libsubtitles-perl 1.04-2 / usr / bin / subs

#!/usr/bin/perl -w
# $Id: subs,v 1.16 2008/03/07 22:24:49 dk Exp $
use strict;
use Subtitles;

my @in;
my $out  = 'out.sub';
my $jtime = 2;
my $a;
my $b;
my $d;
my @eval;
my $split;
my $separate;
my $verbose;
my $codec;
my $rate;
my @points;
my @q = (0, 'end');
my $do_q;
my $zip;
my $inplace;

sub usage
{
	my @codecs = map { s/Subtitles::Codec:://; $_} codecs;
	print <<USAGE;
subs: convert, join, split, and re-time subtitles

format:
   subs [options] subfile [ subfile ... ]

options:
   -a coeff, - a and b coefficients in linear transformation
   -b time     u=at+b, where t and u are src and dest times
               ( default(identity transform) is [a=1,b=0] ).
	       -a can be set as ratio, f.ex. 23.9/25 
   -c codec  - use codec to write file ( one of: '@codecs' )
   -d        - try to prolong duration of quickly disappearing text
   -e code   - run perl code for each line of text in file. On each
               run, the text and time variables are initialized, and 
	       new values, if any, written to the file. The variables
	       are used for:
	       \$_ - subtitle text line
	       \$b - cue beginning
	       \$e - cue end
	       \$i - line number
	       \$n - number of lines
	       \%p - persistent data between runs
	       The -e option can be specified several times
   -h        - display this text
   -i        - edit files in place ( makes backup in .bak files )
   -j sec    - time interval between joins, seconds (default 2)
   -o file   - file to save processed subtitles (default out.sub)
   -O        - separate overlapped lines
   -p t1 t2  - set a control point, where t1 is time of a phrase
               spoken in the film and t2 is time when the same 
	       phrase as appears in the subtitle. Two points are 
	       required for deducing -a and -b coefficients; if 
	       only one point is specified, it is assumed that
	       the other one is [0,0]. Times can be relative, f.ex. 
	       -p 01:00 +3.5 -p -20 1:00:00 
   -P t2 t1    same as -p but reversed arguments ( use if -p was 
               given in wrong order )
   -r rate   - force frame-per-second rate for frame-based subs.
               Useful numbers are 23.976, 24, 25.
   -s time   - split in two parts by time
   -q t1 t2  - restrict changes, if any, in time span t1-t2. Word
               'end' can be used as an alias to the end of the 
	       file. Default values are '0' and 'end'.
   -v        - be verbose
   -z file2  - zip subtitle files so time information is read from
               file2, while text information is read from input file

time format is either [[HH:]MM:]SS[.MSEC] or subtitle format-specific

(bugs to <dmitry\@karasik.eu.org>)

USAGE
	exit(0);
}


# getopt
{
	my ( $i, $do, $ddash, $set_o);
	sub nextarg
	{
		my $arg = $ARGV[++$i];
		die "Option `-$1' requires more parameters\n" unless defined $arg;
		$arg;
	}
	for ( $i = 0; $i < @ARGV; $i++) {
		if ( !$ddash && $ARGV[$i] =~ /^-(.*)$/) {
			if ( $1 eq 'h') {
				usage();
			} elsif ( $1 eq 'i') {
				$inplace = 1;
			} elsif ( $1 eq 'v') {
				$verbose = 1;
			} elsif ( $1 eq '-') {
				$ddash = 1;
			} elsif ( $1 eq 'o') {
				$out	= nextarg;
				$set_o = 1;
			} elsif ( $1 eq 'O') {
				$separate = 1;
				$do = 1;
			} elsif ( $1 eq 'j') {
				$jtime = nextarg;
				die "Invalid -j parameter\n"
					unless $jtime =~ /^[-+]?\d+(\.\d+)?/;
			} elsif ( $1 eq 'd') {
				$d = 1;
				$do = 1;
			} elsif ( $1 eq 'e') {
				push @eval, nextarg;
				$do = 1;
			} elsif ( $1 eq 'a') {
				$a = nextarg;
				$do = 1;
				die "Invalid -a parameter `$a'\n"
					unless $a =~ /^(\-?\d+(?:\.\d+)?)(?:\/(\d+(?:\.\d+)?))?/;
				if ( defined $2) {
					die "Divizion by zero\n" if $2 == 0;
					$a = $1/$2;
				}
			} elsif ( $1 eq 'c') {
				$codec = nextarg;
				my @c = map { s/Subtitles::Codec:://; $_ } codecs;
				my %c = map { $_ => 1 } @c;
				die "Invalid codec name `$codec'; valid are: @c\n"
					unless exists $c{$codec};
				$do = 1;
			} elsif ( $1 eq 'b') {
				$b = nextarg;
				$do = 1;
			} elsif ( $1 eq 'p' or $1 eq 'P') {
				die "Too many control points\n" if 2 == @points;
				my @x = (nextarg, nextarg);
				@x = reverse @x if $1 eq 'P';
				push @points, \@x;
				$do = 1;
			} elsif ( $1 eq 'q') {
				die "Too many brackets\n" if $do_q;
				@q = (nextarg, nextarg);
				$do_q = 1;
			} elsif ( $1 eq 'r') {
				$rate = nextarg;
				die "Invalid rate `$rate'\n"
					unless $rate =~ /^\d+(\.\d+)?$/;
			} elsif ( $1 eq 's') {
				$split = nextarg;
				$do = 1;
			} elsif ( $1 eq 'z') {
				$zip = nextarg;
				$do = 1;
			} else {
				die "Unknown option `-$1'\n";
			}
		} else {
			push @in, $ARGV[$i];
		}
	}
	usage() unless @in;
	$do = 1 if 1 < @in;
	die "Nothing to do!\n" unless $do;
	if ( $inplace) {
		die "-i and -o options are mutually exclusive\n" if $set_o;
		die "Cannot edit in place for more than one input file\n" if 1 < @in;
		die "Cannot edit in place for stdin input\n" if $in[0] eq '-';
		$out = $in[0];
	}
}

