gp_norm_trans_FR.pl 3.49 KB
Newer Older
1
#!/usr/bin/env perl -w
Milos Janda's avatar
Milos Janda committed
2

Arnab Ghoshal's avatar
Arnab Ghoshal committed
3
# Copyright 2012  Milos Janda
Milos Janda's avatar
Milos Janda committed
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#  http://www.apache.org/licenses/LICENSE-2.0
#
# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
# MERCHANTABLITY OR NON-INFRINGEMENT.
# See the Apache 2 License for the specific language governing permissions and
# limitations under the License.


# This script normalizes the GlobalPhone French transcripts that have been 
# extracted in a format where each line contains an utterance ID followed by
# the transcript, e.g:
# FR001_16 C'est un langage oriente+ objet ide+al pour les re+seaux ses programmes
Arnab Ghoshal's avatar
Arnab Ghoshal committed
23
# The normalization is similar to that in 'gp_norm_dict_FR.pl' script.
Milos Janda's avatar
Milos Janda committed
24

Arnab Ghoshal's avatar
Arnab Ghoshal committed
25
my $usage = "Usage: gp_norm_trans_FR.pl [-r|-u] -i transcript > formatted\
Milos Janda's avatar
Milos Janda committed
26
Normalizes transcriptions for GlobalPhone French. The input format is \
Arnab Ghoshal's avatar
Arnab Ghoshal committed
27 28 29 30
assumed to be utterance ID followed by transcript on the same line.
Input is assumed to be ISO-8859-1 encoded, and output is in UTF-8. \
Transcript is lowercased by default, but can be uppercased with the -u option.
\n";
Milos Janda's avatar
Milos Janda committed
31 32 33 34 35 36 37 38

use strict;
use Getopt::Long;
use Unicode::Normalize;
use open ':encoding(iso-8859-1)';
binmode(STDOUT, ":encoding(utf8)");

die "$usage" unless(@ARGV >= 1);
Arnab Ghoshal's avatar
Arnab Ghoshal committed
39
my ($in_trans, $keep_rmn, $uppercase);
Milos Janda's avatar
Milos Janda committed
40
GetOptions ("u" => \$uppercase,    # convert words to uppercase
Arnab Ghoshal's avatar
Arnab Ghoshal committed
41
            "r" => \$keep_rmn,     # keep words in GlobalPhone-style ASCII (rmn)
Milos Janda's avatar
Milos Janda committed
42 43 44 45
	    "i=s" => \$in_trans);  # Input transcription

open(T, "<$in_trans") or die "Cannot open transcription file '$in_trans': $!";
while (<T>) {
Arnab Ghoshal's avatar
Arnab Ghoshal committed
46
  s/\r//g;  # Since files may have CRLF line-breaks!
Milos Janda's avatar
Milos Janda committed
47 48 49 50 51 52 53 54 55 56
  chomp;
  $_ =~ m:^(\S+)\s+(.+): or die "Bad line: $_";
  my $utt_id = $1;
  my $trans = $2;

  $trans =~ s/\"/ /g;  # Remove quotation marks.
  $trans =~ s/[\,\.\?\!\:]/ /g;
  # Normalize spaces
  $trans =~ s/^\s*//; $trans =~ s/\s*$//; $trans =~ s/\s+/ /g;

Arnab Ghoshal's avatar
Arnab Ghoshal committed
57
  $trans = &rmn2utf8_FR($trans) unless (defined($keep_rmn));
Milos Janda's avatar
Milos Janda committed
58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111
  if (defined($uppercase)) {
    $trans = uc($trans);
  } else {
    $trans = lc($trans);
  }

  print "$utt_id $trans\n";
}


sub rmn2utf8_FR {
  my ($in_str) = "@_";

  # CAPITAL LETTERS
  $in_str =~ s/A\`/\x{00C0}/g; # LETTERS WITH GRAVE
  $in_str =~ s/E\`/\x{00C8}/g; 
  $in_str =~ s/U\`/\x{00D9}/g;

  $in_str =~ s/A\^/\x{00C2}/g; # LETTERS WITH CIRCUMFLEX
  $in_str =~ s/E\^/\x{00CA}/g;
  $in_str =~ s/I\^/\x{00CE}/g;
  $in_str =~ s/O\^/\x{00D4}/g;
  $in_str =~ s/U\^/\x{00DB}/g;

  $in_str =~ s/E\:/\x{00C8}/g; # LETTERS WITH DIAERESIS
  $in_str =~ s/I\:/\x{00CF}/g;
  $in_str =~ s/U\:/\x{00DC}/g;
  $in_str =~ s/Y\:/\x{0178}/g;

  $in_str =~ s/E\+/\x{00C9}/g; # E WITH ACUTE
  $in_str =~ s/C\~/\x{00C7}/g; # C WITH CEDILLA


  # SMALL LETTERS
  $in_str =~ s/a\`/\x{00E0}/g; # LETTERS WITH GRAVE
  $in_str =~ s/e\`/\x{00E8}/g; 
  $in_str =~ s/u\`/\x{00F9}/g;

  $in_str =~ s/a\^/\x{00E2}/g; # LETTERS WITH CIRCUMFLEX
  $in_str =~ s/e\^/\x{00EA}/g;
  $in_str =~ s/i\^/\x{00EE}/g;
  $in_str =~ s/o\^/\x{00F4}/g;
  $in_str =~ s/u\^/\x{00FB}/g;

  $in_str =~ s/e\:/\x{00E8}/g; # LETTERS WITH DIAERESIS
  $in_str =~ s/i\:/\x{00EF}/g; 
  $in_str =~ s/u\:/\x{00FC}/g;
  $in_str =~ s/y\:/\x{00FF}/g;

  $in_str =~ s/e\+/\x{00E9}/g; # E WITH ACUTE
  $in_str =~ s/c\~/\x{00E7}/g; # C WITH CEDILLA

  return NFC($in_str);  # recompose & reorder canonically
}