[Pywikipedia-l] SVN: [6270] trunk/pywikipedia
cydeweys at svn.wikimedia.org
cydeweys at svn.wikimedia.org
Mon Jan 19 15:50:18 UTC 2009
Revision: 6270
Author: cydeweys
Date: 2009-01-19 15:50:18 +0000 (Mon, 19 Jan 2009)
Log Message:
-----------
Changing the GeneratorFactory so that instead of handleArg() being called once
and returning a generator, it is called many times to accumulate a list of pages
to work on, and then, at the end, the combined page generator is explicitly
asked for. This makes the GeneratorFactory a lot more useful for a wide variety
of use cases.
Also, GeneratorFactory now understands the -page parameter, so some duplicate
code was removed from bots that handled this individually.
Not in this commit, but coming soon, will be the handling of -namespace in the
GeneratorFactory. This will remove quite a bit of duplicate code from
individual bots and will make most bots more flexible.
Modified Paths:
--------------
trunk/pywikipedia/add_text.py
trunk/pywikipedia/basic.py
trunk/pywikipedia/blockpageschecker.py
trunk/pywikipedia/category.py
trunk/pywikipedia/commonscat.py
trunk/pywikipedia/copyright.py
trunk/pywikipedia/cosmetic_changes.py
trunk/pywikipedia/fixing_redirects.py
trunk/pywikipedia/imagecopy.py
trunk/pywikipedia/imagerecat.py
trunk/pywikipedia/imageuncat.py
trunk/pywikipedia/inline_images.py
trunk/pywikipedia/interwiki.py
trunk/pywikipedia/isbn.py
trunk/pywikipedia/lonelypages.py
trunk/pywikipedia/movepages.py
trunk/pywikipedia/noreferences.py
trunk/pywikipedia/pagegenerators.py
trunk/pywikipedia/piper.py
trunk/pywikipedia/reflinks.py
trunk/pywikipedia/replace.py
trunk/pywikipedia/selflink.py
trunk/pywikipedia/table2wiki.py
trunk/pywikipedia/template.py
trunk/pywikipedia/weblinkchecker.py
Modified: trunk/pywikipedia/add_text.py
===================================================================
--- trunk/pywikipedia/add_text.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/add_text.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -283,7 +283,10 @@
elif arg == '-always':
always = True
else:
- generator = genFactory.handleArg(arg)
+ genFactory.handleArg(arg)
+
+ if not generator:
+ generator = genFactory.getCombinedGenerator()
# Check if there are the minimal settings
if not generator:
raise NoEnoughData('You have to specify the generator you want to use for the script!')
Modified: trunk/pywikipedia/basic.py
===================================================================
--- trunk/pywikipedia/basic.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/basic.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -119,10 +119,7 @@
else:
# check if a standard argument like
# -start:XYZ or -ref:Asdf was given.
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
- else:
+ if not genFactory.handleArg(arg):
pageTitleParts.append(arg)
if pageTitleParts != []:
@@ -131,6 +128,8 @@
page = wikipedia.Page(wikipedia.getSite(), pageTitle)
gen = iter([page])
+ if not gen:
+ gen = genFactory.getCombinedGenerator()
if gen:
# The preloading generator is responsible for downloading multiple
# pages from the wiki simultaneously.
Modified: trunk/pywikipedia/blockpageschecker.py
===================================================================
--- trunk/pywikipedia/blockpageschecker.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/blockpageschecker.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -228,7 +228,7 @@
else:
generator = [wikipedia.Page(wikipedia.getSite(), arg[6:])]
else:
- generator = genFactory.handleArg(arg)
+ genFactory.handleArg(arg)
# Take the right templates to use, the category and the comment
TSP = wikipedia.translate(site, templateSemiProtection)
TTP = wikipedia.translate(site, templateTotalProtection)
@@ -239,6 +239,8 @@
category = wikipedia.translate(site, categoryToCheck)
commentUsed = wikipedia.translate(site, comment)
if not generator:
+ gen = genFactory.getCombinedGenerator()
+ if not generator:
generator = list()
wikipedia.output(u'Loading categories...')