# read files
my $dest;
my @entries;
for my $fn ( grep { defined } ( @in, $zip)) {
	my $entry = Subtitles->new();
	$entry-> rate( $rate) if $rate;
	my $ret;
	if ( $fn eq '-') {
		$ret = $entry-> load(\*STDIN);
		$fn = 'stdin';
	} else {
		open F, "< $fn" or die "Error: cannot open $_:$!\n";
		$ret = $entry-> load(\*F);
		close F;
		$fn = "'$fn'";
	}
	die "Error loading $fn:$@\n" unless $ret;
	if ( $verbose) {
		my ( $c, $l) = ( $entry-> codec, $entry-> lines);
		$c =~ s/Subtitles::Codec:://;
		warn "read $l line(s) from $fn, codec=$c\n" if $verbose;
	}
	push @entries, $entry;
}
$zip = pop @entries if $zip;
$dest = shift @entries;

# validate time-based parameters
if ( defined $split) {
	my $s = $dest-> parse_time( $split);
	die "Cannot parse time `$split'\n" unless defined $s;
	die "`$split' is negative\n" if $s < 0;
	$split = $s;
	warn "split by ". time2str($s) . "\n" if $verbose;
}


# points
for (@points) {
	my ( $p1, $p2) = @$_;
	my ( $s1, $s2);
	$s1 = (($p1 =~ s/^([-+])//) ? $1 : '');
	$s2 = (($p2 =~ s/^([-+])//) ? $1 : '');
	my $t = $dest-> parse_time( $p1);
	die "Cannot parse time `$s1$p1'\n" unless defined $t;
	$p1 = $t;
	$t = $dest-> parse_time( $p2);
	die "Cannot parse time `$s2$p2'\n" unless defined $t;
	$p2 = $t;
	die "Both times in control point [$s1$p1,$s2$p2] are relative\n"
		if length $s1 and length $s2;
	if ( length $s1) { # $p1 is relative
		$p1 = $p2 + $p1 * (( $s1 eq '-') ? -1 : 1);
	} elsif ( length $s2) { # $p2 is relative
		$p2 = $p1 + $p2 * (( $s2 eq '-') ? -1 : 1);
	}
	$_ = [$p1,$p2];
}

unshift @points, [0,0] if 1 == @points;

if ( 2 == @points) {
	my ( $t1, $u1) = @{$points[0]};
	my ( $t2, $u2) = @{$points[1]};
	die "-p option conflicts with -a and -b\n" if defined($a) || defined($b);
	my ( $dt, $du) = ( $t2 - $t1, $u2 - $u1);
	die "Point sets refers to the same time\n" if $dt == 0 || $du == 0;
# 
# 
#    |u(subtitles)
#    |
# u2 |             *
# u1 |	  *		
#    |                      t(speech)
#    ----------------------
#	 t1       t2	  
#
	$a = $dt / $du;
	$b = $t1 - $u1 * $a;

	warn "control points [", time2str($t1), ",", time2str($u1), "], [",
		 time2str($t2), ",", time2str($u2),
		"]\n" if $verbose;
	$b = time2str( $b);
}

