Manual:TikiWiki Conversion

From Linux Web Expert

These steps were originally used to convert pages from a version 1.9.3.1 TikiWiki with both wiki syntax and HTML to a version 1.10.4 Mediawiki. In 2019 the code was adapted to the export of a version 12.6 TikiWiki. The script takes a tar file from a TikiWiki export and returns a MediaWiki XML file. During this adaption the original code, which can still be found at the bottom of this page, was ported to a Git-Repository hosted on GitHub, translated to Python 3 and altered to work with the current application versions. Besides this paragraph and a note for the Python version in the Requirements section, this page was not changed since the steps of ex- and importing stayed the same.

They should work for other versions too where syntax doesn't vary too much but seeing as we only had one small wiki to convert it's not widely tested. It works for all syntax in our TikiWiki though which was a horrible mix of pure wiki syntax, pure html and everything in between.

The known unsupported syntax includes tikiwiki syntax tables

All revisions of pages will be imported with their times and contributors.

The authors of the pages will be set to the TikiWiki authors. These may or may not exist in MediaWiki and anyone who then registers using this username will have those pages attributed to them.

Files uploaded to TikiWiki will have to be uploaded to Mediawiki and their links fixed manually. The script will output a list of pages containing file downloads.

Images from the TikiWiki are included as external images which involves changing the external images settings of your MediaWiki. This can however be set to only allow images from one URL source.

Requirements

  • Admin access to your TikiWiki
  • Sysop access to a MediaWiki installation
  • Access to the TikiWiki database and the capability to perform MySQL commands
  • Python
    • the 2019 version was tested with Python 3.6.1, 3.7.2 and 3.9.3
    • the original code at the bottom of this page used 2.5.1, with this patch from this bug to ensure it's parsed properly
  • The python conversion script

Exporting

  • To export from TikiWiki go to Admin (or Admin Home) then Wiki, and then choose the Export link under Export Wiki Pages.
  • This will produce a tar achive, export.tar, of MIME (RFC2045) message (RFC2822) files.
  • You will also need to create an XML dump of the metadata for the Tiki images. Assuming your database is "tiki" and your table prefix is "tiki_" then the following query will fetch the required data:
SELECT tiki_images.galleryID, tiki_images.imageID, name, filename FROM tiki.tiki_images, tiki.tiki_images_data WHERE tiki_images.imageID=tiki_images_data.imageID
  • If you wish to exclude any pages from the export, perhaps in order to keep them private, then you will need an XML file specifying which pages to exclude. you can do this by hand or use the following query:
SELECT pageName, groupName, permName FROM tiki.users_objectpermissions, tiki.tiki_pages where objectType='wiki page' and objectId=page_id
which will select pages with any restrictions, though you could specify particular restrictions. The script will exclude any pages whose names are contained in a <field name='pageName'> element in that XML file.
  • note: use the mysql commands with the '-X' flag to produce xml output.

Running The Script

  • If you want to keep any pages private you can use the XML file created above.
  • Copy the source to a file named tikiToMwiki.py
  • Run the script :
python tikiToMwiki.py [options] <tiki source URL> [tiki export]
  • where
    • tiki source URL is the URL preceding all tikiwiki page URLs including the trailing slash. For example 'http://yourdomain.com/tiki/'
    • tiki export is the tar file produced by the TikiWiki export. If not specified then the script defaults to reading from stdin.
  • options are
    • -o output file the base name of the file(s) to be output - the script can divide the output into multiple small files so output files will be numbered. For example if output file is 'wiki.xml' the script will produce files called 'wiki0.xml','wiki1.xml' and so on. This will default to the name of tiki export with ".tar" replaced by ".xml" or to stdout if tiki export isn't given. If you specify "-o -" then output will be sent to stdout.
    • -k image file an XML file containing image metadata
    • -i tiki image URL is the (possibly relative) URL used to access images when using their ids or names. For example '/UserFiles/Image/tiki/'
    • The -n flag specifies whether to force pages not to have a table of contents. By default MediaWiki adds a table of contents to pages with three or more headings, you may decide not to use the table of contents to make pages more similar to the original TikiWiki pages.
    • -j new image URL a new URL for the image files (including trailing slash) if this is not set the old image locations will be used. You can use this option if you need to change the TikiWiki image file locations.
    • -m max maximum file size for the new files in MB. The default value is 1. MediaWiki has a default maximum import size of 20M.
    • -p private pages an XML file containing a list of any pages you wish to exclude from the import
  • if your wiki contains pictures you will need to run the script with both the -i and -k options
  • if you provide the tiki export file as a command line argument any internal links which are in the wrong case will be corrected (MediaWiki is case sensitive, TikiWiki isn't). If you use stdin to supply the tiki export then links cannot be corrected.

