mirror of
https://github.com/moses-smt/mosesdecoder.git
synced 2024-12-29 15:04:05 +03:00
ef028446f3
This is not pleasant to read (and much, much less pleasant to write!) but sort of necessary in an open project. Right now it's quite hard to figure out what is licensed how, which doesn't matter much to most people but can suddenly become very important when people want to know what they're being allowed to do. I kept the notices as short as I could. As far as I could see, everything without a clear license notice is LGPL v2.1 or later.
80 lines
2.2 KiB
Perl
Executable File
80 lines
2.2 KiB
Perl
Executable File
#!/usr/bin/env perl
|
|
#
|
|
# This file is part of moses. Its use is licensed under the GNU Lesser General
|
|
# Public License version 2.1 or, at your option, any later version.
|
|
|
|
# $Id$
|
|
# given a list of files, combines them to a single corpus (sent to stdout)
|
|
|
|
use strict;
|
|
use warnings;
|
|
use Getopt::Long;
|
|
use IO::File;
|
|
use File::Basename;
|
|
|
|
binmode(STDIN, ":utf8");
|
|
binmode(STDOUT, ":utf8");
|
|
binmode(STDERR, ":utf8");
|
|
|
|
my @addfactors = @ARGV;
|
|
die "usage: combine_factors.pl corpusfile1 corpusfile2 ..."
|
|
if 0 == scalar @addfactors;
|
|
|
|
my @streams = map {
|
|
my $fn = $_;
|
|
my $opn = ($fn =~ /\.gz$/ ? "zcat $fn |" : "$fn");
|
|
my $stream = new IO::File;
|
|
$stream->open($opn) or die "Can't open '$opn'";
|
|
binmode($stream, ":utf8");
|
|
$stream;
|
|
} @addfactors;
|
|
|
|
my $nr=0;
|
|
my $firststream = shift @streams;
|
|
shift @addfactors; # just to keep the lengths sync'ed
|
|
$_ = readline($firststream);
|
|
while (defined $_) {
|
|
$nr++;
|
|
print STDERR "." if $nr % 10000 == 0;
|
|
print STDERR "($nr)" if $nr % 100000 == 0;
|
|
chomp;
|
|
s/\s+/ /g; s/^ //; s/ $//;
|
|
my @intokens = split / /;
|
|
# load lines of corresponding streams and ensure equal number of words
|
|
my @lines_of_extratoks;
|
|
foreach my $factor (0..$#streams) {
|
|
my $line = readline($streams[$factor]);
|
|
die "Additional factor file $addfactors[$factor] contains too few sentences!"
|
|
if !defined $line;
|
|
chomp($line);
|
|
$line =~ s/\s+/ /g; $line =~ s/^ //; $line =~ s/ $//;
|
|
my @toks = split / /, $line;
|
|
die "Incompatible number of words in factor $factor on line $nr. ($#toks != $#intokens)"
|
|
if $#toks != $#intokens;
|
|
$lines_of_extratoks[$factor] = \@toks;
|
|
}
|
|
|
|
# for every token, print the factors in the order as user wished
|
|
for(my $i=0; $i<=$#intokens; $i++) {
|
|
my $token = $intokens[$i];
|
|
my @outtoken = ();
|
|
push @outtoken, $token; # add the first one
|
|
# print STDERR "Token: $token\n";
|
|
foreach my $factor (0..$#streams) {
|
|
my $f = $lines_of_extratoks[$factor]->[$i];
|
|
die "Missed factor value for word $i+1 on line $nr in $addfactors[$factor]"
|
|
if !defined $f || $f eq "";
|
|
push @outtoken, $f;
|
|
}
|
|
print " " if $i != 0;
|
|
print join("|", @outtoken);
|
|
}
|
|
print "\n";
|
|
$_ = readline($firststream);
|
|
}
|
|
close $firststream;
|
|
print STDERR "Done.\n";
|
|
|
|
|
|
|