minimum length correctly enforced
assert doc_onehot_minlength.shape == (4, 10)
# maximum length correctly enforced
assert doc_onehot_maxlength.shape == (4, 2)
def transform_doc(self, doc, corpus):
hot_doc = data_gen.run_onehot(doc, self.vocab_dict, self.doc_length,
self.doc_length)
hot_corpus = data_gen.run_onehot(corpus, self.vocab_dict,
self.doc_length, self.doc_length)
hot_docs = []
hot_docs.append(hot_doc)
hot_docs.append(hot_corpus)
fake_labels = []
fake_labels.append(self.hot_fake_label)
fake_labels.append(self.hot_fake_label)
#run the network for both docs as once as the input expects a list and it is likely faster
#the labels aren't used so it doesn't matter what those are
connected_layers = self.session.run(self.connected_layer,
feed_dict={
self.x: hot_docs,
self.y: fake_labels
})
return connected_layers
def prep_news_data(self, vocab, min_length, max_length):
from sklearn.datasets import fetch_20newsgroups
newsgroups= fetch_20newsgroups()
documents = [data_gen.run_onehot(normalize.xml_normalize(text), vocab, min_length, max_length)
for text in newsgroups.data]
labels = newsgroups.target
#encode the labels in a dictionary
unique_labels = np.unique(labels)
i = 0
unique_label_dict = {}
for u_c in unique_labels:
unique_label_dict[u_c] = i
i +=1
hot_labels = []
n_classes = len(unique_labels)
for c in labels:
cluster_vect = np.zeros(n_classes, dtype=int)
cluster_vect[unique_label_dict[c]]=1
hot_labels.append(cluster_vect.tolist())
return documents, hot_labels, n_classes