# Define the category if no other generator has been setted
Modified: trunk/pywikipedia/category.py
===================================================================
--- trunk/pywikipedia/category.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/category.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -946,14 +946,19 @@
elif arg == '-recurse':
recurse = True
else:
- gen = genFactory.handleArg(arg)
+ genFactory.handleArg(arg)
if action == 'add':
+ # Note that the add functionality is the only bot that actually uses the
+ # the generator factory. Every other bot creates its own generator exclusively
+ # from the command-line arguments that category.py understands.
if not gen:
- gen = genFactory.handleArg('-links') #default for backwords compatibility
- # The preloading generator is responsible for downloading multiple
- # pages from the wiki simultaneously.
- gen = pagegenerators.PreloadingGenerator(gen)
+ gen = genFactory.getCombinedGenerator()
+ if not gen:
+ genFactory.handleArg('-links') #default for backwords compatibility
+ # The preloading generator is responsible for downloading multiple
+ # pages from the wiki simultaneously.
+ gen = pagegenerators.PreloadingGenerator(genFactory.getCombinedGenerator())
add_category(sort_by_last_name)
elif action == 'remove':
if (fromGiven == False):
Modified: trunk/pywikipedia/commonscat.py
===================================================================
--- trunk/pywikipedia/commonscat.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/commonscat.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -345,8 +345,11 @@
elif arg == '-always':
always = True
else:
- generator = genFactory.handleArg(arg)
+ genFactory.handleArg(arg)
+
if not generator:
+ generator = genFactory.getCombinedGenerator()
+ if not generator:
raise add_text.NoEnoughData('You have to specify the generator you want to use for the script!')
pregenerator = pagegenerators.PreloadingGenerator(generator)
Modified: trunk/pywikipedia/copyright.py
===================================================================
--- trunk/pywikipedia/copyright.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/copyright.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -1145,9 +1145,7 @@
if number < pageNumber:
pageNumber = number
else:
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
+ genFactory.handleArg(arg)
if PageTitles:
pages = [wikipedia.Page(wikipedia.getSite(), PageTitle) for PageTitle in PageTitles]
@@ -1160,6 +1158,8 @@
if ids:
checks_by_ids(ids)
+ if not gen:
+ gen = genFactory.getCombinedGenerator()
if not gen and not ids and not text:
# syntax error, show help text from the top of this file
wikipedia.output(__doc__, 'utf-8')
Modified: trunk/pywikipedia/cosmetic_changes.py
===================================================================
--- trunk/pywikipedia/cosmetic_changes.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/cosmetic_changes.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -393,10 +393,7 @@
genFactory = pagegenerators.GeneratorFactory()
for arg in wikipedia.handleArgs():
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
- else:
+ if not genFactory.handleArg(arg):
pageTitle.append(arg)
# Disabled this check. Although the point is still valid, there
@@ -412,6 +409,8 @@
page = wikipedia.Page(wikipedia.getSite(), ' '.join(pageTitle))
gen = iter([page])
if not gen:
+ gen = genFactory.getCombinedGenerator()
+ if not gen:
wikipedia.showHelp()
elif wikipedia.inputChoice(warning + '\nDo you really want to continue?', ['yes', 'no'], ['y', 'N'], 'N') == 'y':
preloadingGen = pagegenerators.PreloadingGenerator(gen)
Modified: trunk/pywikipedia/fixing_redirects.py
===================================================================
--- trunk/pywikipedia/fixing_redirects.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/fixing_redirects.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -181,10 +181,10 @@
else:
namespace = int(arg[11:])
else:
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
+ genFactory.handleArg(arg)
+ gen = genFactory.getCombinedGenerator()
+
mysite = wikipedia.getSite()
if mysite.sitename() == 'wikipedia:nl':
wikipedia.output(u'\03{lightred}There is consensus on the Dutch Wikipedia that bots should not be used to fix redirects.\03{default}')
Modified: trunk/pywikipedia/imagecopy.py
===================================================================
--- trunk/pywikipedia/imagecopy.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/imagecopy.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -440,15 +440,12 @@
genFactory = pagegenerators.GeneratorFactory()
for arg in wikipedia.handleArgs():
- if arg.startswith('-page'):
- if len(arg) == 5:
- generator = [wikipedia.Page(wikipedia.getSite(), wikipedia.input(u'What page do you want to use?'))]
