def main():
try:
global conf, uid, key, ip, port, sock, status_file
conf = read_json(os.path.join(os.path.dirname(__file__), config_path))
sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
socket.setdefaulttimeout(3)
sock.settimeout(3)
status_file = open(status_path, "w")
#sock.setblocking(False)
ip = conf['remote']
port = conf['remote_port']
if isfile(key_path):
key = read_text(key_path)
else:
prompt_for_key()
if isfile(uid_path):
uid = read_text(uid_path)
else:
uid = gen_id()
write_text(uid_path, uid)
check_alive()
#schedule.every(conf['interval']).seconds.do(check_alive)
schedule.every(5).seconds.do(check_alive)
print("client is running")
while True:
schedule.run_pending()
except KeyboardInterrupt:
print('manual exit')
post_request("GBYE")
sock.close()
def saveResponse(resp: Response):
if resp:
text = resp.text
logger.debug('response:%s', text)
name = util.current_date() + '.json'
file = os.path.join('raw', name)
util.write_text(file, text)
def save_brand_raw_response(resp: Response, category: str):
"""保存品牌榜响应内容
"""
if resp:
content = resp.text
date = util.current_date()
filename = '{}.json'.format(category)
logger.info('save response:%s', filename)
file = os.path.join('raw', date, 'brand', filename)
util.write_text(file, content)
def save_raw_response(resp: Response, filename: str):
"""保存原始响应内容
"""
if resp:
content = resp.text
filename = '{}.json'.format(filename)
logger.info('save response:%s', filename)
date = util.current_date()
file = os.path.join('raw', date, filename)
util.write_text(file, content)
def prompt_for_key():
global key
res = 'NKEY'
while res == 'NKEY':
#while len(key) != 32:
with open(status_path, "w") as status_file:
status_file.write('Please enter valid license key: ')
key = input('Please enter valid license key: ')
res = post_request('HELO')
write_text(key_path, key)
status_file.close()
def get_additional_inputs(self, input_type, available_inputs):
specified_eps = {'soil': 'eps_soil', 'veg_scatterers': 'eps_veg'} #
sub_params = self._params[input_type].keys()
if input_type in specified_eps:
if specified_eps[input_type] not in available_inputs:
util.write_text('"' + specified_eps[input_type] + '" is not provided in config file. Dielectric ' \
'constants will be calculated from specified ' + input_type + ' properties', self.log)
sub_params.remove(specified_eps[input_type])
sub_params += self._params[specified_eps[input_type]].keys()
else:
util.write_text(
'Specified "' + specified_eps[input_type] +
'" used as the dielectric constants for ' +
input_type.upper() + ' layer.', self.log)
return sub_params
def handle_hot_brands(dy: Douyin):
"""热门品牌
"""
categories, resp = dy.get_brand_category()
saveRawResponse(resp, 'brand-category')
time.sleep(1)
brand_map = {}
for category in categories:
id = category['id']
category = category['name']
brands, resp = dy.get_hot_brand(int(id))
time.sleep(1)
saveBrandRawResponse(resp, category)
brand_map[category] = brands
md = generate_brand_md(brand_map)
filename = '{}-brand.md'.format(util.current_date())
file = os.path.join('archives', filename)
util.write_text(file, md)
def saveArchiveMd(md):
logger.debug('archive md:%s', md)
name = util.current_date()+'.md'
file = os.path.join('archives', name)
util.write_text(file, md)
def saveReadme(md):
logger.debug('today md:%s', md)
util.write_text('README.md', md)
def save_raw_content(content: str, filePrefix: str):
filename = '{}-{}.html'.format(filePrefix, util.current_date())
