Revision: 5764 Author: valhallasw Date: 2008-07-27 11:04:03 +0000 (Sun, 27 Jul 2008)

Log Message: ----------- Added support for per-site post-get and pre-put conversions. Added automatic conversion for eowiki X-convention -> unicode

Modified Paths: -------------- trunk/pywikipedia/families/wikipedia_family.py trunk/pywikipedia/family.py trunk/pywikipedia/wikipedia.py

Modified: trunk/pywikipedia/families/wikipedia_family.py =================================================================== --- trunk/pywikipedia/families/wikipedia_family.py 2008-07-26 12:22:14 UTC (rev 5763) +++ trunk/pywikipedia/families/wikipedia_family.py 2008-07-27 11:04:03 UTC (rev 5764) @@ -1,6 +1,6 @@ # -*- coding: utf-8 -*- import urllib -import family, config +import family, config, wikipedia

__version__ = '$Id$'

@@ -954,3 +954,9 @@

def shared_image_repository(self, code): return ('commons', 'commons') + + def post_get_convert(self, site, getText): + if site.lang == 'eo': + return wikipedia.decodeEsperantoX(getText) + else: + return getText

Modified: trunk/pywikipedia/family.py =================================================================== --- trunk/pywikipedia/family.py 2008-07-26 12:22:14 UTC (rev 5763) +++ trunk/pywikipedia/family.py 2008-07-27 11:04:03 UTC (rev 5764) @@ -3183,3 +3183,13 @@ def isPublic(self): """Does the wiki require logging in before viewing it ?""" return True + + def post_get_convert(self, site, getText): + """Does a conversion on the retrieved text from the wiki + i.e. Esperanto X-conversion """ + return getText + + def pre_put_convert(self, site, putText): + """Does a conversion on the text to insert on the wiki + i.e. Esperanto X-conversion """ + return putText

Modified: trunk/pywikipedia/wikipedia.py =================================================================== --- trunk/pywikipedia/wikipedia.py 2008-07-26 12:22:14 UTC (rev 5763) +++ trunk/pywikipedia/wikipedia.py 2008-07-27 11:04:03 UTC (rev 5764) @@ -782,7 +782,13 @@ else: self._isWatched = False # Now process the contents of the textarea - m = self.site().redirectRegex().match(text[i1:i2]) + # Unescape HTML characters, strip whitespace and postconvert + pagetext = text[i1:i2] + pagetext = unescape(pagetext) + pagetext = pagetext.strip() + pagetext = self.site().post_get_convert(pagetext) + + m = self.site().redirectRegex().match(pagetext) if m: # page text matches the redirect pattern if self.section() and not "#" in m.group(1): @@ -801,13 +807,8 @@ self._getexception except AttributeError: raise SectionError # Page has no section by this name - # TODO: Docu and rewrite also this as above. - x = text[i1:i2] - x = unescape(x) - while x and x[-1] in '\n ': - x = x[:-1]

- return x + return pagetext

def getOldVersion(self, oldid, force=False, get_redirect=False, throttle=True, sysop=False, change_edit_time=True): @@ -1277,6 +1278,7 @@ import watchlist watchArticle = watchlist.isWatched(self.title(), site = self.site()) newPage = not self.exists() + newtext = self.site().pre_put_convert(newtext) return self._putPage(newtext, comment, watchArticle, minorEdit, newPage, self.site().getToken(sysop = sysop), sysop = sysop)

@@ -3954,6 +3956,11 @@ linktrail: Return regex for trailing chars displayed as part of a link. disambcategory: Category in which disambiguation pages are listed.

+ post_get_convert: Converts text data from the site immediatly after get + i.e. EsperantoX -> unicode + pre_put_convert: Converts text data from the site immediatly before put + i.e. unicode -> EsperantoX + Methods that yield Page objects derived from a wiki's Special: pages (note, some methods yield other information in a tuple along with the Pages; see method docs for details) -- @@ -5739,7 +5746,13 @@ def linktrail(self): """Return regex for trailing chars displayed as part of a link.""" return self.family.linktrail(self.lang) + + def post_get_convert(self, getText): + return self.family.post_get_convert(self, getText)

+ def pre_put_convert(self, putText): + return self.family.pre_put_convert(self, putText) + def language(self): """Return Site's language code.""" return self.lang