Importing

  • To import the MediaWiki XML file(s) use one of these import methods and upload each XML file.

Changing Settings

If anything goes wrong you can use the deleteBatch maintenance script to delete multiple pages. NB If you import pages which have previously been imported the pages will not be overwritten, you have to delete the pages and then import again.

Re-Use

© Crown copyright 2008 - Rosie Clarkson, Chris Eveleigh (development@planningportal.gov.uk) for the Planning Portal

You may re-use the Crown copyright protected material (not including the Royal Arms and other departmental or agency logos) free of charge in any format. The material must be acknowledged as Crown copyright and the source given.

Code

# -*- coding: utf-8 -*-
#	© Crown copyright 2008 - Rosie Clarkson, Chris Eveleigh (development@planningportal.gov.uk) for the Planning Portal
#
#	You may re-use the Crown copyright protected material (not including the Royal Arms and other departmental or agency logos)
#	free of charge in any format. The material must be acknowledged as Crown copyright and the source given.
#
#
# 23/03/2009 Patched by Miguel Tremblay, Environment Canada
#  Script is now supposed to be functionnal
#   with french characters in UTF-8
# 6 Dec 2012 fix from Andrew White at Typesafe
########################################################
 
import sys,os,time,tarfile
from email.Parser import Parser
from xml.dom import minidom 
import xml.parsers.expat
import xml.sax.saxutils as saxutils 
import htmlentitydefs
import re
from HTMLParser import HTMLParser
from urllib import quote,unquote
from urlparse import urljoin
from optparse import OptionParser
 
#add any other links you may want to map between wikis here
url_maps = {'http://tikiwiki.org/RFCWiki':'http://meta.wikimedia.org/wiki/Cheatsheet'}
 
#checks for HTML tags
class HTMLChecker(HTMLParser):
 
	def handle_starttag(self, tag, attrs):
		global validate
		validate = True
		return True
	def handle_endtag(self, tag):
		global validate
		return True
 
#Mediawiki relies on having the right number of new lines between syntax - for example having two new lines in a list starts a new list.
#The elements that do/don't start a new line in HTML can be controlled by the CSS. The CSS used depends on which skin you're using.
class HTMLToMwiki(HTMLParser):
	global wikitext
	global sourceurl
	global pages
	global uploads
	global headings
	link = False #if the parser is within a link
	src = ''
	innowiki = False
	inem = False #if the parser is within italics
	instrong = False #if the parser is within bold
	inheading = False #if the parser is within a heading
	list=0 #whether the parser is within an ordered list (is numeric to deal with nested lists)
	litem=0 #whether the parser is within a list item - in order to deal with <p> and <br/> tags in ways that wont break it
	ul_count=0 #the number of ul tags used for nested lists
	ol_count=0 #the number of ol tags used for nested lists
	col_count=0
	def handle_starttag(self, tag, attrs):
		if self.innowiki:
			completeTag='<'+tag
			for attr in attrs:
				completeTag+=' '+attr[0]+'="'+attr[1]+'"'
			wikitext.append(completeTag+'>')
		else:
			if tag == 'nowiki':
				wikitext.append('<nowiki>')
				self.innowiki=True
			if tag == 'a':
				self.src=''
				for att in attrs:
					if att[0] == 'href':
						self.src = att[1]
				if self.src in url_maps:
					self.src=url_maps[self.src]
				#deals with uploads
				if 'tiki-download_file.php' in self.src:
					uploads.append(self.src)
				self.link=True
			if tag == 'ol':
				self.ol_count+=1
				self.list+=1
 
