#!/usr/bin/env python
"""
May 27, 2016, Dr. Brian Fristensky, University of Manitoba
delgi.py - remove GI numbers from GenBank flatfiles,
    in compliance with the NCBI phase out of GI numbers.

     http://www.ncbi.nlm.nih.gov/news/03-02-2016-phase-out-of-GI-numbers/
            

 Synopsis:
  delgi.py infile 
       infile - a GenBank flat with GI numbers file
       outfile - a GenBank flat file without GI numbers

       Given an infile whose name is in the form name.gen,
       the original file will be saved as name_ig.gen, and the 
       output will be written to name.gen. 

@modified: May 27 2016
@author: Brian Fristensky
@contact: frist@cc.umanitba.ca
"""


import sys
import os
import re
import shutil
import sys


PROGRAM = "delgi.py: "
USAGE = "\n\t USAGE: delgi.py infile outfile"
 
        
#======================== MAIN PROCEDURE ==========================

def main():
    """
        Called when not in documentation mode.
        """

    #set filenames and open files
    IFN = sys.argv[1]
    OFN = sys.argv[2]
    infile=open(IFN,'r')
    outfile=open(OFN,'w')

    exp1 = '/db_xref=\"GI:'
    exp2 = '(^VERSION[ ]+.+)([ ]+GI:\d+)'
    exp3 = '\g<1>' 

    for line in infile:       
        if not exp1 in line: 
            line = re.sub(exp2, exp3, line)
            outfile.write(line)

    infile.close()
    outfile.close()
    
	
if ( "-test" in sys.argv):
    pass
else:
    main()