#!/usr/bin/perl
#
# Translate input based on a given hash file 
#
# (C) Gal Nov 2003 

use strict;
require "$ENV{HOME}/develop/perl/lib/libhash.pl"; # for load_list

# Init
my $n_items=1;
my $fin = \*STDIN;
my $hash_filename;
my $fieldnum=0;
my $do_field=0;
my $back=0;
my $verbose=0;
my $missing;
my $do_missing=0;
my $do_add=0;
my $delimiter = "\t";
my %hashtab;
my $strrep=0;

# Check flags
while(@ARGV){
  my $arg = shift @ARGV;
  if($arg eq '--help')   {print STDOUT <DATA>; exit(0);
  }elsif($arg eq '-f')   {$hash_filename = shift @ARGV;
  }elsif($arg eq '-n')   {$fieldnum=shift @ARGV;$fieldnum--;$do_field=1;
  }elsif($arg eq '-m')   {$missing=shift @ARGV;$do_missing=1;
  }elsif($arg eq '-b')   {$back=1;
  }elsif($arg eq '-v')   {$verbose=1;
  }elsif($arg eq '-add') {$do_add=1;$delimiter = shift @ARGV;
  }elsif($arg eq '-strrep') {$strrep=1;
  }else{
    die("translate_byhash.pl: Bad argument '$arg'. Use --help for help.");
  }
}
if($verbose){print "verbose turned on\n";}

# Read exiting hash table
if($back==1){
  if (-s $hash_filename){
    %hashtab = load_hash_back($hash_filename);
    $n_items = scalar(keys %hashtab)-1;
  }
}else{
  if (-s $hash_filename){
    %hashtab = load_hash($hash_filename);
    $n_items = scalar(keys %hashtab)-1;
  }
}
my $name;

# Free-text replacement
if($strrep){
  my $len;
  while(my $line = <$fin>){
    while (my ($key, $val) = each(%hashtab)){
      $line =~ s/$key/$val/g;
    }
    print "$line";
  }
  exit(0);
}


# Tabular replacement
while(<$fin>)  {
  chop($_);
  my @row = split(/\t/,$_);

  for(my $i=0;$i<@row;$i++){
    $name = $row[$i];

    if($do_field == 0 | $i == $fieldnum){
      # Print replaced field
      if($do_add){print  "$name$delimiter";}
      if(!defined($hashtab{$name})){	
	if($do_missing){
	  print  "$missing";
	  if($verbose) {print STDERR "$name ==> $missing\n";}
	}else{
	  print  "$name";
	  if($verbose) {print STDERR "'$name' not found\n";}
	}
      }else{	
	print  "$hashtab{$name}";
	if($verbose) {print STDERR "$name ==> $hashtab{$name}\n";}
      }	
      if($i<@row-1){ print "\t";};
    }else{
      # Print original field
      print "$row[$i]\t";
    }
  }
  print "\n";
}






#----------------------------------------------------------------
# load backwards hash from file
#----------------------------------------------------------------
sub load_hash_back
{
  my ($file) = @_;
  open(HASHFILE, "<$file")|| die "Cant open $file: $!\n";
  while (my $line = <HASHFILE>){
    chop($line);
    my @row = split(/\t/,$line);
    my $key = $row[1];
    my $val = $row[0];
    $hashtab{$key} = $val;
  }
  close(HASHFILE);
  return %hashtab;
}
__DATA__

Syntax: translate_byhash.pl  -f FILENAME [OPTIONS] < STDIN
Translate strings in input using a dictionary file. 
input is assumed to be tab delimited.



OPTIONS are: 

   -f FILENAME:
      Filename with hash table. Each row contains two fields:
      key, value

   -m MISSING:
      A string to fill when key is not found in hashfile. the default
      is to leave the string unchanged

   -n N
      Translate only the Nth field in each row (N=1 is the first field)

   -b BACK
      Translate back by the same hash table

   -add DELIMITER
      Instead of replacing each source string with target, print both.
      delimiter by DELIMITER

   -strrep
      In this mode, strings from hashfile are matched and replaced,
      regardless of columns structre. That is, tabs are treated just
      as regular cgaracters.

  Example:
    cat in_file | translate_byhash.pl  -f directory_file > out_file