Fix PennyArcade (fixes #213)

Unfortunatly, images are now split into panels, so some re-assembly
might be required...
This commit is contained in:
Tobias Gruetzmacher 2022-05-27 20:36:02 +02:00
parent 173568f6a4
commit ca2c0cd6ed

View file

@ -1,7 +1,7 @@
# SPDX-License-Identifier: MIT # SPDX-License-Identifier: MIT
# Copyright (C) 2004-2008 Tristan Seligmann and Jonathan Jacobs # Copyright (C) 2004-2008 Tristan Seligmann and Jonathan Jacobs
# Copyright (C) 2012-2014 Bastian Kleineidam # Copyright (C) 2012-2014 Bastian Kleineidam
# Copyright (C) 2015-2020 Tobias Gruetzmacher # Copyright (C) 2015-2022 Tobias Gruetzmacher
# Copyright (C) 2019-2020 Daniel Ring # Copyright (C) 2019-2020 Daniel Ring
from re import compile, escape from re import compile, escape
@ -102,17 +102,14 @@ class PennyAndAggie(_BasicScraper):
class PennyArcade(_ParserScraper): class PennyArcade(_ParserScraper):
url = 'https://www.penny-arcade.com/comic/' url = 'https://www.penny-arcade.com/comic/'
stripUrl = url + '%s' stripUrl = url + '%s'
firstStripUrl = stripUrl % '1998/11/18' firstStripUrl = stripUrl % '1998/11/18/the-sin-of-long-load-times'
imageSearch = '//div[@id="comicFrame"]//img' imageSearch = '//div[d:class("comic-panel")]//img'
prevSearch = '//a[d:class("btnPrev")]' prevSearch = '//a[d:class("older")]'
nextSearch = '//a[d:class("btnNext")]' nextSearch = '//a[d:class("newer")]'
multipleImagesPerStrip = True
starter = bounceStarter starter = bounceStarter
help = 'Index format: yyyy/mm/dd' help = 'Index format: yyyy/mm/dd'
def namer(self, image_url, page_url):
p = page_url.split('/')
return '%04d%02d%02d' % (int(p[4]), int(p[5]), int(p[6]))
class PeppermintSaga(_WPNavi): class PeppermintSaga(_WPNavi):
url = 'http://www.pepsaga.com/' url = 'http://www.pepsaga.com/'