self.lock.release()
def start_downloader(self):
start_time = datetime.now()
self.build_image_folder()
urls = self.build_query_url()
self.pool.map(self.revolveImageURL,urls)
resolve_time = datetime.now()
print('解析網址共花',resolve_time-start_time)
while self.imageURL.qsize():
url = self.imageURL.get()
self.pool.map(self.saveImage,url)
self.pool.close()
self.pool.join()
endtime = datetime.now()
print('下載結束,共花了',endtime-resolve_time)
print('共下載',self.index-self.start_index,'張圖片')
print('共',self.error,'個圖片網址無法獲取')
return self.index-1
if __name__ == '__main__':
if downloader_setting.is_python3():
search = input('關鍵字：')
else:
search = raw_input('關鍵字：')
mydownloader = downloader(search)
mydownloader.start_downloader()
not_image.append(filename)
for filename in not_image:
filenames.remove(filename)
return filenames
def build_subfolder(folder_dir):
name_list = ['train', 'validation']
for name in name_list:
subfolder_dir = os.path.join(folder_dir, name)
if not os.path.exists(subfolder_dir):
os.makedirs(subfolder_dir)
if __name__ == '__main__':
if is_python3():
DIR = input('input folder directory:')
else:
DIR = raw_input('input folder directory:')
build_subfolder(DIR)
filenames = get_filenames(DIR)
print(filenames)
num = len(filenames)
train_num = num * 4 // 5
val_num = num // 5
random.shuffle(filenames)
train = filenames[:train_num]
validation = filenames[train_num:]
folder_list = [train, validation]