- else:
- generator = [wikipedia.Page(wikipedia.getSite(), arg[6:])]
- elif arg == '-always':
+ if arg == '-always':
always = True
else:
- generator = genFactory.handleArg(arg)
+ genFactory.handleArg(arg)
+
+ generator = genFactory.getCombinedGenerator()
if not generator:
raise add_text.NoEnoughData('You have to specify the generator you want to use for the script!')
Modified: trunk/pywikipedia/imagerecat.py
===================================================================
--- trunk/pywikipedia/imagerecat.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/imagerecat.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -313,15 +313,12 @@
site = wikipedia.getSite(u'commons', u'commons')
wikipedia.setSite(site)
for arg in wikipedia.handleArgs():
- if arg.startswith('-page'):
- if len(arg) == 5:
- generator = [wikipedia.Page(site, wikipedia.input(u'What page do you want to use?'))]
- else:
- generator = [wikipedia.Page(site, arg[6:])]
- elif arg == '-onlyfilter':
+ if arg == '-onlyfilter':
onlyfilter = True
else:
- generator = genFactory.handleArg(arg)
+ genFactory.handleArg(arg)
+
+ gen = genFactory.getCombinedGenerator()
if not generator:
generator = pagegenerators.CategorizedPageGenerator(catlib.Category(site, u'Category:Media needing categories'), recurse=True)
Modified: trunk/pywikipedia/imageuncat.py
===================================================================
--- trunk/pywikipedia/imageuncat.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/imageuncat.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -1339,18 +1339,16 @@
site = wikipedia.getSite(u'commons', u'commons')
wikipedia.setSite(site)
for arg in wikipedia.handleArgs():
- if arg.startswith('-page'):
- if len(arg) == 5:
- generator = [wikipedia.Page(site, wikipedia.input(u'What page do you want to use?'))]
- else:
- generator = [wikipedia.Page(site, arg[6:])]
- elif arg.startswith('-yesterday'):
+ if arg.startswith('-yesterday'):
generator = uploadedYesterday(site)
elif arg.startswith('-recentchanges'):
generator = recentChanges(site)
else:
- generator = genFactory.handleArg(arg)
+ genFactory.handleArg(arg)
+
if not generator:
+ generator = genFactory.getCombinedGenerator()
+ if not generator:
wikipedia.output('You have to specify the generator you want to use for the program!')