			if tag == 'ul':
				self.ul_count+=1
			if tag == 'li':
				#append the right no. of # or *s according to the level of nesting
				self.litem+=1
				if self.list>0:
					wikitext.append('\n'+('#'*self.ol_count))
				else:
					wikitext.append('\n'+('*'*self.ul_count))
			if tag == 'img':
				src=''
				for att in attrs:
					if att[0] == 'src':
						src = att[1]
				src = quote(src)
				#we have several different ways of specifying image sources in our tiki
				imagepath = urljoin(sourceurl, src)
				if options.newImagepath != '':
					imagepath=urljoin(options.newImagepath, src.split('/')[-1])
				# the pic tag is used later to identify this as a picture and process the correct mwiki syntax
				wikitext.append('<pic>'+imagepath+' ')	
			if tag == 'table':
				wikitext.append('\n{|')
				for att in attrs:
					#table formatting
					wikitext.append(' '+att[0]+'="'+att[1]+'"')
			if tag == 'tr':	
				wikitext.append('\n|-')
				self.col_count=0
			if tag == 'td':	
				self.col_count+=1
				if self.col_count > 1:
					wikitext.append('\n||')			
				else:
					wikitext.append('\n|')			
			if tag == 'caption':	
				wikitext.append('\n|+')
			if tag in ('strong','b'):
				self.instrong=True
				wikitext.append("'''")
			if tag in('em','i'):
				self.inem=True
				wikitext.append("''")
			if tag =='p':
				#new lines in the middle of lists break the list so we have to use the break tag
				if self.litem==0:
					br = '\n'
				else: 
					br = '<br/>'
				#newlines in the middle of formatted text break the formatting so we have to end and restart the formatting around the new lines
				if self.inem==True:
					br = "''"+br+br+"''"
				if self.instrong==True:
					br = "'''"+br+br+"'''"
				wikitext.append(br)	
			if tag =='h1':
				self.inheading = True
				#headings must start on a new line
				wikitext.append('\n\n==')
				headings.append(tag)
			if tag =='h2':
				self.inheading = True
				wikitext.append('\n\n===')
				headings.append(tag)
			if tag =='h3':
				self.inheading = True
				wikitext.append('\n\n====')
				headings.append(tag)
			else:
				wikitext.append('<'+tag+'>')
 
	def handle_endtag(self, tag):
		if tag =='nowiki':
			wikitext.append('</nowiki>')
			self.innowiki=False
		if not self.innowiki:
			if self.link==True:
				self.src=''
				self.link=False
			if tag == 'img':
				wikitext.append('</pic>')
			if tag == 'ol':
				self.ol_count-=1
				self.list-=1
				wikitext.append('\n\n')
			if tag == 'ul':
				self.ul_count-=1
				wikitext.append('\n\n')
			if tag == 'li':
				self.litem-=1
			if tag == 'table':
				wikitext.append('\n\n|}')
			if tag in('strong','b'):
				self.instrong=False
				wikitext.append("'''")
			if tag in('em','i'):
				self.inem=False
				wikitext.append("''")
			if tag =='h1':
				self.inheading = False
				wikitext.append('==\n\n')
			if tag =='h2':
				self.inheading = False
				wikitext.append('===\n\n')
			if tag =='h3':
				self.inheading = False
				wikitext.append('====\n\n')
			if tag =='p':
				if self.inheading == True:
					br=''
				elif self.litem==0:
					br = '\n'
				else: 
					br = '<br/>'
				if self.inem==True:
					br = " ''"+br+"''"
				if self.instrong==True:
					br = " '''"+br+"'''"
				wikitext.append(br)
			if tag == 'br':
				if self.inheading == True:
					br=''
				elif self.litem==0:
					br = '\n'
				else: 
					br = '<br/>'
				if self.inem==True:
					br = " ''"+br+"''"
				if self.instrong==True:
					br = " '''"+br+"'''"
				wikitext.append(br)
			if tag == 'hr':
				wikitext.append('\n----\n')	
			else:
				wikitext.append('</'+tag+'>')
		else:
			wikitext.append('</'+tag+'>')
	#check for symbols which are mwiki syntax when at the start of a line
	def check_append(self,data):
		stripped = data.lstrip()
		for symbol in ('----','*','#','{|','==','===','===='):
			if stripped.startswith(symbol):
				if len(wikitext) > 2 and wikitext[-3] == '\n':
					if not symbol.startswith('='):
						data= '<nowiki>'+symbol+'</nowiki>'+stripped[len(symbol):]
					else:
						if data.find(symbol,len(symbol)):
							data= '<nowiki>'+symbol+'</nowiki>'+stripped[len(symbol):]
		return data
 
