def from_file(cls, file_path, embed_dim, vocab_size=None):
"""Load word embeddings.
Args:
file_path (str)
embed_dim (int): expected embed_dim
vocab_size (int): max # of words in the vocab. If not specified, uses all available vectors in file.
"""
if vocab_size is None:
vocab_size = num_lines(file_path)
words = []
embeds = []
with codecs.open(file_path, 'r', encoding='utf-8') as f:
lines = verboserate(
f,
desc='Loading embeddings from {}'.format(file_path),
total=vocab_size)
for i, line in enumerate(lines):
if i == vocab_size: break
tokens = line.split()
word, embed = tokens[0], np.array(
[float(tok) for tok in tokens[1:]], dtype=np.float32)
if len(embed) != embed_dim:
raise ValueError('expected {} dims, got {} dims'.format(
embed_dim, len(embed)))
words.append(word)
embeds.append(embed)
vocab = SimpleVocab(words)
embed_matrix = np.stack(embeds)
embed_matrix = embed_matrix.astype(np.float32)
assert embed_matrix.shape == (vocab_size, embed_dim)
return cls(embed_matrix, vocab)
def examples_from_file(path, seq_length_limit):
"""Return list[EditExample] from file path."""
examples = []
# count total lines before loading
total_lines = num_lines(path)
with codecs.open(path, 'r', encoding='utf-8') as f:
lnum = 0
for line in verboserate(f,
desc='Reading data file.',
total=total_lines):
split = line.strip().split('\t')
lnum += 1
input_words = []
try:
for c in config.source_cols:
input_words.append(split[c].split(' '))
trg_words = split[config.target_col].split(
' ') # gold answer
assert len(trg_words) > 0
ex = EditExample(input_words, trg_words)
# skip sequences that are too long, because they use up memory
if max_seq_length(ex) > seq_length_limit:
continue
examples.append(ex)
except:
print 'bad formatting in line ' + str(lnum)
print line
return examples
def file_rows(path, limit):
with codecs.open(path, encoding='utf-8') as f:
reader = unicode_csv_reader(f, delimiter='\t')
for i, row in verboserate(enumerate(reader), total=num_lines(path)):
if i == 0:
continue # skip header
if i > limit:
break
yield row
def examples_from_file(path):
"""Return list[list[unicode]] from file path."""
examples = []
# count total lines before loading
total_lines = num_lines(path)
with codecs.open(path, 'r', encoding='utf-8') as f:
for line in verboserate(f, desc='Reading data file.', total=total_lines):
ex = line.strip().lower()
ex_words = ex.split(' ')
assert len(ex_words) > 0
examples.append(ex_words)
return examples
def load_sentences(path, limit):
"""Load sentences.
Args:
path (str)
limit (int)
Returns:
list[list[unicode]]: a list of sentences (each a sequence of words)
"""
with codecs.open(path, 'r', encoding='utf-8') as f:
sentences = []
total_lines = min(num_lines(path), limit)
lines = verboserate(f, desc='Loading sentences', total=total_lines)
for i, line in enumerate(lines):
if i == limit:
break
sentences.append(line.lower().strip().split())
return sentences