else:
pregenerator = pagegenerators.PreloadingGenerator(generator)
Modified: trunk/pywikipedia/inline_images.py
===================================================================
--- trunk/pywikipedia/inline_images.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/inline_images.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -88,10 +88,7 @@
genFactory = pagegenerators.GeneratorFactory()
for arg in wikipedia.handleArgs():
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
- else:
+ if not genFactory.handleArg(arg):
pageTitle.append(arg)
if pageTitle:
@@ -99,6 +96,8 @@
page = wikipedia.Page(wikipedia.getSite(), ' '.join(pageTitle))
gen = iter([page])
if not gen:
+ gen = genFactory.getCombinedGenerator()
+ if not gen:
wikipedia.showHelp('inline_images')
else:
preloadingGen = pagegenerators.PreloadingGenerator(gen)
Modified: trunk/pywikipedia/interwiki.py
===================================================================
--- trunk/pywikipedia/interwiki.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/interwiki.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -1726,12 +1726,10 @@
elif arg == '-back':
globalvar.nobackonly = True
else:
- generator = genFactory.handleArg(arg)
- if generator:
- hintlessPageGen = generator
- else:
+ if not genFactory.handleArg(arg):
singlePageTitle.append(arg)
+
# ensure that we don't try to change main page
try:
site = wikipedia.getSite()
@@ -1764,6 +1762,8 @@
bot = InterwikiBot()
+ if not hintlessPageGen:
+ hintlessPageGen = genFactory.getCombinedGenerator()
if hintlessPageGen:
if len(namespaces) > 0:
hintlessPageGen = pagegenerators.NamespaceFilterPageGenerator(hintlessPageGen, namespaces)
Modified: trunk/pywikipedia/isbn.py
===================================================================
--- trunk/pywikipedia/isbn.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/isbn.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -1473,16 +1473,15 @@
elif arg == '-format':
format = True
else:
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
- else:
+ if not genFactory.handleArg(arg):
pageTitle.append(arg)
if pageTitle:
page = wikipedia.Page(wikipedia.getSite(), ' '.join(pageTitle))
gen = iter([page])
if not gen:
+ gen = genFactory.getCombinedGenerator()
+ if not gen:
wikipedia.showHelp('isbn')
else:
if namespaces != []:
Modified: trunk/pywikipedia/lonelypages.py
===================================================================
--- trunk/pywikipedia/lonelypages.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/lonelypages.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -134,19 +134,18 @@
nwlimit = int(arg[10:])
generator = wikipedia.getSite().newpages(number = nwlimit)
nwpages = True
- elif arg.startswith('-page'):
- if len(arg) == 5:
- generator = [wikipedia.Page(wikipedia.getSite(), wikipedia.input(u'How many pages do you want to check?'))]
- else:
- generator = [wikipedia.Page(wikipedia.getSite(), arg[6:])]
elif arg == '-always':
always = True
else:
- generator = genFactory.handleArg(arg)
+ genFactory.handleArg(arg)
# Retrive the site
wikiSite = wikipedia.getSite()
+
+ if not generator:
+ generator = genFactory.getCombinedGenerator()
+
# If the generator is not given, use the default one
- if generator == None:
+ if not generator:
generator = wikiSite.lonelypages(repeat = True, number = limit)
# Take the configurations according to our project
comment = wikipedia.translate(wikiSite, commento)
Modified: trunk/pywikipedia/movepages.py
===================================================================
--- trunk/pywikipedia/movepages.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/movepages.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -269,9 +269,7 @@
else:
summary = arg[9:]
else:
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
+ genFactory.handleArg(arg)
if oldName:
wikipedia.output(u'WARNING: -from:%s without -to:' % oldName)
@@ -279,6 +277,9 @@
page = wikipedia.Page(wikipedia.getSite(), pair[0])
bot = MovePagesBot(None, prefix, delete, always, skipredirects, summary)
bot.moveOne(page, pair[1])
+
+ if not gen:
+ gen = genFactory.getCombinedGenerator()
if gen:
preloadingGen = pagegenerators.PreloadingGenerator(gen)
bot = MovePagesBot(preloadingGen, prefix, delete, always, skipredirects, summary)
Modified: trunk/pywikipedia/noreferences.py
===================================================================
--- trunk/pywikipedia/noreferences.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/noreferences.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -453,16 +453,15 @@
elif arg == '-always':
always = True
else:
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
- else:
+ if not genFactory.handleArg(arg):
pageTitle.append(arg)
if pageTitle:
page = wikipedia.Page(wikipedia.getSite(), ' '.join(pageTitle))
gen = iter([page])
if not gen:
+ gen = genFactory.getCombinedGenerator()
+ if not gen:
wikipedia.showHelp('noreferences')
else:
if namespaces != []:
Modified: trunk/pywikipedia/pagegenerators.py
===================================================================
--- trunk/pywikipedia/pagegenerators.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/pagegenerators.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -796,20 +796,32 @@
to work on.