# a & b
$a = 1 unless defined $a;
if ( defined $b) {
	my $bb = $dest-> parse_time( $b);
	die "Cannot parse time `$b'\n" unless defined $bb;
	$b = $bb;
	warn "a=$a,b='". time2str($b) . "'\n" if $verbose;
} else {
	$b = 0;
	warn "a=$a,b=$b\n" if $verbose;
}

# process
$dest-> join( $_, $jtime) for @entries; # join

# zip
if ( $zip) {
	my $f1 = $zip->{from};
	my $t1 = $zip->{to};
	my $f2 = $dest->{from};
	my $t2 = $dest->{to};
	my $x2 = $dest->{text};
	my $n1 = @$f1;
	my $n2 = @$f2;
	if ( $n2 > $n1) {
		warn "zip: $n1 timeslices available while $n2 found in input -- ".
		  "timeframes after ". time2str( $f2->[$n1] ).
		  ", #$n1, will be left unchanged\n"
	} elsif ( $n2 < $n1) {
		warn "zip: $n2 timeslices available while only $n1 found in input -- " .
		"padding with empty lines\n";
		for ( my $i = $n2; $i < $n1; $i++) {
			push @$f2, $$f1[$i];
			push @$t2, $$t1[$i];
			push @$x2, '';
		}
		$n1 = $n2;
	}
	for ( my $i = 0; $i < $n1; $i++) {
		 ($$f2[$i], $$t2[$i]) = ( $$f1[$i], $$t1[$i]);
	}		
}

# brackets and transform
my $vv_q = "brackets [ ";
for ( @q) {
	if ( $_ eq 'end') {
		$_ = $dest-> length;
	} else {
		my $q = $dest-> parse_time( $_);
		die "Cannot parse time `$_'\n" unless defined $q;
		$_ = $q;
	}
	$vv_q .= time2str($_) . ' ';
}
warn "$vv_q ]\n" if $verbose;
$dest-> transform( $a, $b, @q);

# prolong timing
if ( $d) {
	my $i = 0;
	my $from = $dest->{from};
	my $to	= $dest->{to};
	my $text = $dest->{text};
	my $n = @$from;
	my $c = 0;
	my ($qfrom, $qto) = @q;
	$qfrom = 0 unless defined $qfrom;
	$qto	 = $$to[-1] unless defined $qto;
	for ( $i = 0; $i < $n; $i++) {
		my @clob = split("\n", $$text[$i]);
		my $min = 0.8 * @clob;
		next if $$to[$i] - $$from[$i] > $min;
		next if $$from[$i] > $qto || $$to[$i] < $qfrom;
		if ( $i < $n - 1 && $$to[$i] + $min > $$from[$i+1]) {
			$$to[$i] = $$from[$i+1] - 0.01;
		} else {
			$$to[$i] = $$from[$i] + $min;
		}
		$c++;
	}
	warn "$c lines prolonged\n";
}

# -O
if ( $separate) {
	my $from = $dest->{from};
	my $to	= $dest->{to};
	my $text = $dest->{text};
	my $n = @$from;
	my $c = 0;
	my ($qfrom, $qto) = @q;
	$qfrom = 0 unless defined $qfrom;
	$qto	 = $$to[-1] unless defined $qto;
	for ( my $i = 0; $i < $n - 1; $i++) {
		next if $$from[$i] > $qto || $$to[$i + 1] < $qfrom;
		my $d = $$to[$i] - $$from[$i + 1];
		next if $d < 0;
		$d = 0.002 if $d < 0.002;
		$d /= 2.0;
		$$to[$i]		 -= $d;
		$$from[$i+1] += $d + 0.001;
		$c++;
	}
	warn "$c overlapped lines separated\n";
}

# -e
if ( @eval) {
	my $i = 0;
	my $from = $dest->{from};
	my $to	= $dest->{to};
	my $text = $dest->{text};
	my $n = @$from;
	my ($qfrom, $qto) = @q;
	$qfrom = 0 unless defined $qfrom;
	$qto	 = $$to[-1] unless defined $qto;
	my %p;
	
	for my $eval ( @eval) {
		for ( $i = 0; $i < $n; $i++) {
			local $_ = $$text[$i];
			my $b = $$from[$i];
			my $e = $$to[$i];
			next if $b > $qto || $e < $qfrom;
			eval $eval;
			die "error in '$eval': $@" if $@;
			$$text[$i] = $_;
			$$from[$i] = $b;
			$$to[$i] = $e;
		}
	}
}

