pywikibot September 2007

pywikibot@lists.wikimedia.org

24 participants
397 discussions

[Pywikipedia-l] SVN: [4265] trunk/pywikipedia/isbn.py
by wikipedian＠svn.wikimedia.org 13 Sep '07

13 Sep '07

Revision: 4265 Author: wikipedian Date: 2007-09-13 09:34:39 +0000 (Thu, 13 Sep 2007) Log Message: ----------- handle ISBN10s with small x as checksum Modified Paths: -------------- trunk/pywikipedia/isbn.py Modified: trunk/pywikipedia/isbn.py =================================================================== --- trunk/pywikipedia/isbn.py 2007-09-13 09:34:00 UTC (rev 4264) +++ trunk/pywikipedia/isbn.py 2007-09-13 09:34:39 UTC (rev 4265) @@ -1275,7 +1275,7 @@ """ result = [] for c in self.code: - if c.isdigit() or c == 'X': + if c.isdigit() or c in 'Xx': result.append(c) elif c != '-': raise InvalidIsbnException('The ISBN %s contains invalid characters.' % self.code) @@ -1295,13 +1295,13 @@ #print checksum lastDigit = self.digits()[-1] #print lastDigit - if not ((checksum == 10 and lastDigit == 'X') or (lastDigit.isdigit() and checksum == int(lastDigit))): + if not ((checksum == 10 and lastDigit in 'Xx') or (lastDigit.isdigit() and checksum == int(lastDigit))): raise InvalidIsbnException('The ISBN checksum of %s is incorrect.' % self.code) def checkValidity(self): if len(self.digits()) != 10: raise InvalidIsbnException('The ISBN %s is not 10 digits long.' % self.code) - if 'X' in self.digits()[:-1]: + if 'X' in self.digits()[:-1] or 'x' in self.digits()[:-1]: raise InvalidIsbnException('ISBN %s: X is only allowed at the end of the ISBN.' % self.code) self.checkChecksum() @@ -1313,11 +1313,18 @@ ISBN number. """ code = '978-' + self.code[:-1] - + #cs = self.calculateChecksum() #code += str(cs) return ISBN13(code, checksumMissing = True) + def format(self): + # load overridden superclass method + ISBN.format(self) + # capitalize checksum + if self.code[-1] == 'x': + self.code = self.code[:-1] + 'X' + def getIsbn(code): try: i = ISBN13(code) @@ -1342,7 +1349,7 @@ return i.code def hyphenateIsbnNumbers(text): - isbnR = re.compile(r'(?<=ISBN )(?P<code>[\d\-]+X?)') + isbnR = re.compile(r'(?<=ISBN )(?P<code>[\d\-]+[Xx]?)') text = isbnR.sub(_hyphenateIsbnNumber, text) return text @@ -1360,7 +1367,7 @@ return i13.code def convertIsbn10toIsbn13(text): - isbnR = re.compile(r'(?<=ISBN )(?P<code>[\d\-]+X?)') + isbnR = re.compile(r'(?<=ISBN )(?P<code>[\d\-]+[Xx]?)') text = isbnR.sub(_isbn10toIsbn13, text) return text @@ -1371,7 +1378,7 @@ self.to13 = to13 self.format = format self.always = always - self.isbnR = re.compile(r'(?<=ISBN )(?P<code>[\d\-]+X?)') + self.isbnR = re.compile(r'(?<=ISBN )(?P<code>[\d\-]+[Xx]?)') def treat(self, page): try:

1 0

[Pywikipedia-l] SVN: [4264] trunk/pywikipedia/fixes.py
by wikipedian＠svn.wikimedia.org 13 Sep '07

13 Sep '07

Revision: 4264 Author: wikipedian Date: 2007-09-13 09:34:00 +0000 (Thu, 13 Sep 2007) Log Message: ----------- improved 'isbn' fix Modified Paths: -------------- trunk/pywikipedia/fixes.py Modified: trunk/pywikipedia/fixes.py =================================================================== --- trunk/pywikipedia/fixes.py 2007-09-13 08:46:41 UTC (rev 4263) +++ trunk/pywikipedia/fixes.py 2007-09-13 09:34:00 UTC (rev 4264) @@ -270,11 +270,14 @@ 'en': u'Robot: Fixing ISBN format', }, 'replacements': [ - (r'ISBN (\d+) (\d+) (\d+) (\d+) (\d)', r'ISBN \1-\2-\3-\4-\5'), # missing hyphens in ISBN13 - (r'ISBN (\d+) (\d+) (\d+) (\d|X|x)', r'ISBN \1-\2-\3-\4'), # missing hyphens in ISBN10 - (ur'ISBN (\d+)–(\d+)–(\d+)–(\d+)–(\d)', r'ISBN \1-\2-\3-\4-\5'), # dashes instead of hyphens in ISBN13 - (ur'ISBN (\d+)–(\d+)–(\d+)–(\d|X|x)', r'ISBN \1-\2-\3-\4'), # dashes instead of hyphens in ISBN10 - (r'ISBN: (\d+)', r'ISBN \1'), # colon + # colon + (r'ISBN: (\d+)', r'ISBN \1'), + # Spaces, dashes, or dots instead of hyphens as separators, + # or spaces between digits and separators. + # Note that these regular expressions also match valid ISBNs, but + # these won't be changed. + (ur'ISBN (\d+) *[\- –\.] *(\d+) *[\- –\.] *(\d+) *[\- –\.] *(\d+) *[\- –\.] *(\d)(?!\d)', r'ISBN \1-\2-\3-\4-\5'), # ISBN13 + (r'ISBN (\d+) *[\- –\.] *(\d+) *[\- –\.] *(\d+) *[\- –\.] *(\d|X|x)(?!\d)', r'ISBN \1-\2-\3-\4'), # ISBN10 ], }, }

1 0

[Pywikipedia-l] SVN: [4263] trunk/pywikipedia/fixes.py
by wikipedian＠svn.wikimedia.org 13 Sep '07

13 Sep '07

Revision: 4263 Author: wikipedian Date: 2007-09-13 08:46:41 +0000 (Thu, 13 Sep 2007) Log Message: ----------- added 'isbn' fix Modified Paths: -------------- trunk/pywikipedia/fixes.py Modified: trunk/pywikipedia/fixes.py =================================================================== --- trunk/pywikipedia/fixes.py 2007-09-13 08:46:31 UTC (rev 4262) +++ trunk/pywikipedia/fixes.py 2007-09-13 08:46:41 UTC (rev 4263) @@ -262,6 +262,19 @@ u'[[17. Juni 1953]]', u'[[11. September 2001]]', ] - }, + 'isbn': { + 'regex': True, + 'msg': { + 'de': u'Bot: Korrigiere ISBN-Format', + 'en': u'Robot: Fixing ISBN format', + }, + 'replacements': [ + (r'ISBN (\d+) (\d+) (\d+) (\d+) (\d)', r'ISBN \1-\2-\3-\4-\5'), # missing hyphens in ISBN13 + (r'ISBN (\d+) (\d+) (\d+) (\d|X|x)', r'ISBN \1-\2-\3-\4'), # missing hyphens in ISBN10 + (ur'ISBN (\d+)–(\d+)–(\d+)–(\d+)–(\d)', r'ISBN \1-\2-\3-\4-\5'), # dashes instead of hyphens in ISBN13 + (ur'ISBN (\d+)–(\d+)–(\d+)–(\d|X|x)', r'ISBN \1-\2-\3-\4'), # dashes instead of hyphens in ISBN10 + (r'ISBN: (\d+)', r'ISBN \1'), # colon + ], + }, }