	def handle_data(self,data):
		if self.link==True:
			#sometimes spaces are in the piped data (probably because of our editor) so we need to make sure we add that before the link
			space = ''
			if data.startswith(' '):
				space = ' '
			if self.src.startswith(sourceurl+'tiki-download_file.php'):
				wikitext.append(space+'['+self.src+' '+data+']')
			elif self.src.startswith(sourceurl):
				if 'page=' in self.src:
					ptitle = self.src.split('page=')
					page = ptitle[1].replace('+',' ')
					for file in pages:
						#mwiki is case sensitive to page names and tikiwiki isn't so check that the file actually exists
						if file.lower()==page.lower():
							page = file
					wikitext.append(space+'[['+page+'|'+data+']]') 
			else:
				#catch relative urls
				if self.src.startswith('..'):
					self.src = urljoin(sourceurl, self.src)
				wikitext.append(space+'['+self.src+' '+data+']')
		elif self.litem:
			# if we're in a list put nowiki tags around data begining with * or # so it isnt counted as nesting
			if data[0] in ('*', '#'):
				data = '<nowiki>'+data[0]+'</nowiki>'+data[1:]
			wikitext.append(data)
		else:
			data = self.check_append(data)
			wikitext.append(data)
 
	def handle_entityref(self,data):
		data="&amp;"+data+";"
		if self.link:
			wikitext.append(' '+data)
		elif self.litem:
			wikitext.append(data)
		else:
			wikitext.append(data)
 
	def handle_charref(self,data):
		data="&amp;"+data+";"
		if self.link:
			wikitext.append(' '+data)
		elif self.litem:
			wikitext.append(data)
		else:
			wikitext.append(data)
 
def insertImage(word,words):
	global image 
	global imagenames
	global imageids
	global imagepath
	#there are even more ways to specify pic sources in our tiki
	if 'name=' in word:
		parts = word.split('=')
		try:
			filename = imagenames[parts[2]]
		except KeyError:
			sys.stderr.write(parts[2]+' doesn\'t exist in your image XML file and won\'t be displayed properly\n')
			filename=parts[2]
		filename = quote(filename)
		imagepath = urljoin(urljoin(sourceurl,imageurl), filename)
		if options.newImagepath != '':
			imagepath=urljoin(options.newImagepath, filename)
		words.append('<pic>'+imagepath)
	if 'id=' in word:
		parts = word.split('=')
		try:
			filename = imageids[parts[2]]
		except KeyError:
			sys.stderr.write( 'The image with ID '+parts[2]+' doesn\'t exist in your image XML file and won\'t be displayed properly\n')
			filename=parts[2]
		filename = quote(filename)
		imagepath = urljoin(urljoin(sourceurl,imageurl), filename)
		if options.newImagepath != '':
			imagepath=urljoin(options.newImagepath, filename)
		words.append('<pic>'+imagepath)
	if '}' in word:
		bracket=word.find('}')
		if word[-1]!='}':
			if word[bracket+1]!=' ':
				word=word.replace('}','</pic> ')
			else: 
				word=word.replace('}','</pic>')
		word=word.replace('}','</pic>')
		words.append(word)
		image = False
 
	return words
 
def insertLink(word):
	global intLink
	global page
	global words
	global pages
	first=False
	#the link may be split if it contains spaces so it may be sent in parts
	brackets=word.find('((')
	if brackets != -1:
		word = word.replace('((','[[')
		page = word[brackets:]
		words.append(word[:brackets])
		if '))' in word:
			word = word.replace('))',']]')
			end=word.find(']]')
			text = word[brackets+2:end]
			#again check the filenames to ensure case sensitivity is ok
			for file in pages:
				if unicode(file, "Latin-1").lower() \
					== text.lower():
					text = file
			text = '[['+text+word[end:]
			if text[-1]!='\n':
				words.append(text+' ')
			else:
				words.append(text)
			page = ''
			intLink=False
 
	elif '))' in word:
		word = word.replace('))',']]')
		page += ' '+word
		pipe = page.find('|')
		if pipe != -1:
			end=pipe
			text= page[2:pipe]
		else:
			brackets=page.find(']]')
			end=brackets
			text= page[2:brackets]
		for file in pages:
			if unicode(file, "latin-1").lower()==text.lower():
				page=page[:2]+file+page[end:]
		if page[-1]!='\n':
			words.append(page+' ')
		else:
			words.append(page)
		page = ''
		intLink=False
	else:
		first=False
		page += ' '+word
 
parser = OptionParser()
parser.add_option("-n", "--notableofcontents",
		  action="store_true", dest="notoc", default=False, 
		  help="disable all automatic contents tables")
parser.add_option("-m", "--maxfilesize",
                  action="store", type="int", dest="max", default=1,
                  help="the maximum import file size")
parser.add_option("-j", "--newimageurl",
                  action="store", type="string", dest="newImagepath", default='',
                  help="the new location of any images (inc. trailing slash)")
parser.add_option("-i", "--imageurl",
                  action="store", type="string", dest="imageurl", default='',
                  help="the relative URL used in tiki to access images (inc. trailing slash)")
parser.add_option("-p", "--privatepages",
                  action="store", type="string", dest="privatexml", default='',
                  help="an XML file containing any private pages not to be added to the wiki")
parser.add_option("-o", "--outputfile",
                  action="store", type="string", dest="outputFile", default='',
                  help="the name of the output wiki XML file(s)")
parser.add_option("-k", "--imagexml",
                  action="store", type="string", dest="imagexml", default='',
                  help="an XML file containing metadata for the images in the tiki")
 
