+ body = re.sub(r'\s*()+\s*', r'\n\n', body)
+
+ # Because a leading or trailing non break tag will break the following code, we have to mess around rather badly for a few lines.
+ body = body.replace(u'[',u'&squareBracketStart;')
+ body = body.replace(u']',u'&squareBracketEnd;')
+ body = body.replace(u'
',u'[br /]')
+
+ breaksRegexp = [
+ re.compile(r'([^\]])(\[br\ \/\])([^\[])'),
+ re.compile(r'([^\]])(\[br\ \/\]){2}([^\[])'),
+ re.compile(r'([^\]])(\[br\ \/\]){3}([^\[])'),
+ re.compile(r'([^\]])(\[br\ \/\]){4}([^\[])'),
+ re.compile(r'([^\]])(\[br\ \/\]){5}([^\[])'),
+ re.compile(r'([^\]])(\[br\ \/\]){6}([^\[])'),
+ re.compile(r'([^\]])(\[br\ \/\]){7}([^\[])'),
+ re.compile(r'([^\]])(\[br\ \/\]){8}([^\[])'),
+ re.compile(r'(\[br\ \/\]){9,}')]
+
+ breaksCount = [
+ len(breaksRegexp[0].findall(body)),
+ len(breaksRegexp[1].findall(body)),
+ len(breaksRegexp[2].findall(body)),
+ len(breaksRegexp[3].findall(body)),
+ len(breaksRegexp[4].findall(body)),
+ len(breaksRegexp[5].findall(body)),
+ len(breaksRegexp[6].findall(body)),
+ len(breaksRegexp[7].findall(body))]
+
+ breaksMax = 0
+ breaksMaxIndex = 0;
+
+ for i in range(len(breaksCount)):
+ if breaksCount[i] > breaksMax:
+ breaksMax = breaksCount[i]
+ breaksMaxIndex = i
+
+ # Find all instances of consecutive breaks less than otr equal to the max count use most often
+ # replase those tags to inverted p tag pairs, those with more connsecutive breaks are replaced them with a horisontal line
+ for i in range(len(breaksCount)):
+ if i <= breaksMaxIndex:
+ body = breaksRegexp[i].sub(r'\1\n\3', body)
+ else:
+ body = breaksRegexp[i].sub(r'\1
\n
\n\3', body)
+
+ body = breaksRegexp[8].sub(r'
\n
\n', body)
+
+ # Reverting the square brackets
+ body = body.replace(u'[', u'<')
+ body = body.replace(u']', u'>')
+ body = body.replace(u'&squareBracketStart;', u'[')
+ body = body.replace(u'&squareBracketEnd;', u']')
+
+ # If for some reason, a third break makes its way inside the paragraph, preplace that with the empty paragraph for the additional linespaing.
+ body = re.sub(r'
\s*()+', r'
\n', body)
+
+ # change empty p tags to include a br to force spacing.
+ body = re.sub(r'
\s*
', r'
', body)
+
+ # Clean up hr tags, and add inverted p tag pairs
+ body = re.sub(r'\s*\s*', r'\n
\n', body)
+
+ # Because the previous regexp may cause trouble if the hr tag already had a p tag pair around it, w nee dot repair that.
+ # Repeated opening p tags are condenced to one. As we added the extra leading opening p tags, we can safely assume that
+ # the last in such a chain must be the original. Lets keep its attributes if they are there.
+ body = re.sub(r'\s*(]*>\s*)+]*)>\s*', r'\n', body)
+ # Repeated closing p tags are condenced to one
+ body = re.sub(r'\s*(<\/\s*p>\s*){2,}', r'
\n', body)
+
+ # superflous cleaning, remove whitespaces traling opening p tags. These does affect formatting.
+ body = re.sub(r']*)>\s*', r'', body)
+ # superflous cleaning, remove whitespaces leading closing p tags. These does not affect formatting.
+ body = re.sub(r'\s*', r'', body)
+
+ # re-wrap in div tag.
+ body = u'\n' + body + u'\n
'
+
+ return body
+