1 0

[Pywikipedia-l] SVN: [4262] trunk/pywikipedia/isbn.py
by wikipedian＠svn.wikimedia.org 13 Sep '07

13 Sep '07

Revision: 4262 Author: wikipedian Date: 2007-09-13 08:46:31 +0000 (Thu, 13 Sep 2007) Log Message: ----------- bugfix Modified Paths: -------------- trunk/pywikipedia/isbn.py Modified: trunk/pywikipedia/isbn.py =================================================================== --- trunk/pywikipedia/isbn.py 2007-09-13 08:46:15 UTC (rev 4261) +++ trunk/pywikipedia/isbn.py 2007-09-13 08:46:31 UTC (rev 4262) @@ -1356,8 +1356,8 @@ except InvalidIsbnException: # don't change return code - i.format() - return i.code + i13 = i.toISBN13() + return i13.code def convertIsbn10toIsbn13(text): isbnR = re.compile(r'(?<=ISBN )(?P<code>[\d\-]+X?)') @@ -1386,6 +1386,7 @@ newText = oldText if self.to13: newText = self.isbnR.sub(_isbn10toIsbn13, newText) + if self.format: newText = self.isbnR.sub(_hyphenateIsbnNumber, newText) self.save(page, newText)

1 0

[Pywikipedia-l] SVN: [4261] trunk/pywikipedia/editarticle.py
by wikipedian＠svn.wikimedia.org 13 Sep '07

13 Sep '07

Revision: 4261 Author: wikipedian Date: 2007-09-13 08:46:15 +0000 (Thu, 13 Sep 2007) Log Message: ----------- made edit summary prefix/suffix translatable Modified Paths: -------------- trunk/pywikipedia/editarticle.py Modified: trunk/pywikipedia/editarticle.py =================================================================== --- trunk/pywikipedia/editarticle.py 2007-09-12 22:12:11 UTC (rev 4260) +++ trunk/pywikipedia/editarticle.py 2007-09-13 08:46:15 UTC (rev 4261) @@ -15,8 +15,6 @@ __metaclass__ = type __version__ = "$Id$" -sig = u" (edited with editarticle.py 0.4)" - import sys import os import string @@ -26,6 +24,11 @@ import wikipedia import config +msg = { + 'de': 'Manuelle Bearbeitung: %s', + 'en': 'Manual edit: %s', +} + class TextEditor: def __init__(self): pass @@ -142,35 +145,6 @@ if not self.options.edit_redirect and self.page.isRedirectPage(): self.page = wikipedia.Page(site, self.page.getRedirectTarget()) - def repair(self, content): - """ - Removes single newlines. - """ - ##### - # This method was disabled because its functionality belong into - # cosmetic_changes.py, not here. - return content - # - ##### - if self.options.join_lines: - lines = content.splitlines() - result = [] - for i, line in enumerate(lines): - try: - nextline = lines[i+1] - except IndexError: - nextline = "last" - result.append(line) - if line.strip() == "" or line[0] not in self.joinchars or \ - nextline.strip() == "" or nextline[0] not in self.joinchars: - result.append('\n') - else: - result.append(" ") - s = "".join(result) - else: - s = content - return s - def handle_edit_conflict(self): fn = os.path.join(tempfile.gettempdir(), self.page.title()) fp = open(fn, 'w') @@ -186,9 +160,9 @@ textEditor = TextEditor() new = textEditor.edit(old) if new and old != new: - new = self.repair(new) wikipedia.showDiff(old, new) - comment = wikipedia.input(u"What did you change?") + sig + changes = wikipedia.input(u"What did you change?") + comment = wikipedia.translate(wikipedia.getSite(), msg) % changes try: self.page.put(new, comment = comment, minorEdit = False, watchArticle=self.options.watch) except wikipedia.EditConflict:

1 0

[Pywikipedia-l] [ pywikipediabot-Bugs-1773949 ] wrong argument description of movepages.py
by SourceForge.net 13 Sep '07

13 Sep '07

Bugs item #1773949, was opened at 2007-08-14 15:31 Message generated for change (Comment added) made by falk_steinhauer You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=603138&aid=1773949&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None >Status: Open Resolution: Accepted >Priority: 3 Private: No Submitted By: Falk Steinhauer (falk_steinhauer) Assigned to: Leonardo Gregianin (leogregianin) Summary: wrong argument description of movepages.py Initial Comment: I am using snapshot 2007-06-19. The commandline help of movepages.py shows an option called "-addprefix", but the respective option in the sourcecode is spelled "-prefix". Users that are not able to understand the sourcecode might not be able to understand how to call the script. Another thing is, that this script is not using page title highlighting like replace.py when user interaction is desired. The first line in "def treat(self,page)" should be: colors = [None] * 6 + [13] * len(page.title()) + [None] * 4 ---------------------------------------------------------------------- >Comment By: Falk Steinhauer (falk_steinhauer) Date: 2007-09-13 08:06 Message: Logged In: YES user_id=1810075 Originator: YES Bug still exists in snapshot 2007-08-11. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=603138&aid=1773949&group_…

1 0

[Pywikipedia-l] SVN: [4260] trunk/pywikipedia
by wikipedian＠svn.wikimedia.org 12 Sep '07