(options, args) = parser.parse_args()
 
# the tar file containing the tiki file export - if not specified read from stdin
#stdin doesn't work at the moment and fails after you've used extractfile as this returns nothing
if len(args)>1:
	archive = tarfile.open(args[1])
	#add all files in the export tar to the list of pages	
	pages = archive.getnames()
	if options.outputFile=='':
		outputFile= args[1].replace('.tar','.xml')
	else:
		outputFile=options.outputFile
else:
	pages=[]
	#if reading from stdin you can't iterate through the files again so pages is left empty and links are not corrected
	archive = tarfile.open(name= sys.stdin.name, mode='r|', fileobj=sys.stdin)
	#if you're reading from stdin and don't specify an output file output to stdout
	if options.outputFile=='':
		options.outputFile='-'
p = Parser()
 
#multiple files may be created so this is added to the output file string to identify them
fileCount=0
 
#the string to name all outputfiles the fileCount is added to this
if options.outputFile=='-':
	mwikixml=sys.stdout
else:
	mwikixml = open(outputFile[:-4]+str(fileCount)+outputFile[-4:], 'wb')
	sys.stderr.write('Creating new wiki xml file '+outputFile[:-4]+str(fileCount)+outputFile[-4:])
 
#the source URL of the tiki - in the form http://[your url]/tiki/
sourceurl = args[0]
 
#the relative address used to access pictures in TikiWiki
imageurl = options.imageurl
 
privatePages=[]
if options.privatexml !='':
	privateparse = 	minidom.parse(options.privatexml)
	rows=privateparse.getElementsByTagName('row')
	for row in rows:
		fields = row.getElementsByTagName('field')
		for field in fields:
			if field.getAttribute('name')=='pageName':
				privatePages.append(field.firstChild.data)
#fill the lookup table with the image information
#a file containing an xml dump from the tiki DB
imagenames={}
imageids={}
if options.imagexml !='':
	imagexml = options.imagexml
	lookup = minidom.parse(imagexml)
 
	rows = lookup.getElementsByTagName('row')
	for row in rows:
		fields = row.getElementsByTagName('field')
		for field in fields:
			if field.getAttribute('name')=='name':
				iname = field
			if field.getAttribute('name')=='filename':
				ifile = field
			if field.getAttribute('name')=='imageID':
				iid = field
		imagenames[iname.firstChild.data] = ifile.firstChild.data
		imageids[iid.firstChild.data] = ifile.firstChild.data
 
 
 
#list of users who have edited pages	
authors = []
filepages = {}
totalSize=0
pagecount = 0
versioncount = 0
 
#write mediawiki xml file
mwikixml.write('<mediawiki xml:lang="en">\n')
 
for member in archive:
	if member.name not in privatePages:
		#add each file in the tiki export directory
		tikifile = archive.extractfile(member)
		mimefile = p.parse(tikifile)
		mwikixml.write('<page>\n')
		partcount=0
		uploads = []
	 
		if not mimefile.is_multipart():
			partcount =1
		for part in mimefile.walk():
			outputpage=''
			if partcount == 1:
				title = unquote(part.get_param('pagename'))
				outputpage += '<title>'+title+'</title>'
			partcount+=1
			if part.get_params() is not None and \
				('application/x-tikiwiki','') in part.get_params():
				versioncount +=1
				headings =[]
				if part.get_param('lastmodified') == None:
					break
				outputpage += '<revision>\n'
				outputpage += '<timestamp>'+ \
					time.strftime('%Y-%m-%dT%H:%M:%SZ',time.gmtime(eval(part.get_param('lastmodified'))))+'</timestamp>\n'
				outputpage += '<contributor><username>'+part.get_param('author')+'</username></contributor>\n'
				# add author to list of contributors to be output at the end
				if part.get_param('author')not in authors:
					authors.append(part.get_param('author'))
				outputpage += '<text xml:space="preserve">\n'
				mwiki = ''
				#we add the tiki description to the page in bold and italic (much as it was in tikiwiki)
				#for them to function properly we need to ensure that these strings are followed by a new line
				# the </br> is used as a placeholder and is converted to \n later
				if part.get_param('description') not in(None,''):
					mwiki+="'''''"+unquote(part.get_param('description'))+"'''''</br>"
				#then add the table of contents (or specify none)
				if options.notoc: mwiki=mwiki + "__NOTOC__</br>"
				else: mwiki+="__TOC__</br>"
				mwiki += part.get_payload().decode('utf-8')
	 
