def __init__(self, excel_name, parent_folder=None):
"""
:param excel_name: excel文件名
:param parent_folder: 文件夹
"""
self.excel_name = '{}_{}.xlsx'.format(excel_name, t)
self.parent_folder = parent_folder
self.work_book = Workbook()
make_dir(self.parent_folder)
def save_image(imagetuple, itemtuple):
folder = helper.make_dir(itemtuple[-2:]) # make_dir should return a dir, not create one
filename = imagetuple[0]
image = imagetuple[1]
path = os.path.join(config.path, 'image', folder, filename)
print "saving image", filename
f = open(path, 'wb')
f.write(image)
f.close()
def download_new_item(latest_item):
screen_geometry = helper.get_screen_geometry()
url = latest_item[0]
folder = helper.make_dir(latest_item[-2:])
conn = urllib.urlopen(url)
page = conn.read()
conn.close()
images = re.findall(pattern_image, page)
images = list(set(images))
for image in images:
m = re.match('.*/(.*)', image)
filename = m.groups()[0]
m = re.match('.*?(\d+)x(\d+).*', filename)
(width, height) = list(map(int, m.groups()[:2]))
if not (width, height) == screen_geometry:
continue
if os.path.exists(os.path.join(config.path, 'image', folder, filename)):
continue
logging.debug('downloading image: %s\n' % image)
conn = urllib.urlopen(image)
image_data = conn.read()
conn.close()
yield (filename, image_data)