12 Sep '07

Revision: 4260 Author: wikipedian Date: 2007-09-12 22:12:11 +0000 (Wed, 12 Sep 2007) Log Message: ----------- created methods to correctly set hyphens in ISBNs fixed ISBN-10 -> ISBN-13 conversion activated ISBN hyphen setting in cosmetic_changes Modified Paths: -------------- trunk/pywikipedia/cosmetic_changes.py trunk/pywikipedia/isbn.py Modified: trunk/pywikipedia/cosmetic_changes.py =================================================================== --- trunk/pywikipedia/cosmetic_changes.py 2007-09-12 19:54:07 UTC (rev 4259) +++ trunk/pywikipedia/cosmetic_changes.py 2007-09-12 22:12:11 UTC (rev 4260) @@ -25,7 +25,7 @@ all of them, but be careful if you do. """ __version__ = '$Id$' -import wikipedia, pagegenerators +import wikipedia, pagegenerators, isbn import sys import re @@ -91,6 +91,7 @@ text = self.resolveHtmlEntities(text) text = self.validXhtml(text) text = self.removeUselessSpaces(text) + text = isbn.hyphenateIsbnNumbers(text) if self.debug: wikipedia.showDiff(oldText, text) return text @@ -280,6 +281,7 @@ return text + def cleanUpSectionHeaders(self, text): """ For better readability of section header source code, puts a space Modified: trunk/pywikipedia/isbn.py =================================================================== --- trunk/pywikipedia/isbn.py 2007-09-12 19:54:07 UTC (rev 4259) +++ trunk/pywikipedia/isbn.py 2007-09-12 22:12:11 UTC (rev 4260) @@ -2,52 +2,1162 @@ # -*- coding: utf-8 -*- """ -This bot goes over multiple pages of the home wiki, and converts all ISBN-10 -codes to the ISBN-13 format. +This script goes over multiple pages of the home wiki, and reports invalid +ISBN numbers. -This script understands various command-line arguments: +Additionally, it can convert all ISBN-10 codes to the ISBN-13 format, and +correct the ISBN format by placing hyphens. - -start: used as -start:page_name, specifies that the robot should - go alphabetically through all pages on the home wiki, - starting at the named page. +These command line parameters can be used to specify which pages to work on: - -file: used as -file:file_name, read a list of pages to treat - from the named textfile. Page titles should be enclosed - in [[double-squared brackets]]. +&params; - -ref: used as -start:page_name, specifies that the robot should - touch all pages referring to the named page. +-namespace:n Number or name of namespace to process. The parameter can be + used multiple times. It works in combination with all other + parameters, except for the -start parameter. If you e.g. + want to iterate over all categories starting at M, use + -start:Category:M. - -links: used as -links:page_name, specifies that the robot should - touch all pages referred to from the named page. +Furthermore, the following command line parameters are supported: - -cat: used as -cat:category_name, specifies that the robot should - touch all pages in the named category. +-to13 Converts all ISBN-10 codes to ISBN-13. + NOTE: This needn't be done, as MediaWiki still supports + (and will keep supporting) ISBN-10, and all libraries and + bookstores will most likely do so as well. -All other parameters will be regarded as a page title; in this case, the bot -will only touch a single page. +-format Corrects the hyphenation. + NOTE: This is in here for testing purposes only. Usually + it's not worth to create an edit for such a minor issue. + The recommended way of doing this is enabling + cosmetic_changes, so that these changes are made on-the-fly + to all pages that are modified. -##################################################################### -# ATTENTION # -##################################################################### -# The ISBN-13 standard is scheduled to be used as of 2007-01-01. # -# I expect that some libraries and online bookstores will have some # -# problems with the conversion, so we shouldn't start changing # -# ISBNs at new year's eve. # -# As most online bookstores and library catalogs will probably be # -# backwards compatible, we shouldn't change anything before all # -# major sites have managed the conversion process. # -##################################################################### +-always Don't prompt you for each replacement. -The bot code is unfinished, although the ISBN conversion code seems -to work fine already. --Daniel """ __version__='$Id$' import wikipedia, pagegenerators, catlib, config -import sys +import sys, re +docuReplacements = { + '&params;': pagegenerators.parameterHelp, +} + +# Summary messages in different languages +msg = { + 'de': 'Bot: Formatiere ISBN', + 'en': 'Robot: Formatting ISBN', +} + +# Maps each group number to the list of its publisher number ranges. +# Taken from http://www.isbn-international.org/converter/ranges.htm +ranges = { + '0': [ # English speaking area + ('00', '19'), + ('200', '699'), + ('7000', '8499'), + ('85000', '89999'), + ('900000', '949999'), + ('9500000', '9999999'), + ], + '1': [ # English speaking area + ('00', '09'), + ('100', '399'), + ('4000', '5499'), + ('55000', '86979'), + ('869800', '998999'), + ], + '2': [ # French speaking area + ('00', '19'), + ('200', '349'), + ('35000', '39999'), + ('400', '699'), + ('7000', '8399'), + ('84000', '89999'), + ('900000', '949999'), + ('9500000', '9999999'), + ], + '3': [ # German speaking area + ('00', '02'), + ('030', '033'), + ('0340', '0369'), + ('03700', '03999'), + ('04', '19'), + ('200', '699'), + ('7000', '8499'), + ('85000', '89999'), + ('900000', '949999'), + ('9500000', '9999999'), + ], + '4': [ # Japan + ('00', '19'), + ('200', '699'), + ('7000', '8499'), + ('85000', '89999'), + ('900000', '949999'), + ('9500000', '9999999'), + ], + '5': [ # Russian Federation + ('00', '19'), + ('200', '699'), + ('7000', '8499'), + ('85000', '89999'), + ('900000', '909999'), + ('91000', '91999'), + ('9200', '9299'), + ('93000', '94999'), + ('9500', '9799'), + ('98000', '98999'), + ('9900000', '9909999'), + ('9910', '9999'), + ], + '600': [ # Iran + ('00', '09'), + ('100', '499'), + ('5000', '8999'), + ('90000', '99999'), + ], + '601': [ # Kazakhstan + ('00', '19'), + ('200', '699'), + ('7000', '7999'), + ('80000', '84999'), + ('85', '99'), + ], + '602': [ # Indonesia + ('00', '19'), + ('200', '799'), + ('8000', '9499'), + ('95000', '99999'), + ], + '603': [ # Saudi Arabia + ('00', '04'), + ('500', '799'), + ('8000', '8999'), + ('90000', '99999'), + ], + '604': [ # Vietnam + ('0', '4'), + ('50', '89'), + ('900', '979'), + ('9800', '9999'), + ], + '605': [ # Turkey + ('00', '09'), + ('100', '399'), + ('4000', '5999'), + ('60000', '89999'), + ], + '7': [ # China, People's Republic + ('00', '09'), + ('100', '499'), + ('5000', '7999'), + ('80000', '89999'), + ('900000', '999999'), + ], + '80': [ # Czech Republic; Slovakia + ('00', '19'), + ('200', '699'), + ('7000', '8499'), + ('85000', '89999'), + ('900000', '999999'), + ], + '81': [ # India + ('00', '19'), + ('200', '699'), + ('7000', '8499'), + ('85000', '89999'), + ('900000', '999999'), + ], + '82': [ # Norway + ('00', '19'), + ('200', '699'), + ('7000', '8999'), + ('90000', '98999'), + ('990000', '999999'), + ], + '83': [ # Poland + ('00', '19'), + ('200', '599'), + ('60000', '69999'), + ('7000', '8499'), + ('85000', '89999'), + ('900000', '999999'), + ], + '84': [ # Spain + ('00', '19'), + ('200', '699'), + ('7000', '8499'), + ('85000', '89999'), + ('9000', '9199'), + ('920000', '923999'), + ('92400', '92999'), + ('930000', '949999'), + ('95000', '96999'), + ('9700', '9999'), + ], + '85': [ # Brazil + ('00', '19'), + ('200', '599'), + ('60000', '69999'), + ('7000', '8499'), + ('85000', '89999'), + ('900000', '979999'), + ('98000', '99999'), + ], + '86': [ # Serbia and Montenegro + ('00', '29'), + ('300', '599'), + ('6000', '7999'), + ('80000', '89999'), + ('900000', '999999'), + ], + '87': [ # Denmark + ('00', '29'), + ('400', '649'), + ('7000', '7999'), + ('85000', '94999'), + ('970000', '999999'), + ], + '88': [ # Italian speaking area + ('00', '19'), + ('200', '599'), + ('6000', '8499'), + ('85000', '89999'), + ('900000', '949999'), + ('95000', '99999'), + ], + '89': [ # Korea + ('00', '24'), + ('250', '549'), + ('5500', '8499'), + ('85000', '94999'), + ('950000', '999999'), + ], + '90': [ # Netherlands, Belgium (Flemish) + ('00', '19'), + ('200', '499'), + ('5000', '6999'), + ('70000', '79999'), + ('800000', '849999'), + ('8500', '8999'), + ('900000', '909999'), + ('940000', '949999'), + ], + '91': [ # Sweden + ('0', '1'), + ('20', '49'), + ('500', '649'), + ('7000', '7999'), + ('85000', '94999'), + ('970000', '999999'), + ], + '92': [ # International Publishers (Unesco, EU), European Community Organizations + ('0', '5'), + ('60', '79'), + ('800', '899'), + ('9000', '9499'), + ('95000', '98999'), + ('990000', '999999'), + ], + '93': [ # India - no ranges fixed yet + ], + '950': [ # Argentina + ('00', '49'), + ('500', '899'), + ('9000', '9899'), + ('99000', '99999'), + ], + '951': [ # Finland + ('0', '1'), + ('20', '54'), + ('550', '889'), + ('8900', '9499'), + ('95000', '99999'), + ], + '952': [ # Finland + ('00', '19'), + ('200', '499'), + ('5000', '5999'), + ('60', '65'), + ('6600', '6699'), + ('67000', '69999'), + ('7000', '7999'), + ('80', '94'), + ('9500', '9899'), + ('99000', '99999'), + ], + '953': [ # Croatia + ('0', '0'), + ('10', '14'), + ('150', '549'), + ('55000', '59999'), + ('6000', '9499'), + ('95000', '99999'), + ], + '954': [ # Bulgaria + ('00', '29'), + ('300', '799'), + ('8000', '8999'), + ('90000', '92999'), + ('9300', '9999'), + ], + '955': [ # Sri Lanka + ('0', '0'), + ('1000', '1999'), + ('20', '54'), + ('550', '799'), + ('8000', '9499'), + ('95000', '99999'), + ], + '956': [ # Chile + ('00', '19'), + ('200', '699'), + ('7000', '9999'), + ], + '957': [ # Taiwan, China + ('00', '02'), + ('0300', '0499'), + ('05', '19'), + ('2000', '2099'), + ('21', '27'), + ('28000', '30999'), + ('31', '43'), + ('440', '819'), + ('8200', '9699'), + ('97000', '99999'), + ], + '958': [ # Colombia + ('00', '59'), + ('600', '799'), + ('8000', '9499'), + ('95000', '99999'), + ], + '959': [ # Cuba + ('00', '19'), + ('200', '699'), + ('7000', '8499'), + ], + '960': [ # Greece + ('00', '19'), + ('200', '659'), + ('6600', '6899'), + ('690', '699'), + ('7000', '8499'), + ('85000', '99999'), + ], + '961': [ # Slovenia + ('00', '19'), + ('200', '599'), + ('6000', '8999'), + ('90000', '94999'), + ], + '962': [ # Hong Kong + ('00', '19'), + ('200', '699'), + ('7000', '8499'), + ('85000', '86999'), + ('8700', '8999'), + ('900', '999'), + ], + '963': [ # Hungary + ('00', '19'), + ('200', '699'), + ('7000', '8499'), + ('85000', '89999'), + ('9000', '9999'), + ], + '964': [ # Iran + ('00', '14'), + ('150', '249'), + ('2500', '2999'), + ('300', '549'), + ('5500', '8999'), + ('90000', '96999'), + ('970', '989'), + ('9900', '9999'), + ], + '965': [ # Israel + ('00', '19'), + ('200', '599'), + ('7000', '7999'), + ('90000', '99999'), + ], + '966': [ # Ukraine + ('00', '19'), + ('2000', '2999'), + ('300', '699'), + ('7000', '8999'), + ('90000', '99999'), + ], + '967': [ # Malaysia + ('00', '29'), + ('300', '499'), + ('5000', '5999'), + ('60', '89'), + ('900', '989'), + ('9900', '9989'), + ('99900', '99999'), + ], + '968': [ # Mexico + ('01', '39'), + ('400', '499'), + ('5000', '7999'), + ('800', '899'), + ('9000', '9999'), + ], + '969': [ # Pakistan + ('0', '1'), + ('20', '39'), + ('400', '799'), + ('8000', '9999'), + ], + '970': [ # Mexico + ('01', '59'), + ('600', '899'), + ('9000', '9099'), + ('91000', '96999'), + ('9700', '9999'), + ], + '971': [ #Philippines? + ('000', '019'), + ('02', '02'), + ('0300', '0599'), + ('06', '09'), + ('10', '49'), + ('500', '849'), + ('8500', '9099'), + ('91000', '99999'), + ], + '972': [ # Portugal + ('0', '1'), + ('20', '54'), + ('550', '799'), + ('8000', '9499'), + ('95000', '99999'), + ], + '973': [ # Romania + ('0', '0'), + ('100', '169'), + ('1700', '1999'), + ('20', '54'), + ('550', '759'), + ('7600', '8499'), + ('85000', '88999'), + ('8900', '9499'), + ('95000', '99999'), + ], + '974': [ # Thailand + ('00', '19'), + ('200', '699'), + ('7000', '8499'), + ('85000', '89999'), + ('90000', '94999'), + ('9500', '9999'), + ], + '975': [ # Turkey + ('00000', '00999'), + ('01', '24'), + ('250', '599'), + ('6000', '9199'), + ('92000', '98999'), + ('990', '999'), + ], + '976': [ # Caribbean Community + ('0', '3'), + ('40', '59'), + ('600', '799'), + ('8000', '9499'), + ('95000', '99999'), + ], + '977': [ # Egypr + ('00', '19'), + ('200', '499'), + ('5000', '6999'), + ('700', '999'), + ], + '978': [ # Nigeria + ('000', '199'), + ('2000', '2999'), + ('30000', '79999'), + ('8000', '8999'), + ('900', '999'), + ], + '979': [ # Indonesia + ('000', '099'), + ('1000', '1499'), + ('15000', '19999'), + ('20', '29'), + ('3000', '3999'), + ('400', '799'), + ('8000', '9499'), + ('95000', '99999'), + ], + '980': [ # Venezuela + ('00', '19'), + ('200', '599'), + ('6000', '9999'), + ], + '981': [ # Singapore + ('00', '19'), + ('200', '299'), + ('3000', '9999'), + ], + '982': [ # South Pacific + ('00', '09'), + ('100', '699'), + ('70', '89'), + ('9000', '9999'), + ], + '983': [ # Malaysia + ('00', '01'), + ('020', '199'), + ('2000', '3999'), + ('40000', '44999'), + ('45', '49'), + ('50', '79'), + ('800', '899'), + ('9000', '9899'), + ('99000', '99999'), + ], + '984': [ # Bangladesh + ('00', '39'), + ('400', '799'), + ('8000', '8999'), + ('90000', '99999'), + ], + '985': [ # Belarus + ('00', '39'), + ('400', '599'), + ('6000', '8999'), + ('90000', '99999'), + ], + '986': [ # Taiwan, China + ('00', '11'), + ('120', '559'), + ('5600', '7999'), + ('80000', '99999'), + ], + '987': [ # Argentina + ('00', '09'), + ('1000', '1999'), + ('20000', '29999'), + ('30', '49'), + ('500', '899'), + ('9000', '9499'), + ('95000', '99999'), + ], + '988': [ # Hongkong + ('00', '16'), + ('17000', '19999'), + ('200', '799'), + ('8000', '9699'), + ('97000', '99999'), + ], + '989': [ # Portugal + ('0', '1'), + ('20', '54'), + ('550', '799'), + ('8000', '9499'), + ('95000', '99999'), + ], + '9937': [ # Nepal + ('0', '2'), + ('30', '49'), + ('500', '799'), + ('8000', '9999'), + ], + '9938': [ # Tunisia + ('00', '79'), + ('800', '949'), + ('9500', '9999'), + ], + '9939': [ # Armenia + ('0', '4'), + ('50', '79'), + ('800', '899'), + ('9000', '9999'), + ], + '9940': [ # Montenegro + ('0', '1'), + ('20', '49'), + ('500', '899'), + ('9000', '9999'), + ], + '9941': [ # Georgia + ('0', '0'), + ('10', '39'), + ('400', '899'), + ('9000', '9999'), + ], + '9942': [ # Ecuador + ('00', '89'), + ('900', '994'), + ('9950', '9999'), + ], + '9943': [ # Uzbekistan + ('00', '29'), + ('300', '399'), + ('4000', '9999'), + ], + '9944': [ # Turkey + ('0', '2'), + ('300', '499'), + ('5000', '5999'), + ('60', '89'), + ('900', '999'), + ], + '9945': [ # Dominican Republic + ('00', '00'), + ('010', '079'), + ('08', '39'), + ('400', '569'), + ('57', '57'), + ('580', '849'), + ('8500', '9999'), + ], + '9946': [ # Korea, P.D.R. + ('0', '1'), + ('20', '39'), + ('400', '899'), + ('9000', '9999'), + ], + '9947': [ # Algeria + ('0', '1'), + ('20', '79'), + ('800', '999'), + ], + '9948': [ # United Arab Emirates + ('00', '39'), + ('400', '849'), + ('8500', '9999'), + ], + '9949': [ # Estonia + ('0', '0'), + ('10', '39'), + ('400', '899'), + ('9000', '9999'), + ], + '9950': [ # Palestine + ('00', '29'), + ('300', '840'), + ('8500', '9999'), + ], + '9951': [ # Kosova + ('00', '39'), + ('400', '849'), + ('8500', '9999'), + ], + '9952': [ # Azerbaijan + ('0', '1'), + ('20', '39'), + ('400', '799'), + ('8000', '9999'), + ], + '9953': [ # Lebanon + ('0', '0'), + ('10', '39'), + ('400', '599'), + ('60', '89'), + ('9000', '9999'), + ], + '9954': [ # Morocco + ('0', '1'), + ('20', '39'), + ('400', '799'), + ('8000', '9999'), + ], + '9955': [ # Lithuania + ('00', '39'), + ('400', '929'), + ('9300', '9999'), + ], + '9956': [ # Cameroon + ('0', '0'), + ('10', '39'), + ('400', '899'), + ('9000', '9999'), + ], + '9957': [ # Jordan + ('00', '39'), + ('400', '699'), + ('70', '84'), + ('8500', '9999'), + ], + '9958': [ # Bosnia and Herzegovina + ('0', '0'), + ('10', '49'), + ('500', '899'), + ('9000', '9999'), + ], + '9959': [ # Libya + ('0', '1'), + ('20', '79'), + ('800', '949'), + ('9500', '9999'), + ], + '9960': [ # Saudi Arabia + ('00', '59'), + ('600', '899'), + ('9000', '9999'), + ], + '9961': [ # Algeria + ('0', '2'), + ('30', '69'), + ('700', '949'), + ('9500', '9999'), + ], + '9962': [ # Panama + ('00', '54'), + ('5500', '5599'), + ('56', '59'), + ('600', '849'), + ('8500', '9999'), + ], + '9963': [ # Cyprus + ('0', '2'), + ('30', '54'), + ('550', '749'), + ('7500', '9999'), + ], + '9964': [ # Ghana + ('0', '6'), + ('70', '94'), + ('950', '999'), + ], + '9965': [ # Kazakhstan + ('00', '39'), + ('400', '899'), + ('9000', '9999'), + ], + '9966': [ # Kenya + ('00', '69'), + ('7000', '7499'), + ('750', '959'), + ('9600', '9999'), + ], + '9967': [ # Kyrgyzstan + ('00', '39'), + ('400', '899'), + ('9000', '9999'), + ], + '9968': [ # Costa Rica + ('00', '49'), + ('500', '939'), + ('9400', '9999'), + ], + '9970': [ # Uganda + ('00', '39'), + ('400', '899'), + ('9000', '9999'), + ], + '9971': [ # Singapore + ('0', '5'), + ('60', '89'), + ('900', '989'), + ('9900', '9999'), + ], + '9972': [ # Peru + ('00', '09'), + ('1', '1'), + ('200', '249'), + ('2500', '2999'), + ('30', '59'), + ('600', '899'), + ('9000', '9999'), + ], + '9973': [ # Tunisia + ('0', '05'), + ('060', '089'), + ('0900', '0999'), + ('10', '69'), + ('700', '969'), + ('9700', '9999'), + ], + '9974': [ # Uruguay + ('0', '2'), + ('30', '54'), + ('550', '749'), + ('7500', '9499'), + ('95', '99'), + ], + '9975': [ # Moldova + ('0', '0'), + ('100', '399'), + ('4000', '4499'), + ('45', '89'), + ('900', '949'), + ('9500', '9999'), + ], + '9976': [ # Tanzania + ('0', '5'), + ('60', '89'), + ('900', '989'), + ('9990', '9999'), + ], + '9977': [ # Costa Rica + ('00', '89'), + ('900', '989'), + ('9900', '9999'), + ], + '9978': [ # Ecuador + ('00', '29'), + ('300', '399'), + ('40', '94'), + ('950', '989'), + ('9900', '9999'), + ], + '9979': [ # Iceland + ('0', '4'), + ('50', '64'), + ('650', '659'), + ('66', '75'), + ('760', '899'), + ('9000', '9999'), + ], + '9980': [ # Papua New Guinea + ('0', '3'), + ('40', '89'), + ('900', '989'), + ('9900', '9999'), + ], + '9981': [ # Morocco + ('00', '09'), + ('100', '159'), + ('1600', '1999'), + ('20', '79'), + ('800', '949'), + ('9500', '9999'), + ], + '9982': [ # Zambia + ('00', '79'), + ('800', '989'), + ('9900', '9999'), + ], + '9983': [ # Gambia + ('80', '94'), + ('950', '989'), + ('9900', '9999'), + ], + '9984': [ # Latvia + ('00', '49'), + ('500', '899'), + ('9000', '9999'), + ], + '9985': [ # Estonia + ('0', '4'), + ('50', '79'), + ('800', '899'), + ('9000', '9999'), + ], + '9986': [ # Lithuania + ('00', '39'), + ('400', '899'), + ('9000', '9399'), + ('940', '969'), + ('97', '99'), + ], + '9987': [ # Tanzania + ('00', '39'), + ('400', '879'), + ('8800', '9999'), + ], + '9988': [ # Ghana + ('0', '2'), + ('30', '54'), + ('550', '749'), + ('7500', '9999'), + ], + '9989': [ # Macedonia + ('0', '0'), + ('100', '199'), + ('2000', '2999'), + ('30', '59'), + ('600', '949'), + ('9500', '9999'), + ], + '99901': [ # Bahrain + ('00', '49'), + ('500', '799'), + ('80', '99'), + ], + '99902': [ # Gabon - no ranges fixed yet + ], + '99903': [ # Mauritius + ('0', '1'), + ('20', '89'), + ('900', '999'), + ], + '99904': [ # Netherlands Antilles; Aruba, Neth. Ant + ('0', '5'), + ('60', '89'), + ('900', '999'), + ], + '99905': [ # Bolivia + ('0', '3'), + ('40', '79'), + ('800', '999'), + ], + '99906': [ # Kuwait + ('0', '2'), + ('30', '59'), + ('600', '699'), + ('70', '89'), + ('9', '9'), + ], + '99908': [ # Malawi + ('0', '0'), + ('10', '89'), + ('900', '999'), + ], + '99909': [ # Malta + ('0', '3'), + ('40', '94'), + ('950', '999'), + ], + '99910': [ # Sierra Leone + ('0', '2'), + ('30', '89'), + ('900', '999'), + ], + '99911': [ # Lesotho + ('00', '59'), + ('600', '999'), + ], + '99912': [ # Botswana + ('0', '3'), + ('400', '599'), + ('60', '89'), + ('900', '999'), + ], + '99913': [ # Andorra + ('0', '2'), + ('30', '35'), + ('600', '604'), + ], + '99914': [ # Suriname + ('0', '4'), + ('50', '89'), + ('900', '949'), + ], + '99915': [ # Maldives + ('0', '4'), + ('50', '79'), + ('800', '999'), + ], + '99916': [ # Namibia + ('0', '2'), + ('30', '69'), + ('700', '999'), + ], + '99917': [ # Brunei Darussalam + ('0', '2'), + ('30', '89'), + ('900', '999'), + ], + '99918': [ # Faroe Islands + ('0', '3'), + ('40', '79'), + ('800', '999'), + ], + '99919': [ # Benin + ('0', '2'), + ('40', '69'), + ('900', '999'), + ], + '99920': [ # Andorra + ('0', '4'), + ('50', '89'), + ('900', '999'), + ], + '99921': [ # Qatar + ('0', '1'), + ('20', '69'), + ('700', '799'), + ('8', '8'), + ('90', '99'), + ], + '99922': [ # Guatemala + ('0', '3'), + ('40', '69'), + ('700', '999'), + ], + '99923': [ # El Salvador + ('0', '1'), + ('20', '79'), + ('800', '999'), + ], + '99924': [ # Nicaragua + ('0', '2'), + ('30', '79'), + ('800', '999'), + ], + '99925': [ # Paraguay + ('0', '3'), + ('40', '79'), + ('800', '999'), + ], + '99926': [ # Honduras + ('0', '0'), + ('10', '59'), + ('600', '999'), + ], + '99927': [ # Albania + ('0', '2'), + ('30', '59'), + ('600', '999'), + ], + '99928': [ # Georgia + ('0', '0'), + ('10', '79'), + ('800', '999'), + ], + '99929': [ # Mongolia + ('0', '4'), + ('50', '79'), + ('800', '999'), + ], + '99930': [ # Armenia + ('0', '4'), + ('50', '79'), + ('800', '999'), + ], + '99931': [ # Seychelles + ('0', '4'), + ('50', '79'), + ('800', '999'), + ], + '99932': [ # Malta + ('0', '0'), + ('10', '59'), + ('600', '699'), + ('7', '7'), + ('80', '99'), + ], + '99933': [ # Nepal + ('0', '2'), + ('30', '59'), + ('600', '999'), + ], + '99934': [ # Dominican Republic + ('0', '1'), + ('20', '79'), + ('800', '999'), + ], + '99935': [ # Haiti + ('0', '2'), + ('7', '8'), + ('30', '59'), + ('600', '699'), + ('90', '99'), + ], + '99936': [ # Bhutan + ('0', '0'), + ('10', '59'), + ('600', '999'), + ], + '99937': [ # Macau + ('0', '1'), + ('20', '59'), + ('600', '999'), + ], + '99938': [ # Srpska + ('0', '1'), + ('20', '59'), + ('600', '899'), + ('90', '99'), + ], + '99939': [ # Guatemala + ('0', '5'), + ('60', '89'), + ('900', '999'), + ], + '99940': [ # Georgia + ('0', '0'), + ('10', '69'), + ('700', '999'), + ], + '99941': [ # Armenia + ('0', '2'), + ('30', '79'), + ('800', '999'), + ], + '99942': [ # Sudan + ('0', '4'), + ('50', '79'), + ('800', '999'), + ], + '99943': [ # Alsbania + ('0', '2'), + ('30', '59'), + ('600', '999'), + ], + '99944': [ # Ethiopia + ('0', '4'), + ('50', '79'), + ('800', '999'), + ], + '99945': [ # Namibia + ('0', '5'), + ('60', '89'), + ('900', '999'), + ], + '99946': [ # Nepal + ('0', '2'), + ('30', '59'), + ('600', '999'), + ], + '99947': [ # Tajikistan + ('0', '2'), + ('30', '69'), + ('700', '999'), + ], + '99948': [ # Eritrea + ('0', '4'), + ('50', '79'), + ('800', '999'), + ], + '99949': [ # Mauritius + ('0', '1'), + ('20', '89'), + ('900', '999'), + ], + '99950': [ # Cambodia + ('0', '4'), + ('50', '79'), + ('800', '999'), + ], + '99951': [ # Congo - no ranges fixed yet + ], + '99952': [ # Mali + ('0', '4'), + ('50', '79'), + ('800', '999'), + ], + '99953': [ # Paraguay + ('0', '2'), + ('30', '79'), + ('800', '999'), + ], + '99954': [ # Bolivia + ('0', '2'), + ('30', '69'), + ('700', '999'), + ], + '99955': [ # Srpska + ('0', '1'), + ('20', '59'), + ('600', '899'), + ('90', '99'), + ], + '99956': [ # Albania + ('00', '59'), + ('600', '999'), + ], +} + class IsbnBot: def __init__(self, generator): self.generator = generator @@ -65,64 +1175,65 @@ except wikipedia.LockedPage: print "Page %s is locked?!" % page.aslink() -def mainBackup(): - #page generator - gen = None - # If the user chooses to work on a single page, this temporary array is - # used to read the words from the page title. The words will later be - # joined with spaces to retrieve the full title. - pageTitle = [] - for arg in wikipedia.handleArgs(): - if arg.startswith('-start:'): - page = wikipedia.Page(wikipedia.getSite(), arg[7:]) - gen = pagegenerators.AllpagesPageGenerator(page.titleWithoutNamespace(), namespace = page.namespace()) - elif arg.startswith('-ref:'): - referredPage = wikipedia.Page(wikipedia.getSite(), arg[5:]) - gen = pagegenerators.ReferringPageGenerator(referredPage) - elif arg.startswith('-links:'): - linkingPage = wikipedia.Page(wikipedia.getSite(), arg[7:]) - gen = pagegenerators.LinkedPageGenerator(linkingPage) - elif arg.startswith('-file:'): - gen = pagegenerators.TextfilePageGenerator(arg[6:]) - elif arg.startswith('-cat:'): - cat = catlib.Category(wikipedia.getSite(), arg[5:]) - gen = pagegenerators.CategorizedPageGenerator(cat) - # TODO: add -xml: option - else: - pageTitle.append(arg) - if pageTitle: - # work on a single page - page = wikipedia.Page(wikipedia.getSite(), ' '.join(pageTitle)) - gen = iter([page]) - if not gen: - wikipedia.showHelp('isbn') - else: - preloadingGen = pagegenerators.PreloadingGenerator(gen) - bot = IsbnBot(preloadingGen) - bot.run() - class InvalidIsbnException(wikipedia.Error): """Invalid ISBN""" + def __init__(self, message): + self.message = message -class ISBN13: - def __init__(self, isbn10): +class ISBN: + """ + Abstract superclass + """ + + def format(self): """ - Creates a 13-digit ISBN from a 10-digit ISBN by prefixing the GS1 - prefix '978' and recalculating the checksum. - The hyphenation structure is taken from the format of the original - ISBN number. - TODO: Find out if there is a feasible way to fix hyphenation mistakes. - Seems to be difficult, as each country/region can set its own - hyphenation rules. However, the converter at isbn.org knows where to - place hyphens. + Puts hyphens into this ISBN number. """ - self.isbn10 = isbn10 - self.isbn10.checkValidity() - self.code = '978-' + self.isbn10.code[:-1] - cs = self.calculateChecksum() - self.code += str(cs) - + result = '' + rest = '' + for digit in self.digits(): + rest += str(digit) + # Determine the prefix (if any) + for prefix in self.possiblePrefixes(): + if rest.startswith(prefix): + result += prefix + '-' + rest = rest[len(prefix):] + + # Determine the group + for groupNumber in ranges.iterkeys(): + if rest.startswith(groupNumber): + result += groupNumber + '-' + rest = rest[len(groupNumber):] + publisherRanges = ranges[groupNumber] + break + else: + raise InvalidIsbnException('ISBN %s: group number unknown.' % self.code) + + # Determine the publisher + for (start, end) in publisherRanges: + length = len(start) # NOTE: start and end always have equal length + if rest[:length] > start and rest[:length] <= end: + result += rest[:length] + '-' + rest = rest[length:] + break + else: + raise InvalidIsbnException('ISBN %s: publisher number unknown.' % self.code) + + # The rest is the item number and the 1-digit checksum. + result += rest[:-1] + '-' + rest[-1] + self.code = result + +class ISBN13(ISBN): + def __init__(self, code, checksumMissing = False): + self.code = code + if checksumMissing: + self.code += str(self.calculateChecksum()) + self.checkValidity() + + def possiblePrefixes(self): + return ['978', '979'] + def digits(self): """ Returns a list of the digits in the ISBN code. @@ -132,24 +1243,32 @@ if c.isdigit(): result.append(int(c)) elif c != '-': - raise InvalidIsbnException('The ISBN contains invalid characters.') + raise InvalidIsbnException('The ISBN %s contains invalid characters.' % self.code) return result - + + def checkValidity(self): + if