	 
				#does the validator do anything?!
				validate = False
				validator = HTMLChecker()
				validator.feed(mwiki)
				#fixes pages that end up on a single line (these were probably created by our WYSWYG editor being used on windows and linux)
				if not validate:
					mwiki=mwiki.replace('\t', '    ')
					mwiki=mwiki.replace('  ', ' &nbsp;')
					mwiki=mwiki.replace('<', '&lt;')
					mwiki=mwiki.replace('>', '&gt;')

					#make sure newlines after headings are preserved
					next = 0
					while '\r\n!' in mwiki[next:] or '&lt;/br&gt;!' in mwiki[next:] or mwiki[next:].startswith('!'):
						if mwiki[next:].startswith('!'): found= next
						else:
							foundreturn = mwiki.find('\r\n!', next)
							foundbreak = mwiki.find('&lt;/br&gt;!', next)
							if (foundreturn != -1 and foundreturn<foundbreak) or foundbreak==-1:
								found = foundreturn+2
							else:
								found = foundbreak+11

						next = mwiki.find('\r\n', found)
						if next == -1: break
						mwiki=mwiki[:next]+'</br>'+mwiki[next+2:]
						next += 5

					#as validate is false the page does not contain any html so whitespace needs to be preserved
					mwiki=mwiki.replace('\r\n', '</br>')

				#double escape < and > entities so that &lt; is not unescaped to < which is then treated as HTML tags
				#mwiki=mwiki.replace('&amp;', '&amp;amp;')
				mwiki=mwiki.replace('&amp;lt;', '&amp;amp;lt;')
				mwiki=mwiki.replace('&amp;gt;', '&amp;amp;gt;')
				mwiki=mwiki.replace('&lt;', '&amp;lt;')
				mwiki=mwiki.replace('&gt;', '&amp;gt;')
				mwiki=mwiki.replace(u'\ufffd', '&nbsp;') 
	 
				# unescape XML entities
				entitydefs = dict( ("&"+k+";", unichr(v)) for k, v in htmlentitydefs.name2codepoint.items() )
				entitydefs.pop("&amp;")
				entitydefs.pop("&gt;")
				entitydefs.pop("&lt;")
				mwiki=saxutils.unescape(mwiki, entitydefs)

				# replace tiki syntax that will be interpreted badly with tiki syntax the parser will understand
				#empty formatting tags will be converted to many "'"s which then confuses mwiki
				mwiki = mwiki.replace('[[','~np~[~/np~')	
				#need to replace no wiki tags here in case any html/xml is inside them that we want to keep
				mwiki=mwiki.replace('~np~', '<nowiki>')
				mwiki=mwiki.replace('~/np~', '</nowiki>')
				mwiki=mwiki.replace('<em></em>', '')
				mwiki=mwiki.replace('<em><em>', '<em>')
				mwiki=mwiki.replace('</em></em>', '</em>')
				mwiki=mwiki.replace('<strong></strong>', '')
				mwiki=mwiki.replace('<strong><strong>', '<strong>')
				mwiki=mwiki.replace('</strong></strong>', '</strong>')
				#this makes sure definitions keep their preceding newline
				mwiki=mwiki.replace('\n;', '</br>;')
				mwiki=mwiki.replace('\n', ' ')
				mwiki=mwiki.replace('</br>', '\n')
				mwiki=mwiki.replace('&lt;/br&gt;', '\n')
				mwiki=mwiki.replace('\r', ' ')
				mwiki=mwiki.replace('\t', ' ')

				# Mediawiki automatically creates a table of content
				mwiki=mwiki.replace('Table of content', '')
				mwiki=mwiki.replace('{maketoc}', '')

				# convert === underline syntax before the html converter as 
				# headings in mwiki use =s and h3 tags will become ===heading===
				next = 0
				while '===' in mwiki[next:]:
					start = mwiki.find('===', next)
					end = mwiki.find('===', start+3)

					if end != -1:
						mwiki = mwiki[:start]+'<u>'+mwiki[start+3:end]+'</u>'+mwiki[end+3:]
					next = start +1
					#if there is another === convert them both

