import re,sys, os, cPickle

titleReg = re.compile(r'<title>([^<]+)<\/title>', re.MULTILINE | re.DOTALL)
#articleReg = re.compile(r'<text xml:space="preserve">(.*?)</text>', re.MULTILINE | re.DOTALL)

articleReg = re.compile(r'<text xml:space="preserve">(.*?)</text>|<text xml:space="preserve"\s?/>', re.MULTILINE | re.DOTALL)

def replaceEntities(input):
	input = input.replace('&lt;','<')
	input = input.replace('&gt;','>')
	input = input.replace('&apos;',"'")
	input = input.replace('&quot;','"')
	input = input.replace('&amp;','&')
	return input

def removetitledash(input):
	input = input.replace('/','_')
	return input


def leniter(iterator):
    """leniter(iterator): return the length of an iterator,
    consuming it."""
    if hasattr(iterator, "__len__"):
        return len(iterator)
    nelements = 0
    for _ in iterator:
        nelements += 1
    return nelements

def folderIncrement(i):
	stringInt = str(i)
	while True:
		if len(stringInt) < 3:
			stringInt = '0'+stringInt
		else:
			return stringInt




def main():
    if len(sys.argv) < 2:
    	print("Provide wikidump name")
    	sys.exit(0)
    f = open(sys.argv[-1],'r')
    breakPointFile = open('breakPoints.py','r')
    breakPoints = cPickle.load(breakPointFile)
    
    currentBreak = 0
    #int = 0
    folderName = 'filedumps/'
    if not os.path.exists(folderName):
				os.mkdir(folderName)
    
    breakNum = 0
    for breakPoint in breakPoints:
    	print "Break number %s of %s " % (breakNum, len(breakPoints))
    	breakNum = breakNum + 1
    	actualBreak = breakPoint - currentBreak
    	#int,folderName = parseWiki(f,actualBreak,int,folderName)
    	dump = f.read(actualBreak)
    	newSplit = open(folderName+"/"+folderIncrement(breakNum)+".txt",'w')
    	newSplit.write(dump)
    	newSplit.close()
    
    	
    	#print "Titleiterlength is %s" % str(int)
    	currentBreak = breakPoint
    
    dump = f.read()
    newSplit = open(folderName+"/"+folderIncrement(breakNum+1)+".txt",'w')
    newSplit.write(dump)
    newSplit.close()
    
    f.close()
    
            


if __name__ == '__main__':
    main()