Ricordisamoa has uploaded a new change for review.
https://gerrit.wikimedia.org/r/184543
Change subject: [WIP] script to fix "Pages using duplicate arguments in
template calls"
......................................................................
[WIP] script to fix "Pages using duplicate arguments in template calls"
TODO:
* catch possible typos with TemplateData & Levenshtein distance
* documentation!
Change-Id: I2505afc34eebf32eb8a52bfbf47b3543d923d884
---
A scripts/duplicate_arguments.py
1 file changed, 127 insertions(+), 0 deletions(-)
git pull ssh://gerrit.wikimedia.org:29418/pywikibot/core
refs/changes/43/184543/1
diff --git a/scripts/duplicate_arguments.py b/scripts/duplicate_arguments.py
new file mode 100644
index 0000000..92eef02
--- /dev/null
+++ b/scripts/duplicate_arguments.py
@@ -0,0 +1,127 @@
+# -*- coding: utf-8 -*-
+"""Script to fix template calls with duplicate arguments."""
+#
+# (C) Pywikibot team, 2015
+#
+# Distributed under the terms of the MIT license.
+#
+
+import pywikibot
+from pywikibot import i18n, pagegenerators, Bot
+from pywikibot.data import api
+import mwparserfromhell
+
+
+class DuplicateArgumentsBot(Bot):
+
+ """Bot to fix template calls with duplicate arguments."""
+
+ comment = {
+ 'en': u'remove duplicate arguments in template calls'
+ }
+
+ def __init__(self, generator, **kwargs):
+ self.availableOptions.update({
+ 'action_if_all_empty': None,
+ })
+ super(DuplicateArgumentsBot, self).__init__(**kwargs)
+ self.generator = generator
+
+ def treat(self, page):
+ self.current_page = page
+ code = mwparserfromhell.parse(page.text)
+ edited = False # to prevent unwanted changes
+ for template in code.ifilter_templates():
+ params = {}
+ for param in template.params:
+ params.setdefault(param.name.strip(), []).append(param)
+ for name, items in params.items():
+ if len(items) == 1:
+ # unique parameters
+ continue
+ empty = [param for param in items if param.value.strip() == '']
+ diff = len(items) - len(empty)
+ if diff == 0:
+ choice = self.getOption('action_if_all_empty')
+ if not choice:
+ choices = []
+ for index, param in enumerate(items):
+ choices.append((u'keep only occurrence
#{index}'.format(index=index), str(index)))
+ choices += [(u'keep only the last occurrence', 'l'),
+ (u'remove all occurrences', 'r'),
+ (u'skip these params', 's')]
+ choice = pywikibot.input_choice(
+ u'Found {count} params with name "{name}", all
empty.\n'
+ u'What to do?'.format(count=len(items), name=name),
+ choices)
+ if choice == 's':
+ continue
+ to_remove = items[:]
+ if choice != 'r':
+ to_remove.pop(-1 if choice == 'l' else int(choice))
+ for param in to_remove:
+ template.remove(param)
+ edited = True
+ else:
+ pywikibot.output(
+ u'Found {count} params with name "{name}", '
+ u'of which {empty} are empty.'.format(
+ count=len(items), name=name, empty=len(empty)))
+ if edited:
+ text = unicode(code)
+ comment = i18n.translate(page.site.lang, self.comment,
fallback=True)
+ self.userPut(page, page.text, text, comment=comment)
+
+
+def main(*args):
+ """
+ Process command line arguments and invoke bot.
+
+ If args is an empty list, sys.argv is used.
+
+ @param args: command line arguments
+ @type args: list of unicode
+ """
+ local_args = pywikibot.handle_args(args)
+ genFactory = pagegenerators.GeneratorFactory()
+ message_key = 'duplicate-args-category'
+ tracking = False
+
+ for arg in local_args:
+ if arg == '-tracking':
+ tracking = True
+ else:
+ genFactory.handleArg(arg)
+
+ tracking_gen = None
+ if tracking:
+ site = pywikibot.Site()
+ # TODO FIXME: make APISite.mediawiki_messages() support
'amenableparser'
+ message_query = api.QueryGenerator(
+ site=site,
+ meta='allmessages',
+ ammessages=message_key,
+ amlang=site.lang,
+ amenableparser=True
+ )
+ message = None
+ for item in message_query:
+ if item['name'] == message_key and '*' in item:
+ message = item['*']
+ break
+ if message:
+ category = pywikibot.Category(site, message)
+ tracking_gen = pagegenerators.CategorizedPageGenerator(category)
+ else:
+ raise KeyError(u'Message "{key}" not found on site {site}'.format(
+ key=message_key, site=site))
+ gen = genFactory.getCombinedGenerator(gen=tracking_gen)
+ if gen:
+ bot = DuplicateArgumentsBot(gen)
+ bot.run()
+ else:
+ pywikibot.showHelp()
+
+
+if __name__ == '__main__':
+ main()
--
To view, visit https://gerrit.wikimedia.org/r/184543
To unsubscribe, visit https://gerrit.wikimedia.org/r/settings
Gerrit-MessageType: newchange
Gerrit-Change-Id: I2505afc34eebf32eb8a52bfbf47b3543d923d884
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
