Rename latestUrl in url
This commit is contained in:
parent
7f78bea1af
commit
1451047877
33 changed files with 49 additions and 59 deletions
|
@ -24,17 +24,17 @@ def regexNamer(regex):
|
|||
return _namer
|
||||
|
||||
|
||||
def bounceStarter(latestUrl, nextSearch):
|
||||
def bounceStarter(url, nextSearch):
|
||||
"""Get start URL by "bouncing" back and forth one time."""
|
||||
@classmethod
|
||||
def _starter(cls):
|
||||
"""Get bounced start URL."""
|
||||
url = fetchUrl(latestUrl, cls.prevSearch, session=cls.session)
|
||||
if not url:
|
||||
raise ValueError("could not find prevSearch pattern %r in %s" % (cls.prevSearch.pattern, latestUrl))
|
||||
url2 = fetchUrl(url, nextSearch, session=cls.session)
|
||||
url1 = fetchUrl(url, cls.prevSearch, session=cls.session)
|
||||
if not url1:
|
||||
raise ValueError("could not find prevSearch pattern %r in %s" % (cls.prevSearch.pattern, url))
|
||||
url2 = fetchUrl(url1, nextSearch, session=cls.session)
|
||||
if not url2:
|
||||
raise ValueError("could not find nextSearch pattern %r in %s" % (nextSearch.pattern, url))
|
||||
raise ValueError("could not find nextSearch pattern %r in %s" % (nextSearch.pattern, url1))
|
||||
return url2
|
||||
return _starter
|
||||
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
from re import compile
|
||||
from ..scraper import make_scraper
|
||||
from ..util import tagre, getQueryParams, fetchUrl
|
||||
|
@ -44,6 +44,7 @@ def add(name, shortName, imageFolder=None, lastStrip=None):
|
|||
prevSearch=_prevSearch,
|
||||
help='Index format: n',
|
||||
namer=namer,
|
||||
url=baseUrl,
|
||||
)
|
||||
if lastStrip is None:
|
||||
attrs['starter'] = _starter
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile, IGNORECASE
|
||||
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile, IGNORECASE, MULTILINE
|
||||
|
||||
|
@ -121,11 +121,11 @@ class Fallen(_BasicScraper):
|
|||
|
||||
|
||||
class FredoAndPidjin(_BasicScraper):
|
||||
homepage = 'http://www.pidjin.net/'
|
||||
url = 'http://www.pidjin.net/'
|
||||
stripUrl = None
|
||||
help = 'Index format: yyyy/mm/dd/name'
|
||||
imageSearch = compile(tagre('img', 'src', '(http://cdn\.pidjin\.net/wp-content/uploads/\d+/\d+/[^"]+\.png)'))
|
||||
multipleImagesPerStrip = True
|
||||
prevSearch = compile(tagre('a', 'href', '([^"]+)')+"Prev</a>")
|
||||
starter = indirectStarter(homepage,
|
||||
compile(tagre('a', 'href', "("+homepage+r'\d\d\d\d/\d\d/\d\d/[^"]+/)')))
|
||||
starter = indirectStarter(url,
|
||||
compile(tagre('a', 'href', "("+url+r'\d\d\d\d/\d\d/\d\d/[^"]+/)')))
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
|
||||
|
@ -68,9 +68,10 @@ class Goats(_BasicScraper):
|
|||
|
||||
|
||||
class GoneWithTheBlastwave(_BasicScraper):
|
||||
starter = indirectStarter('http://www.blastwave-comic.com/index.php?p=comic&nro=1',
|
||||
url = 'http://www.blastwave-comic.com/index.php?p=comic&nro=1'
|
||||
starter = indirectStarter(url,
|
||||
compile(r'href="(index.php\?p=comic&nro=\d+)"><img src="images/page/default/latest'))
|
||||
stripUrl = 'http://www.blastwave-comic.com/index.php?p=comic&nro=%s'
|
||||
stripUrl = url[:-1] + '%s'
|
||||
imageSearch = compile(r'<img.+src=".+(/comics/.+?)"')
