Ricordisamoa has uploaded a new change for review.
https://gerrit.wikimedia.org/r/149219
Change subject: standardize spamremove.py
......................................................................
standardize spamremove.py
- SpamRemoveBot extends pywikibot.Bot (with built-in 'always' option)
- use GeneratorFactory to provide -namespace: filtering
and other standard arguments
Change-Id: If8e5f694721fcf28256c868f2b1a8eb230227208
---
M scripts/spamremove.py
1 file changed, 84 insertions(+), 71 deletions(-)
git pull ssh://gerrit.wikimedia.org:29418/pywikibot/core
refs/changes/19/149219/1
diff --git a/scripts/spamremove.py b/scripts/spamremove.py
index 60906f1..f0d8c09 100755
--- a/scripts/spamremove.py
+++ b/scripts/spamremove.py
@@ -18,100 +18,113 @@
-always Do not ask, but remove the lines automatically. Be very
careful in using this option!
--namespace: Filters the search to a given namespace. If this is specified
- multiple times it will search all given namespaces
+In addition, these arguments can be used to restrict changes to some pages:
+
+¶ms;
"""
#
-# (C) Pywikipedia bot team, 2007-2014
+# (C) Pywikibot team, 2007-2014
#
# Distributed under the terms of the MIT license.
#
__version__ = '$Id$'
-#
import pywikibot
-from pywikibot import pagegenerators, i18n
+from pywikibot import i18n, pagegenerators, Bot
from pywikibot.editor import TextEditor
+
+docuReplacements = {
+ '¶ms;': pagegenerators.parameterHelp
+}
+
+
+class SpamRemoveBot(Bot):
+ def __init__(self, generator, spamSite, **kwargs):
+ super(SpamRemoveBot, self).__init__(**kwargs)
+ self.generator = generator
+ self.spamSite = spamSite
+ self.summary = i18n.twtranslate(pywikibot.Site(), 'spamremove-remove',
+ {'url': self.spamSite})
+
+ def run(self):
+ for i, p in enumerate(self.generator, 1):
+ text = p.text
+ if self.spamSite not in text:
+ continue
+ # Show the title of the page we're working on.
+ # Highlight the title in purple.
+ pywikibot.output(u'\n\n>>> \03{lightpurple}%s\03{default} <<<'
+ % p.title())
+ lines = text.split('\n')
+ newpage = []
+ lastok = ''
+ for line in lines:
+ if self.spamSite in line:
+ if lastok:
+ pywikibot.output(lastok)
+ pywikibot.output('\03{lightred}%s\03{default}' % line)
+ lastok = None
+ else:
+ newpage.append(line)
+ if line.strip():
+ if lastok is None:
+ pywikibot.output(line)
+ lastok = line
+ if self.getOption('always'):
+ answer = 'y'
+ else:
+ answer = pywikibot.inputChoice(u'\nDelete the red lines?',
+ ['yes', 'no', 'edit'],
+ ['y', 'N', 'e'], 'n')
+ if answer == 'n':
+ continue
+ elif answer == 'e':
+ editor = TextEditor()
+ newtext = editor.edit(text, highlight=self.spamSite,
+ jumpIndex=text.find(self.spamSite))
+ else:
+ newtext = '\n'.join(newpage)
+ if newtext != text:
+ p.text = newtext
+ p.save(self.summary)
+ else:
+ if 'i' not in locals():
+ pywikibot.output('No page found.')
+ elif i == 1:
+ pywikibot.output('1 page done.')
+ else:
+ pywikibot.output('%d pages done.' % i)
def main():
- always = False
- namespaces = []
- spamSite = ''
- for arg in pywikibot.handleArgs():
- if arg == "-always":
- always = True
- elif arg.startswith('-namespace:'):
- try:
- namespaces.append(int(arg[len('-namespace:'):]))
- except ValueError:
- namespaces.append(arg[len('-namespace:'):])
+ spamSite = None
+ options = {}
+
+ local_args = pywikibot.handleArgs()
+ genFactory = pagegenerators.GeneratorFactory()
+
+ for arg in local_args:
+ if arg == '-always':
+ options['always'] = True
+ elif genFactory.handleArg(arg):
+ continue
else:
spamSite = arg
if not spamSite:
pywikibot.showHelp()
- pywikibot.output(u"No spam site specified.")
+ pywikibot.output(u'No spam site specified.')
return
- mysite = pywikibot.Site()
- pages = mysite.exturlusage(spamSite)
- if namespaces:
- pages = pagegenerators.NamespaceFilterPageGenerator(pages, namespaces)
- pages = pagegenerators.PreloadingGenerator(pages)
+ genFactory.gens.append(pagegenerators.LinksearchPageGenerator(spamSite))
+ generator = genFactory.getCombinedGenerator()
+ generator = pagegenerators.PreloadingGenerator(generator)
- summary = i18n.twtranslate(mysite, 'spamremove-remove',
- {'url': spamSite})
- for i, p in enumerate(pages, 1):
- text = p.text
- if spamSite not in text:
- continue
- # Show the title of the page we're working on.
- # Highlight the title in purple.
- pywikibot.output(u"\n\n>>> \03{lightpurple}%s\03{default} <<<"
- % p.title())
- lines = text.split('\n')
- newpage = []
- lastok = ""
- for line in lines:
- if spamSite in line:
- if lastok:
- pywikibot.output(lastok)
- pywikibot.output('\03{lightred}%s\03{default}' % line)
- lastok = None
- else:
- newpage.append(line)
- if line.strip():
- if lastok is None:
- pywikibot.output(line)
- lastok = line
- if always:
- answer = "y"
- else:
- answer = pywikibot.inputChoice(u'\nDelete the red lines?',
- ['yes', 'no', 'edit'],
- ['y', 'N', 'e'], 'n')
- if answer == "n":
- continue
- elif answer == "e":
- editor = TextEditor()
- newtext = editor.edit(text, highlight=spamSite,
- jumpIndex=text.find(spamSite))
- else:
- newtext = "\n".join(newpage)
- if newtext != text:
- p.text = newtext
- p.save(summary)
- else:
- if "i" not in locals():
- pywikibot.output('No page found.')
- elif i == 1:
- pywikibot.output('1 pages done.')
- else:
- pywikibot.output('%d pages done.' % i)
+ bot = SpamRemoveBot(generator, spamSite, **options)
+ bot.run()
if __name__ == '__main__':
--
To view, visit https://gerrit.wikimedia.org/r/149219
To unsubscribe, visit https://gerrit.wikimedia.org/r/settings
Gerrit-MessageType: newchange
Gerrit-Change-Id: If8e5f694721fcf28256c868f2b1a8eb230227208
Gerrit-PatchSet: 1
Gerrit-Project: pywikibot/core
Gerrit-Branch: master
Gerrit-Owner: Ricordisamoa <[email protected]>
_______________________________________________
MediaWiki-commits mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/mediawiki-commits