$dest-> codec( "Subtitles::Codec::$codec") if defined $codec;
my @write;

if ( defined $split) {
# split & save
# no inplace logic - original file is never overwritten
	my ( $s1, $s2) = $dest-> split( $split);
	my $root = $out;
	$root =~ s/(\.[^\.]*)$//;
	my $tail = defined($1) ? $1 : '';
	warn "write ".$s1->lines." line(s) in '$root.1$tail'\n" if $verbose;
	open F, "> $root.1$tail" or die "Cannot open $root.1$tail:$!\n";
	$s1-> save(\*F) or die "Error saving $root.2$tail:$@\n";
	close F;
	warn "write ".$s2->lines." line(s) in '$root.2$tail'\n" if $verbose;
	open F, "> $root.2$tail" or die "Cannot open $root.2$tail:$!\n";
	$s2-> save(\*F) or die "Error saving $root.2$tail:$@\n"; 
	close F;
} else {
# just save
	warn "write ".$dest->lines." line(s) in '$out'\n" if $verbose;
	my $rename = $inplace or -f $out;
	if ( $rename) {
		rename $out, "$out.bak" or die "Cannot rename $out to $out.bak:$!\n";
	}
	eval {
		open F, "> $out" or die "Cannot open $out:$!\n";
		$dest-> save(\*F) or die "Error saving $out:$@\n";
		close F;
	};
	if ( $@) {
		rename "$out.bak", $out if $rename;
		die $@;
	}
}
# done
exit(0);


__DATA__

=pod

=head1 NAME

subs - convert, join, split, and re-time subtitles

=head1 FORMAT

	subs [options] subfile [ subfile ... ]

=head1 OPTIONS

=over

=item -a coeff, -b time

a and b coefficients in linear transformation u=at+b, where t and u are src and
dest times ( default(identity transform) is [a=1,b=0] ).  -a can be set as
ratio, f.ex. 23.9/25 

=item -c codec  

Use codec to write file. Run 'subs -h' for list of installed codecs.

=item -d
	
Try to prolong duration of quickly disappearing text.
'Quickly' is less than 0.8 second per line of text.

=item -e command

Run perl code for each line of text in file. On each run, the text and time
variables are initialized, and new values, if any, written to the file. The
variables are used for:

=over

=item $_

subtitle text line

=item $b

cue beginning

=item $e

cue end

=item $i

line number

=item $n

number of lines

=item %p

persistent data between runs

=back

The -e option can be specified several times

=item -h

Display help

=item -i

Edit files in place ( makes backup in .bak files )

=item -j sec

Time interval between joins, seconds (default 2)

=item -o file

File to save processed subtitles (default out.sub)

=item -O

Separate overlapped lines

=item -p t1 t2 or -P t2 t1

Set a control point, where t1 is time of a phrase spoken in the film and t2 is
time when the same phrase as appears in the subtitle. Two points are required
for deducing -a and -b coefficients; if only one point is specified, it is
assumed that the other one is [0,0].

Times can be relative, f.ex. -p 01:00 +3.5 -p -20 1:00:00

Options -P and -p are the same except the argument sequence is reversed.  -P is
to be used when arguments to -p were typed manually and in wrong order.

=item -q t1 t2

Restrict changes, if any, in time span t1-t2. Word 'end' can be used as an
alias to the end of the file. Default values are '0' and 'end'.

=item -r rate

Force frame-per-second rate for frame-based subs

=item -s time

Split in two parts by time

=item -v

Be verbose

=item -z file.sub
   
Zip subtitle files so time information is read from file.sub, while text
information is read from the input file(s).

=back

=head1 NOTES

The time format is either [[HH:]MM:]SS[.MSEC] or subtitle format-specific

=head1 EXAMPLES

Warning: -i is a great feature, but use it with certain caution.

If subtitles are shown too early ( 5 seconds):

  subs -i -b 5 file.sub

If subtitles are for a movie in 25 fps, need to be for 24
( actual for frame-based formats only ).

  subs -i -a 24/25 file.sub

If subtitles start ok, but in 1 hour are late in 7 seconds:

  subs -i -p 0 0 -p 1:00:00 +7 file.sub

Join two parts with 15-second gap
  
  subs -o joined.sub -j 15 part1.sub part2.sub 

Split in two after 50 minutes and half a second ( makes basename.1.sub
and basename.2.sub ).

  subs -o basename.sub -s 50:00.5 toobig.sub

Remove closed caption-specific comments such as '[Sneezing]' or '[Music playing]'

  subs -e 's/[\s-]*\[.*\]\s*\n*//gs' sub.sub

=head1 BUGS

Subtitles written as C<.smi> format may differ from
original.

=head1 SEE ALSO

L<Subtitles> - backend module for this program

=head1 AUTHOR

Dmitry Karasik, E<lt>dmitry@karasik.eu.orgE<gt>.

=cut