len(self.digits()) != 13: + raise InvalidIsbnException('The ISBN %s is not 13 digits long.' % self.code) + if self.calculateChecksum() != self.digits()[-1]: + raise InvalidIsbnException('The ISBN checksum of %s is incorrect.' % self.code) + def calculateChecksum(self): # See http://en.wikipedia.org/wiki/ISBN#Check_digit_in_ISBN_13 sum = 0 - for i in range(0, len(self.digits()), 2): + for i in range(0, 13 - 1, 2): sum += self.digits()[i] - for i in range(1, len(self.digits()), 2): + for i in range(1, 13 - 1, 2): sum += 3 * self.digits()[i] - return 10 - (sum % 10) - - -class ISBN10: + return (10 - (sum % 10)) % 10 + +class ISBN10(ISBN): def __init__(self, code): self.code = code self.checkValidity() + def possiblePrefixes(self): + return [] + def digits(self): """ Returns a list of the digits and Xs in the ISBN code. @@ -159,7 +1278,7 @@ if c.isdigit() or c == 'X': result.append(c) elif c != '-': - raise InvalidIsbnException('The ISBN contains invalid characters.') + raise InvalidIsbnException('The ISBN %s contains invalid characters.' % self.code) return result def checkChecksum(self): @@ -177,24 +1296,192 @@ lastDigit = self.digits()[-1] #print lastDigit if not ((checksum == 10 and lastDigit == 'X') or (lastDigit.isdigit() and checksum == int(lastDigit))): - raise InvalidIsbnException('The checksum is incorrect.') + raise InvalidIsbnException('The ISBN checksum of %s is incorrect.' % self.code) def checkValidity(self): if len(self.digits()) != 10: - raise InvalidIsbnException('The ISBN is not 10 digits long.') + raise InvalidIsbnException('The ISBN %s is not 10 digits long.' % self.code) if 'X' in self.digits()[:-1]: - raise InvalidIsbnException('X is only allowed at the end of the ISBN.') + raise InvalidIsbnException('ISBN %s: X is only allowed at the end of the ISBN.' % self.code) self.checkChecksum() + def toISBN13(self): + """ + Creates a 13-digit ISBN from this 10-digit ISBN by prefixing the GS1 + prefix '978' and recalculating the checksum. + The hyphenation structure is taken from the format of the original + ISBN number. + """ + code = '978-' + self.code[:-1] + + #cs = self.calculateChecksum() + #code += str(cs) + return ISBN13(code, checksumMissing = True) + +def getIsbn(code): + try: + i = ISBN13(code) + except InvalidIsbnException, e13: + try: + i = ISBN10(code) + except InvalidIsbnException, e10: + raise InvalidIsbnException(u'ISBN-13: %s / ISBN-10: %s' % (e13.message, e10.message)) + return i + +def _hyphenateIsbnNumber(match): + """ + Helper function to deal with a single ISBN + """ + code = match.group('code') + try: + i = getIsbn(code) + except InvalidIsbnException: + # don't change + return code + i.format() + return i.code + +def hyphenateIsbnNumbers(text): + isbnR = re.compile(r'(?<=ISBN )(?P<code>[\d\-]+X?)') + text = isbnR.sub(_hyphenateIsbnNumber, text) + return text + +def _isbn10toIsbn13(match): + """ + Helper function to deal with a single ISBN + """ + code = match.group('code') + try: + i = getIsbn(code) + except InvalidIsbnException: + # don't change + return code + i.format() + return i.code + +def convertIsbn10toIsbn13(text): + isbnR = re.compile(r'(?<=ISBN )(?P<code>[\d\-]+X?)') + text = isbnR.sub(_isbn10toIsbn13, text) + return text + +class IsbnBot: + + def __init__(self, generator, to13 = False, format = False, always = False): + self.generator = generator + self.to13 = to13 + self.format = format + self.always = always + self.isbnR = re.compile(r'(?<=ISBN )(?P<code>[\d\-]+X?)') + + def treat(self, page): + try: + oldText = page.get() + for match in self.isbnR.finditer(oldText): + code = match.group('code') + try: + getIsbn(code) + except InvalidIsbnException, e: + wikipedia.output(e.message) + + newText = oldText + if self.to13: + newText = self.isbnR.sub(_isbn10toIsbn13, newText) + if self.format: + newText = self.isbnR.sub(_hyphenateIsbnNumber, newText) + self.save(page, newText) + except wikipedia.NoPage: + wikipedia.output(u"Page %s does not exist?!" % page.aslink()) + except wikipedia.IsRedirectPage: + wikipedia.output(u"Page %s is a redirect; skipping." % page.aslink()) + except wikipedia.LockedPage: + wikipedia.output(u"Page %s is locked?!" % page.aslink()) + + def save(self, page, text): + if text != page.get(): + # Show the title of the page we're working on. + # Highlight the title in purple. + wikipedia.output(u"\n\n>>> \03{lightpurple}%s\03{default} <<<" % page.title()) + wikipedia.showDiff(page.get(), text) + if not self.always: + choice = wikipedia.inputChoice(u'Do you want to accept these changes?', ['Yes', 'No', 'Always yes'], ['y', 'N', 'a'], 'N') + if choice == 'n': + return + elif choice == 'a': + self.always = True + + if self.always: + try: + page.put(text) + except wikipedia.EditConflict: + wikipedia.output(u'Skipping %s because of edit conflict' % (page.title(),)) + except wikipedia.SpamfilterError, e: + wikipedia.output(u'Cannot change %s because of blacklist entry %s' % (page.title(), e.url)) + except wikipedia.LockedPage: + wikipedia.output(u'Skipping %s (locked page)' % (page.title(),)) + else: + # Save the page in the background. No need to catch exceptions. + page.put_async(text) + + + def run(self): + comment = wikipedia.translate(wikipedia.getSite(), msg) + wikipedia.setAction(comment) + + for page in self.generator: + self.treat(page) + + def main(): - # test code - isbn10 = ISBN10(u'3-86640-001-2') - isbn13 = ISBN13(isbn10) - print isbn13.code + #page generator + gen = None + # This temporary array is used to read the page title if one single + # page to work on is specified by the arguments. + pageTitle = [] + # Which namespaces should be processed? + # default to [] which means all namespaces will be processed + namespaces = [] + # This factory is responsible for processing command line arguments + # that are also used by other scripts and that determine on which pages + # to work on. + genFactory = pagegenerators.GeneratorFactory() + # Never ask before changing a page + always = False + to13 = False + format = False + for arg in wikipedia.handleArgs(): + if arg.startswith('-namespace:'): + try: + namespaces.append(int(arg[11:])) + except ValueError: + namespaces.append(arg[11:]) + elif arg == '-always': + always = True + elif arg == '-to13': + to13 = True + elif arg == '-format': + format = True + else: + generator = genFactory.handleArg(arg) + if generator: + gen = generator + else: + pageTitle.append(arg) + + if pageTitle: + page = wikipedia.Page(wikipedia.getSite(), ' '.join(pageTitle)) + gen = iter([page]) + if not gen: + wikipedia.showHelp('isbn') + else: + if namespaces != []: + gen = pagegenerators.NamespaceFilterPageGenerator(gen, namespaces) + preloadingGen = pagegenerators.PreloadingGenerator(gen) + bot = IsbnBot(preloadingGen, to13 = to13, format = format, always = always) + bot.run() + if __name__ == "__main__": try: main() finally: wikipedia.stopme() -