|
||||
prevSearch = compile(r'href="(index.php\?p=comic&nro=\d+)"><img src="images/page/default/previous')
|
||||
help = 'Index format: n'
|
||||
|
|
|
@ -1,4 +1,4 @@
|
|||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import _BasicScraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import _BasicScraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import _BasicScraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile, IGNORECASE
|
||||
from ..scraper import _BasicScraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile, IGNORECASE
|
||||
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import _BasicScraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import make_scraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import _BasicScraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
from re import compile
|
||||
from ..scraper import make_scraper
|
||||
from ..util import tagre
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import make_scraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import _BasicScraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import _BasicScraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile, MULTILINE, IGNORECASE, sub
|
||||
from os.path import splitext
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import make_scraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile, IGNORECASE
|
||||
from ..scraper import _BasicScraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile, IGNORECASE
|
||||
|
||||
|
|
|
@ -1,5 +1,5 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import make_scraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile, IGNORECASE, DOTALL
|
||||
from ..scraper import make_scraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..util import tagre
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import _BasicScraper
|
||||
|
|
|
@ -1,6 +1,6 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
# Copyright (C) 2004-2005 Tristan Seligmann and Jonathan Jacobs
|
||||
# Copyright (C) 2012 Bastian Kleineidam
|
||||
# Copyright (C) 2012-2013 Bastian Kleineidam
|
||||
|
||||
from re import compile
|
||||
from ..scraper import _BasicScraper
|
||||
|
|
|
@ -11,8 +11,8 @@ from .output import out
|
|||
class _BasicScraper(object):
|
||||
'''Base class with scrape functions for comics.
|
||||
|
||||
@type latestUrl: C{string}
|
||||
@cvar latestUrl: The URL for the latest comic strip.
|
||||
@type url: C{string}
|
||||
@cvar url: The URL for the comic strip.
|
||||
@type stripUrl: C{string}
|
||||
@cvar stripUrl: A string that is interpolated with the strip index
|
||||
to yield the URL for a particular strip.
|
||||
|
|
|
@ -36,13 +36,7 @@ description: a list of comic strips supported by Dosage
|
|||
<div id="comics">
|
||||
%(content)s
|
||||
</div>
|
||||
<script type="text/javascript">
|
||||
(function() {
|
||||
var po = document.createElement('script'); po.type = 'text/javascript'; po.async = true;
|
||||
po.src = 'https://apis.google.com/js/plusone.js';
|
||||
var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(po, s);
|
||||
})();
|
||||
</script>
|
||||
<script type="text/javascript" src="https://apis.google.com/js/plusone.js"></script>
|
||||
<script>
|
||||
window.onload = function() {
|
||||
var wall = new Masonry(document.getElementById('comics'), {
|
||||
|
@ -81,13 +75,7 @@ title: Dosage comic %(name)s
|
|||
</tr>
|
||||
</table>
|
||||
<div class="g-plusone" data-size="standard" data-annotation="inline" data-width="300"></div>
|
||||
<script type="text/javascript">
|
||||
(function() {
|
||||
var po = document.createElement('script'); po.type = 'text/javascript'; po.async = true;
|
||||
po.src = 'https://apis.google.com/js/plusone.js';
|
||||
var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(po, s);
|
||||
})();
|
||||
</script>
|
||||
<script type="text/javascript" src="https://apis.google.com/js/plusone.js"></script>
|
||||
</section>
|
||||
{%% endblock content %%}
|
||||
"""
|
||||
|
@ -153,10 +141,10 @@ def get_testinfo(filename, modified):
|
|||
key, entry = get_testentry(line)
|
||||
keys.append(key)
|
||||
update_testentry(key, entry, testinfo)
|
||||
if num_tests % 5 == 0:
|
||||
print(num_tests, end=" ", file=sys.stderr)
|
||||
elif add_error and line.startswith(" E "):
|
||||
entry["error"] = line[3:].strip()
|
||||
if num_tests % 5 == 0:
|
||||
print(num_tests, end=" ", file=sys.stderr)
|
||||
orphan_entries(keys, testinfo)
|
||||
save_result(testinfo, json_file)
|
||||
return testinfo
|
||||
|
|
Loading…
Reference in a new issue