file = os.path.join('raw', filename)
util.write_text(file, content)
def save_archive_md(md):
logger.debug('archive md:%s', md)
name = '{}.md'.format(util.current_date())
file = os.path.join('archives', name)
util.write_text(file, md)
def save_readme(md):
logger.info('today md:%s', md)
util.write_text('README.md', md)
pbar.close()
if (ep - START_EP + 1) % 10 == 0 and frozen_model:
adjust_learning_rate(optimizer, 0.5)
memory.update_rate *= 0.5
if ep > 2 and ep % 3 == 0:
## random eval
no_batch = int(0.2 * len(image_files) / BATCH_SIZE)
# simple
eval_loader = get_beauty_loader(image_files, labels, bg_list, simple_image_transform, load_image,
batch_size=BATCH_SIZE, shuffle=True, num_workers=NUM_PROC)
simple_val_acc = estimate_topk_accuracy(model, memory, eval_loader, k=TOPK, no_batch=no_batch)
print("\n\n---- Simple Val Acc = {}".format(simple_val_acc))
write_text(log_file, "Epoch {} : simple val acc {} \n".format(ep, simple_val_acc))
# normal
eval_loader = get_beauty_loader(image_files, labels, bg_list, normal_image_transform, load_image,
batch_size=BATCH_SIZE, shuffle=True, num_workers=NUM_PROC)
normal_val_acc = estimate_topk_accuracy(model, memory, eval_loader, k=TOPK, no_batch=no_batch)
print("---- Simple Val Acc = {}".format(normal_val_acc))
write_text(log_file, "Epoch {} : normal val acc {} \n".format(ep, normal_val_acc))
# Complex 1
eval_loader = get_beauty_loader(image_files, labels, bg_list, complex_image_transform, load_rgba_image,
batch_size=BATCH_SIZE, shuffle=True, num_workers=NUM_PROC)
complex_val_acc1 = estimate_topk_accuracy(model, memory, eval_loader, k=TOPK, no_batch=no_batch)
print("---- Complex Val Acc 1 with load_rgba = {}".format(complex_val_acc1))
write_text(log_file, "Epoch {} : complex val acc 1 with load_rgba {} \n".format(ep, complex_val_acc1))
def saveRawResponse(content: str):
logger.debug('raw content:%s', content)
name = util.current_date()+'.json'
file = os.path.join('raw', name)
util.write_text(file, content)
def save_readme(md):
logger.debug('readme:%s', md)
util.write_text('README.md', md)
final_distances = []
for d in distances:
tmp_indices = [indices1[i] for i in find_element(d, distances1)
] + [indices2[i] for i in find_element(d, distances2)]
for a in tmp_indices:
if a not in final_indices:
final_indices.append(a)
final_distances.append(d)
return final_distances[0:topk], final_indices[0:topk]
model_name = sys.argv[1] # MultiScaleDense121, MultiScaleDense121Edge, Combine
test_fol = sys.argv[2] # /testset
save_file = sys.argv[3] # /result/predictions.csv
write_text(save_file, "Validation Image ID,Training Image ID\n", mode='w')
# image_list = [os.path.join(test_fol, f) for f in os.listdir(test_fol) if f.endswith(('jpg', 'png', 'jpeg', 'JPG', 'PNG', 'JPEG'))]
image_list = [os.path.join(test_fol, f) for f in os.listdir(test_fol)]
image_list.sort()
if model_name in ["MultiScaleDense121", "model1"]:
print("*** Generate predictions by MultiScaleDense121\n\n")
elif model_name in ["MultiScaleDense121Edge", "model2"]:
print("*** Generate predictions by MultiScaleDense121Edge\n\n")
else:
print(
"*** Generate predictions by the combination of (MultiScaleDense121 + MultiScaleDense121Edge)\n\n"
)
for tmp_image in image_list:
print(tmp_image)
def saveRawContent(content: str, filePrefix: str, fileSuffix='json'):
logger.debug('raw content:%s', content)
name = '{}-{}.{}'.format(filePrefix, util.current_date(), fileSuffix)
file = os.path.join('raw', name)
util.write_text(file, content)