				#print mwiki


				wikitext=[]
	 
				#convert any HTML tags to mediawiki syntax
				htmlConverter = HTMLToMwiki()
				htmlConverter.feed(mwiki)
	 
				mwiki = ''.join(wikitext)

				#replace tiki syntax with mwiki
				mwiki=mwiki.replace('__',"'''")
	 
				#split the text into lines and then strings to parse 
				words=[]
				image=False
				intLink=False
				box=False
				colour = False
				inColourTag=False
				page = ''
				centre=False
				for line in mwiki.splitlines(True):
					# Convert external links to mediawiki syntax
					m = re.match(r'(.*)\[(.*)\|(.*)\](.*)', line)
					if m:
						line = m.group(1) + "[" + re.sub(r'(.*)&amp;(.*);(.*)',r'\1&\2\3', m.group(2)) + " " + m.group(3) + "]" + m.group(4) + "\n"

					# Convert 'CODE' samples
					line = re.sub(r'(.*){CODE\(caption=&amp;gt;(.*)\)}(.*)', r'\1<!-- \2 --><pre>\3', line)
					line = re.sub(r'(.*){CODE}(.*)', r'\1</pre>\2', line)

					# Convert anchor
					line = re.sub(r'(.*){ANAME\(\)}(.*){ANAME}(.*)', r'\1<span id=&quot;\2&quot;></span>\3', line)
					# Convert anchor links
					line = re.sub(r'(.*){ALINK\(aname=(?:")?([^"]*)(?:")?\)}(.*){ALINK}(.*)', r'\1[[#\2|\3]]\4', line)

					heading = False
					noCentre = False
					#if there are an odd no. of ::s don't convert to centered text
					if line.count('::') % 2 != 0:
						noCentre =True
					count =0
					spl = line.split(' ')
					if spl[0].find('!') == 0: heading = True
					for word in spl:
						#handle headings
						if heading is True:
							if count == 0 and word:
								#replace !s
								bangs =0;
								while word[bangs]== '!':
									word=word.replace('!','=',1)
									bangs+=1
									if bangs >= len(word):
										if len(spl) == 1: bangs /= 2
										break
							if count is len(spl)-1:
								#add =s to end
								end = word.find('\n')
								if end != -1:
									word=word[:end]+(bangs*'=')+word[end:]
								else:
									word=word[:end]+(bangs*'=')
						#handle centered text
						if '::' in word and not noCentre:
							next=0
							while '::' in word[next:]:
								next = word.find('::')
								if centre:
									centre = False
									word = word.replace('::','</center>',1)
								else:
									centre = True
									word = word.replace('::','<center>',1)
						#handle font colours
						if inColourTag:
							colon = word.find(':')
							if colon != -1:
								word = word[:colon]+'">'+word[colon+1:]
								inColourTag=False
						if '~~' in word:
							next=0
							while '~~' in word[next:]:
								next = word.find('~~')
								if colour == True:
									#end span
									colour = False
									word = word.replace('~~','</span>',1)
								else:
									#start span
									colour=True
									colon = word.find(':',next)
									extratext =''
									if colon != -1:
										word = word[:next]+"<span style='color:"+word[next+2:colon]+"'>"+word[colon+1:]
									else:
										word = word[:next]+'<span style="color:'+word[next+2:]
										inColourTag=True
								next += 1
						# handle boxes
						if '^' in word:
							hats = word.count('^')
							for hat in range(1, hats+1):
								index = word.find('^')
								if not box:
									word = word[:index]+'<pre>'+word[index+1:]
									box=True
								else:
									word = word[:index]+'</pre>'+word[index+1:]
									box=False							
						if '{img' in word:
							image = True
						if '((' in word:
							intLink = True
						if image:
							words = insertImage(word,words)
						elif intLink:
							insertLink(word)
						else:
							#stops mwiki automatically creating links (which can then be broken by formatting
							if ('http' in word or 'ftp://' in word) and '[' not in word and ']' not in word and '<pic>' not in word and '<pre>' not in word and '</pre>' not in word and not box:
								index = 0
								format = False
								formatted = ''
								for char in word:
									index+=1
									if char == "'":
										if not format:
											format= True
											formatted = formatted+'</nowiki>'
									else:
										if format:
											format = False
											formatted = formatted+'<nowiki>'
	 
									formatted +=char
	 
								word = '<nowiki>'+formatted+'</nowiki>'
							if word != '':
								if '\n' in word[-1]:	
									words.append(word)
								else:
									words.append(word+' ')
						count+=1
	 
				mwiki = ''.join(words)
				#get rid of pic placeholder tags
				mwiki=mwiki.replace("<pic>", "")
				mwiki=mwiki.replace("</pic>", "")
	 
