#!/usr/bin/env python
##coding=utf-8

import re
import os
import sys
import argparse

parser = argparse.ArgumentParser(description="\n\033[1;32;40m \
Usage: python vcfstat.py -i [sample.anno.hg19_multianno.vcf] -o [stat.sv.txt]\033[0m ")

parser.add_argument('-i', '--inputvcf', help = 'the sample vcf information file', required = True)
parser.add_argument('-o', '--output', help = 'the analysis of the pathway', required = True)
argv = vars(parser.parse_args())
inF = os.path.abspath(argv['inputvcf'].strip())
ouF = os.path.abspath(argv['output'].strip())

def drawvcf(INF,OUF):
    ouc=open(OUF,'w')
    ouc.write("SV_Lenght\tSV_type\n")
    with open(INF,'r') as l:
        for li in l:
            lis=li.strip().split('\t')
            if str(lis[0]).startswith("#"):
                continue
            else:
                info=str(lis[7]).strip().split(";")
                for infol in info:
                    if "SVTYPE" in str(infol):
                        infoli=str(infol).strip().split("=")
                        typekeyi=infoli[1]
                    elif str(infol).startswith("SVLEN"):
                        infoli=str(infol).strip().split("=")
                        svlen=abs(int(infoli[1]))
                        ouc.write(str(svlen)+"\t"+str(typekeyi)+"\n")

statvcf(inF,ouF)