"""
def __init__(self):
- pass
+ self.gens = []
- def setCategoryGen(self, arg, length, recurse = False):
+ """
+ This function returns the combination of all accumulated generators
+ that have been created in the process of handling arguments.
+ Only call it after all arguments have been parsed.
+ """
+ def getCombinedGenerator(self):
+ if (len(self.gens) == 0):
+ return None
+ elif (len(self.gens) == 1):
+ return DuplicateFilterPageGenerator(self.gens[0])
+ else:
+ return DuplicateFilterPageGenerator(CombinedPageGenerator(self.gens))
+
+ def getCategoryGen(self, arg, length, recurse = False):
if len(arg) == length:
categoryname = wikipedia.input(u'Please enter the category name:')
else:
categoryname = arg[length + 1:]
ind = categoryname.find('|')
+ startfrom = None
if ind > 0:
startfrom = categoryname[ind + 1:]
categoryname = categoryname[:ind]
- else:
- startfrom = None
cat = catlib.Category(wikipedia.getSite(), 'Category:%s' % categoryname)
return CategorizedPageGenerator(cat, start = startfrom, recurse = recurse)
@@ -830,6 +842,14 @@
cat = catlib.Category(wikipedia.getSite(), 'Category:%s' % categoryname)
return SubCategoriesPageGenerator(cat, start = startfrom, recurse = recurse)
+ """
+ This function parses one argument at a time. If it is recognized as an
+ argument that specifies a generator, a generator is created and added
+ to the accumulation list, and the function returns true. Otherwise, it
+ returns false, so that callee can try parsing the argument.
+ Call getCombinedGenerator() after all arguments have been parsed to get
+ the final output generator.
+ """
def handleArg(self, arg):
gen = None
if arg.startswith('-filelinks'):
@@ -872,16 +892,23 @@
textfilename = wikipedia.input(u'Please enter the local file name:')
gen = TextfilePageGenerator(textfilename)
elif arg.startswith('-catr'):
- gen = self.setCategoryGen(arg, 5, recurse = True)
+ gen = self.getCategoryGen(arg, 5, recurse = True)
elif arg.startswith('-cat'):
- gen = self.setCategoryGen(arg, 4)
+ gen = self.getCategoryGen(arg, len('-cat'))
+ elif arg.startswith('-category'):
+ gen = self.getCategoryGen(arg, len('-category'))
elif arg.startswith('-subcatsr'):
gen = self.setSubCategoriesGen(arg, 9, recurse = True)
elif arg.startswith('-subcats'):
gen = self.setSubCategoriesGen(arg, 8)
# This parameter is deprecated, catr should be used instead.
elif arg.startswith('-subcat'):
- gen = self.setCategoryGen(arg, 7, recurse = True)
+ gen = self.getCategoryGen(arg, 7, recurse = True)
+ elif arg.startswith('-page'):
+ if len(arg) == len('-page'):
+ gen = [wikipedia.Page(wikipedia.getSite(), wikipedia.input(u'What page do you want to use?'))]
+ else:
+ gen = [wikipedia.Page(wikipedia.getSite(), arg[len('-page:'):])]
elif arg.startswith('-uncatfiles'):
gen = UnCategorizedImageGenerator()
elif arg.startswith('-uncatcat'):
@@ -958,10 +985,12 @@
elif arg.startswith('-yahoo'):
gen = YahooSearchPageGenerator(arg[7:])
else:
- return None
- # make sure all yielded pages are unique
- gen = DuplicateFilterPageGenerator(gen)
- return gen
+ pass
+ if gen:
+ self.gens.append(gen)
+ return True
+ else:
+ return False
if __name__ == "__main__":
try:
Modified: trunk/pywikipedia/piper.py
===================================================================
--- trunk/pywikipedia/piper.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/piper.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -188,10 +188,7 @@
else:
# check if a standard argument like
# -start:XYZ or -ref:Asdf was given.
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
- else:
+ if not genFactory.handleArg(arg):
pageTitleParts.append(arg)
if pageTitleParts != []:
@@ -200,6 +197,8 @@
page = wikipedia.Page(wikipedia.getSite(), pageTitle)
gen = iter([page])
+ if not gen:
+ gen = genFactory.getCombinedGenerator()
if gen:
# The preloading generator is responsible for downloading multiple
# pages from the wiki simultaneously.
Modified: trunk/pywikipedia/reflinks.py
===================================================================
--- trunk/pywikipedia/reflinks.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/reflinks.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -678,9 +678,7 @@
namespaces = []
generator = None
for arg in wikipedia.handleArgs():
- if arg.startswith('-page:'):
- PageTitles.append(arg[6:])
- elif arg.startswith('-namespace:'):
+ if arg.startswith('-namespace:'):
try:
namespaces.append(int(arg[11:]))
except ValueError:
@@ -706,7 +704,7 @@
else:
xmlFilename = arg[5:]
else:
- generator = genFactory.handleArg(arg)
+ genFactory.handleArg(arg)
if xmlFilename:
try:
@@ -714,10 +712,9 @@
except NameError:
xmlStart = None
generator = XmlDumpPageGenerator(xmlFilename, xmlStart, namespaces)
- elif PageTitles:
- pages= [wikipedia.Page(wikipedia.getSite(), PageTitle) for PageTitle in PageTitles]
- generator = iter(pages)
if not generator:
+ generator = genFactory.getCombinedGenerator()
+ if not generator:
# syntax error, show help text from the top of this file
wikipedia.showHelp('reflinks')
return
Modified: trunk/pywikipedia/replace.py
===================================================================
--- trunk/pywikipedia/replace.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/replace.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -16,10 +16,6 @@
Argument can also be given as "-page:pagetitle". You can
give this parameter multiple times to edit multiple pages.
--category Works on all of the pages in a specific category. Specify
- this argument multiple times to work on multiple categories
- simultaneously.
-
Furthermore, the following command line parameters are supported:
-regex Make replacements using regular expressions. If this argument
@@ -515,8 +511,6 @@
# Between a regex and another (using -fix) sleep some time (not to waste
# too much CPU
sleep = None
- # A list of categories whose pages we should process.
- categories = []
# Read commandline parameters.
for arg in wikipedia.handleArgs(*args):
@@ -536,8 +530,6 @@
xmlFilename = arg[5:]
elif arg =='-sql':
useSql = True
- elif arg.startswith('-category'):
- categories.append(arg[len('-category:'):])
elif arg.startswith('-page'):
if len(arg) == 5:
PageTitles.append(wikipedia.input(
@@ -581,10 +573,7 @@
elif arg.startswith('-allowoverlap'):
allowoverlap = True
else:
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
- else:
+ if not genFactory.handleArg(arg):
commandline_replacements.append(arg)
if (len(commandline_replacements) % 2):
@@ -704,15 +693,14 @@
%s
LIMIT 200""" % (whereClause, exceptClause)
gen = pagegenerators.MySQLPageGenerator(query)
- elif categories:
- gens = [pagegenerators.CategorizedPageGenerator(catlib.Category(wikipedia.getSite(), 'Category:' + t)) for t in categories]
- gen = pagegenerators.DuplicateFilterPageGenerator(pagegenerators.CombinedPageGenerator(gens))
elif PageTitles:
pages = [wikipedia.Page(wikipedia.getSite(), PageTitle)
for PageTitle in PageTitles]
gen = iter(pages)
if not gen:
+ gen = genFactory.getCombinedGenerator()
+ if not gen:
# syntax error, show help text from the top of this file
wikipedia.showHelp('replace')
return
Modified: trunk/pywikipedia/selflink.py
===================================================================
--- trunk/pywikipedia/selflink.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/selflink.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -224,16 +224,15 @@
