# catch label and screenshot img and segment them into smaller size
img, label = None, None
catch_success = False
if is_catch_element and '.com' in links.iloc[index]:
# set the format of libel
libel_format = pd.read_csv(os.path.join(data_position, 'format.csv'),
index_col=0)
url = 'http://' + links.iloc[index] if 'http://' not in links.iloc[
index] else links.iloc[index]
try:
img, label = catch.catch(url, label_path, img_org_path,
libel_format, driver_path)
except FunctionTimedOut:
print('Catch Time Out')
continue
# segment the lengthy images
if is_segment and img is not None:
seg.segment_img(img, 600, img_segment_path, 0)
# read and draw label on segment img
if is_draw_label and img is not None and label is not None:
draw.label(label, img, img_drawn_path)
end_time = time.clock()
print("*** %d Time taken:%ds ***\n" % (index, int(end_time - start_time)))
if index > end_pos:
break
if browser == 'PhantomJS':
driver = webdriver.PhantomJS(
executable_path=os.path.join(driver_path, 'phantomjs.exe'))
elif browser == 'Chrome':
options = webdriver.ChromeOptions()
options.add_argument(
'--headless') # do not show the browser every time
driver = webdriver.Chrome(executable_path=os.path.join(
driver_path, 'chromedriver.exe'),
options=options)
# set the format of label
label_format = pd.read_csv('data/format.csv', index_col=0)
try:
img, label = catch.catch(url, out_html, out_elements, out_img,
label_format, driver)
# read and draw label on segment img
if is_draw_label and img is not None and label is not None:
img_drawn_path = 'data/0_drawn.png'
draw.label(label, img, img_drawn_path)
# segment the lengthy images
if is_segment and img is not None:
img_segment_dir = '/segment'
seg.segment_img(img, 600, img_segment_dir, 0)
except FunctionTimedOut:
print('Catch Time Out')
end_time = time.clock()
print("*** Time taken:%ds ***" % int(end_time - start_time))