#!/usr/bin/env python

"""
convert nt and wgs data (fasta format) to giNumber_seqLen
run formatdb in the command line: gunzip -c nt.gz |formatdb -i stdin -p F -n "nt.chunk" -v 2000
"""

import os, sys, math

if __name__ == '__main__':
    
    seq = [] 
    len_seq = 0
    invalid_lines = 0
    
    for i, line in enumerate(sys.stdin):
        line = line.rstrip('\r\n')
        if line.startswith('>'):
            if len_seq > 0:
                print ">%s_%d" %(gi, len_seq)
                print "\n".join(seq)
            title = line
            fields = title.split('|')
            if len(fields) >= 2 and fields[0] == '>gi':
                gi = fields[1]
            else:
                gi = 'giunknown'
                invalid_lines += 1
            len_seq = 0
            seq = []
        else:
            seq.append(line)
            len_seq += len(line)
    if len_seq > 0:
        print ">%s_%d" %(gi, len_seq)
        print "\n".join(seq)
        
    print >> sys.stderr, "Unable to find gi number for %d sequences, the title is replaced as giunknown" %(invalid_lines)