1 0

[Pywikipedia-l] SVN: [4259] trunk/pywikipedia/movepages.py
by leogregianin＠svn.wikimedia.org 12 Sep '07

12 Sep '07

Revision: 4259 Author: leogregianin Date: 2007-09-12 19:54:07 +0000 (Wed, 12 Sep 2007) Log Message: ----------- pt message delete Modified Paths: -------------- trunk/pywikipedia/movepages.py Modified: trunk/pywikipedia/movepages.py =================================================================== --- trunk/pywikipedia/movepages.py 2007-09-12 13:27:56 UTC (rev 4258) +++ trunk/pywikipedia/movepages.py 2007-09-12 19:54:07 UTC (rev 4259) @@ -53,16 +53,14 @@ deletesummary={ 'de': u'Bot: Lösche nach Seitenverschiebung nicht benötigte Umleitung', - 'en': u'Robot: Deleting redirect after page has been moved' + 'en': u'Robot: Deleting redirect after page has been moved', + 'pt': u'Bot: Página apagada depois de movida', # These are too unspecific: #'ar': u'حذف الصفحة بواسطة البوت', - #'en': u'Robot: deleted page', - #'de': u'Bot: Seite gelöscht', #'el': u'Διαγραφή σελίδων με bot', #'fr': u'Page supprimée par bot', #'nl': u'Pagina verwijderd door robot', #'pl': u'Usunięcie artykułu przez robota', - #'pt': u'Página apagada por bot', } class MovePagesBot: @@ -229,4 +227,4 @@ try: main() finally: - wikipedia.stopme() \ No newline at end of file + wikipedia.stopme()

1 0

[Pywikipedia-l] [ pywikipediabot-Bugs-1788099 ] featured.py inserting {{Link FA}} before categories
by SourceForge.net 12 Sep '07

12 Sep '07

Bugs item #1788099, was opened at 2007-09-04 18:40 Message generated for change (Settings changed) made by leogregianin You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=603138&aid=1788099&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: other Group: None >Status: Closed >Resolution: Invalid Priority: 6 Private: No Submitted By: André Malafaya Baptista (malafaya) >Assigned to: Leonardo Gregianin (leogregianin) Summary: featured.py inserting {{Link FA}} before categories Initial Comment: According to the comment in the source code of featured.py: # insert just before interwiki , the featured template should be inserted just before the interwiki links. I got a complaint on the bot moving the featured templates to above the categories. Please check: http://ka.wikipedia.org/w/index.php?title=%E1%83%90%E1%83%A0%E1%83%A2%E1%83… (Categories are [[კატეგორია:***]] in this Wikipedia) Thanks. ---------------------------------------------------------------------- Comment By: André Malafaya Baptista (malafaya) Date: 2007-09-12 15:33 Message: Logged In: YES user_id=1037345 Originator: YES And this sorting is the same for all wikis? There is no distinction among languages? If so, please do close this issue. ---------------------------------------------------------------------- Comment By: Leonardo Gregianin (leogregianin) Date: 2007-09-10 12:46 Message: Logged In: YES user_id=1136737 Originator: NO This comment only work on old featured.py. Now featured.py organize interwikis and categories same the interwiki.py ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=603138&aid=1788099&group_…

1 0

[Pywikipedia-l] [ pywikipediabot-Bugs-1788099 ] featured.py inserting {{Link FA}} before categories
by SourceForge.net 12 Sep '07

12 Sep '07

Bugs item #1788099, was opened at 2007-09-04 23:40 Message generated for change (Comment added) made by malafaya You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=603138&aid=1788099&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: other Group: None Status: Open Resolution: None Priority: 6 Private: No Submitted By: André Malafaya Baptista (malafaya) Assigned to: Nobody/Anonymous (nobody) Summary: featured.py inserting {{Link FA}} before categories Initial Comment: According to the comment in the source code of featured.py: # insert just before interwiki , the featured template should be inserted just before the interwiki links. I got a complaint on the bot moving the featured templates to above the categories. Please check: http://ka.wikipedia.org/w/index.php?title=%E1%83%90%E1%83%A0%E1%83%A2%E1%83… (Categories are [[კატეგორია:***]] in this Wikipedia) Thanks. ---------------------------------------------------------------------- >Comment By: André Malafaya Baptista (malafaya) Date: 2007-09-12 20:33 Message: Logged In: YES user_id=1037345 Originator: YES And this sorting is the same for all wikis? There is no distinction among languages? If so, please do close this issue. ---------------------------------------------------------------------- Comment By: Leonardo Gregianin (leogregianin) Date: 2007-09-10 17:46 Message: Logged In: YES user_id=1136737 Originator: NO This comment only work on old featured.py. Now featured.py organize interwikis and categories same the interwiki.py ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=603138&aid=1788099&group_…

1 0

← Newer
1
...
21
22
23
24
25
26
27
...
40
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

pywikibot September 2007