except ValueError:
namespaces.append(arg[11:])
else:
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
- else:
+ if not genFactory.handleArg(arg):
pageTitle.append(arg)
if pageTitle:
page = wikipedia.Page(wikipedia.getSite(), ' '.join(pageTitle))
gen = iter([page])
if not gen:
+ gen = genFactory.getCombinedGenerator()
+ if not gen:
wikipedia.showHelp('selflink')
else:
if namespaces != []:
Modified: trunk/pywikipedia/table2wiki.py
===================================================================
--- trunk/pywikipedia/table2wiki.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/table2wiki.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -584,10 +584,7 @@
elif arg.startswith('-debug'):
debug = True
else:
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
- else:
+ if not genFactory.handleArg(arg):
page_title.append(arg)
# if the page is given as a command line argument,
@@ -596,7 +593,10 @@
page_title = ' '.join(page_title)
page = wikipedia.Page(wikipedia.getSite(), page_title)
gen = iter([page])
+
if not gen:
+ gen = genFactory.getCombinedGenerator()
+ if not gen:
# show help
wikipedia.showHelp('table2wiki')
sys.exit(0)
Modified: trunk/pywikipedia/template.py
===================================================================
--- trunk/pywikipedia/template.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/template.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -29,10 +29,6 @@
-always Don't bother asking to confirm any of the changes, Just Do It.
--page: Only edit a specific page. You can use this argument multiple times to work
- on multiple pages. If the page title has spaces in it, enclose the entire
- page name in quotes.
-
-category: Appends the given category to every page that is edited. This is useful when
a category is being broken out from a template parameter or when templates are
being upmerged but more information must be preserved.
@@ -359,7 +355,7 @@
editSummary = ''
addedCat = ''
acceptAll = False
- pageTitles = []
+ genFactory = pagegenerators.GeneratorFactory()
# If xmlfilename is None, references will be loaded from the live wiki.
xmlfilename = None
# read command line parameters
@@ -384,13 +380,9 @@
editSummary = arg[len('-summary:'):]
elif arg.startswith('-always'):
acceptAll = True
- elif arg.startswith('-page'):
- if len(arg) == len('-page'):
- pageTitles.append(wikipedia.input(u'Which page do you want to chage?'))
- else:
- pageTitles.append(arg[len('-page:'):])
else:
- templateNames.append(arg)
+ if not genFactory.handleArg(arg):
+ templateNames.append(arg)
if subst or remove:
for templateName in templateNames:
@@ -411,10 +403,9 @@
if xmlfilename:
gen = XmlDumpTemplatePageGenerator(oldTemplates, xmlfilename)
- elif pageTitles:
- pages = [wikipedia.Page(wikipedia.getSite(), pageTitle) for pageTitle in pageTitles]
- gen = iter(pages)
else:
+ gen = genFactory.getCombinedGenerator()
+ if not gen:
gens = []
gens = [pagegenerators.ReferringPageGenerator(t, onlyTemplateInclusion = True) for t in oldTemplates]
gen = pagegenerators.CombinedPageGenerator(gens)
Modified: trunk/pywikipedia/weblinkchecker.py
===================================================================
--- trunk/pywikipedia/weblinkchecker.py 2009-01-19 12:53:43 UTC (rev 6269)
+++ trunk/pywikipedia/weblinkchecker.py 2009-01-19 15:50:18 UTC (rev 6270)
@@ -799,10 +799,7 @@
elif arg.startswith('-ignore:'):
HTTPignore.append(int(arg[8:]))
else:
- generator = genFactory.handleArg(arg)
- if generator:
- gen = generator
- else:
+ if not genFactory.handleArg(arg):
singlePageTitle.append(arg)
if singlePageTitle:
@@ -810,6 +807,8 @@
page = wikipedia.Page(wikipedia.getSite(), singlePageTitle)
gen = iter([page])
+ if not gen:
+ gen = genFactory.getCombinedGenerator()
if gen:
if namespaces != []:
gen = pagegenerators.NamespaceFilterPageGenerator(gen, namespaces)
More information about the Pywikipedia-l
mailing list