def _collate_fn(batch, token2idx, label_type, LFR_m=1, LFR_n=1):
"""
Args:
batch: list, len(batch) = 1. See AudioDataset.__getitem__()
Returns:
xs_pad: N x Ti x D, torch.Tensor
ilens : N, torch.Tentor
ys_pad: N x To, torch.Tensor
"""
# batch should be located in list
assert len(batch) == 1
batch = load_inputs_and_targets(batch[0],
token2idx,
label_type,
LFR_m=LFR_m,
LFR_n=LFR_n)
xs, ys = batch
# TODO: perform subsamping
# get batch of lengths of input sequences
ilens = np.array([x.shape[0] for x in xs])
# perform padding and convert to tensor
xs_pad = pad_list([torch.from_numpy(x).float() for x in xs], 0)
ilens = torch.from_numpy(ilens)
ys_pad = pad_list([torch.from_numpy(y).long() for y in ys], 0)
return xs_pad, ilens, ys_pad
def f_xy_pad(batch):
xs_pad = pad_list([torch.tensor(sample[0]).long() for sample in batch[0]],
0)
ys_pad = pad_list([torch.tensor(sample[1]).long() for sample in batch[0]],
0)
# xs_pad = pad_to_batch([sample for sample in batch[0][0]], 0)
# ys_pad = pad_to_batch([sample for sample in batch[0][1]], 0)
return xs_pad, ys_pad
def _collate_fn(batch, LFR_m=1, LFR_n=1, model_choose='baseline3'):
"""
Args:
batch: list, len(batch) = 1. See AudioDataset.__getitem__()
Returns:
xs_pad: N x Ti x D, torch.Tensor
ilens : N, torch.Tentor
ys_pad: N x To, torch.Tensor
"""
# batch should be located in list
assert len(batch) == 1
batch = load_inputs_and_targets(batch[0], LFR_m=LFR_m, LFR_n=LFR_n)
xs, dialect_labels = batch
import math
if model_choose in ['baseline2', 'baseline4']:
ilens = np.array([int(math.ceil(x.shape[0] / 4)) for x in xs])
else:
ilens = np.array([int(math.ceil(x.shape[0])) for x in xs])
# perform padding and convert to tensor
xs_pad = pad_list([torch.from_numpy(x).float() for x in xs], 0)
ilens = torch.from_numpy(ilens)
dialect_labels = torch.from_numpy(dialect_labels)
return xs_pad, ilens, dialect_labels
def preprocess(self, padded_input):
"""Generate decoder input and output label from padded_input
Add <sos> to decoder input, and add <eos> to decoder output label
"""
ys = [y[y != IGNORE_ID] for y in padded_input] # parse padded ys
# prepare input and output word sequences with sos/eos IDs
eos = ys[0].new([self.eos_id])
sos = ys[0].new([self.sos_id])
ys_in = [torch.cat([sos, y], dim=0) for y in ys]
ys_out = [torch.cat([y, eos], dim=0) for y in ys]
# padding for ys with -1
# pys: utt x olen
ys_in_pad = pad_list(ys_in, self.eos_id)
ys_out_pad = pad_list(ys_out, IGNORE_ID)
assert ys_in_pad.size() == ys_out_pad.size()
return ys_in_pad, ys_out_pad
def extract_content_gsc(file_name):
tree = html.fromstring(parse_file(file_name))
title = tree.xpath('//h1[@class="product_title entry-title"]/text()')[0]
price_from = tree.xpath(
'//div[@class="summary entry-summary"]/p/span[1]/span/text() | //div[@class="summary entry-summary"]/p/span[1]/text()')
price_from = "".join(price_from)
price_to = tree.xpath(
'//div[@class="summary entry-summary"]/p/span[2]/span/text() | //div[@class="summary entry-summary"]/p/span[2]/text()')
price_to = "".join(price_to)
description = tree.xpath('//div[@class="woocommerce-product-details__short-description"]/p/text()')
description = ("".join(description)).replace("\n", "")
category = tree.xpath('//span[@class="posted_in"]/a/text()')[0]
tags = tree.xpath('//span[@class="tagged_as"]/a/text()')
attributes = [attr.strip() for attr in
tree.xpath('//table[@class="table table-hover variations"]/thead/tr/th[not(@*)]/text()')]
var_attr = {}
for attr in attributes:
var_attr[attr] = tree.xpath(
f'//table[@class="table table-hover variations"]/tbody/tr/td[@data-title="{attr}"]/text()')
separate_list_prices = tree.xpath(
'//span[@class="price"]//span[1]/span/text() | //span[@class="price"]//span[1]/text()')
c = 0
separate_list_price = []
while c < len(separate_list_prices) - 1:
separate_list_price.append(f"{separate_list_prices[c]}{separate_list_prices[c + 1]}")
c += 2
separate_discount_prices = pad_list(tree.xpath('//ins/span/span/text() | //ins/span/text()'), var_attr["Model"],
"currency_xpath")
c = 0
separate_discount_price = []
while c < len(separate_discount_prices) - 1:
separate_discount_price.append(f"{separate_discount_prices[c]}{separate_discount_prices[c + 1]}")
c += 2
variations = zip(separate_list_price, separate_discount_price)
results = generate_json_gsc(title, price_from, price_to, description, category, tags, var_attr, variations)
print(json.dumps(results, indent=4))
def f_x_pad(batch):
return pad_list([torch.tensor(sample).long() for sample in batch[0]], 0)
