n50.py

import sys
import argparse
import glob, os



def main():
    parser = argparse.ArgumentParser(description='Reference Culling')
    parser.add_argument('-i', '--insize', help='Path to output directory.', required=True)
    parser.add_argument('-r', '--ref_file', help='Path to coordinate file generated by align_reads.sh.', required=True)
    args = parser.parse_args()
    ref=args.ref_file
    in_file=args.insize
    size_list=[]
    with open(in_file) as file:
        for line in file:
            line=line.replace("\n", "")
            temp=line.split('\t')
            size_list.append(int(temp[1]))
    size_ref=[]
    with open(ref) as file:
        for line in file:
            line=line.replace("\n", "")
            temp=line.split('\t')
            size_ref.append(int(temp[1]))
    calc_n50(size_list,sum(size_ref))




def calc_n50(array,ref_size):
    array.sort(reverse = True)
    #print(array)
    n50 = 0
    n = 0 
    half = ref_size/2
    for val in array:
        n50 += val
        if n50 >= half:
            n = val
            break 
    print("\n NG50:\n"+str(n))



if __name__ == "__main__":
    main()