				#make sure there are no single newlines - mediawiki just ignores them. Replace multiple lines with single and then single with double.
				while "\n\n" in mwiki or "\n \n" in mwiki: 
					mwiki=mwiki.replace("\n\n", "\n")
					mwiki=mwiki.replace("\n \n", "\n")
				mwiki=mwiki.replace('\n', '\n\n')
	 
				#replace multiple lines with single where they would break formatting - such as in a list
				mwiki=mwiki.replace('\n\n#', '\n#')
				mwiki=mwiki.replace('\n\n*', '\n*')
				mwiki=mwiki.replace('*<br/>', '*')
				mwiki=mwiki.replace('#<br/>', '#')
				mwiki = mwiki.lstrip('\n')
	 
				lines=[]
				for line in mwiki.splitlines(True):
					if line.startswith(':'):
						line='<nowiki>:</nowiki>'+line[1:]
					lines.append(line)
				mwiki=''.join(lines)
	 
				entitydefs = dict( (unichr(k), "&amp;"+v+";") for k, v in htmlentitydefs.codepoint2name.items() )
				entitydefs.pop('<')
				entitydefs.pop('>')
				entitydefs.pop('&')
				mwiki=saxutils.escape(mwiki, entitydefs)
	 
				for n in range(len(mwiki)):
					if mwiki[n]< " " and mwiki[n]!='\n' and mwiki[n]!='\r' and mwiki[n]!='\t':
						mwiki=mwiki[:n]+"?"+mwiki[n+1:]
	 
				mwiki=mwiki.replace('amp;lt;','lt;')
				mwiki=mwiki.replace('amp;gt;','gt;')
	 
				while "  " in mwiki: 
					mwiki=mwiki.replace("  ", " ")
				mwiki=mwiki.replace('&lt;!--','<!--')
				mwiki=mwiki.replace('--&gt;','-->')
	 
				# the table of contents will have been seen as bold formatting
				if len(headings)>=3: 
					mwiki = mwiki.replace("'''TOC'''", '__TOC__')
					mwiki = mwiki.replace("'''NOTOC'''", '__NOTOC__')
				else:
					mwiki = mwiki.replace("'''TOC'''\n\n", '')
					mwiki = mwiki.replace("'''NOTOC'''\n\n", '')
					mwiki = mwiki.replace("'''TOC'''\n", '')#if it's before bullets/numbers the second \n will have gone
					mwiki = mwiki.replace("'''NOTOC'''\n", '')

				outputpage = unicode(outputpage, "Latin-1")
				outputpage+=mwiki+'</text>\n'	
				outputpage+='</revision>\n'
				outputpage=outputpage.encode('utf-8')
				totalSize+=len(outputpage)
	 
				#mediawiki has a maximum import file size so start a new file after that limit
				if options.outputFile!='-':
					if totalSize > options.max*1024*1024:
						totalSize=len(unicode(outputpage, "Latin-1"))
						mwikixml.write('</page>')
						mwikixml.write('</mediawiki>')
						fileCount += 1
						mwikixml = open(outputFile[:-4]+str(fileCount)+outputFile[-4:], 'wb') 
						sys.stderr.write('Creating new wiki xml file '+outputFile[:-4]+str(fileCount)+outputFile[-4:]+'\n')
						mwikixml.write('<mediawiki xml:lang="en">\n')
						#if this isn't the first part write page and title
						mwikixml.write('<page>\n')
						mwikixml.write('<title>'+title+'</title>')				
					mwikixml.write(outputpage)
				else:
					mwikixml.write(outputpage)
			else:
				if partcount != 1:
					if sys.stdout == False:
						sys.stderr.write(str(part.get_param('pagename'))+' version '+str(part.get_param('version'))+' wasn\'t counted')
	 
		mwikixml.write('</page>')
		if uploads != []:
			filepages[title] = uploads
		pagecount+=1
mwikixml.write('</mediawiki>')
sys.stderr.write('\nnumber of pages = '+str(pagecount)+' number of versions = '+str(versioncount)+'\n')
sys.stderr.write('with contributions by '+str(authors)+'\n')
sys.stderr.write('and file uploads on these pages: '+str(filepages.keys())+'\n')

See also