#!/usr/bin/python import sys import getopt import re def compare(f1,f2,o1,o2,si_line): lines_count=0; in1 = open(f1,"r") in2 = open(f2,"r") ou1 = open(o1,"w") ou2 = open(o2,"w") while 1: line1 = in1.readline() line2 = in2.readline() if line1 and line2: lines_count=lines_count+1 print >>ou1,re.sub(r' ',"",line1) ou2.write(line2) if si_line<lines_count: break else: break f1.close() f2.close() o1.close() o2.close() return lines_count def main(): opts,args = getopt.getopt(sys.argv[1:],'h:',['fq1=','fq2=','out1=','out2=','Rlen=','Dataset=']) fq1=opts[0][1] fq2=opts[1][1] out1=opts[2][1] out2=opts[3][1] readlength = opts[4][1] datasize = opts[5][1] size_lines = 4*int(datasize)*1e9/(int(readlength)*2) count = compare(fq1,fq2,out1,out2,size_lines)-1 real_size = (count/4)*2*int(readlength)/1e9 print datasize print real_size if __name__ == "__main__": main()