#!/usr/bin/perl
#twolout.pl
$| = 1;
use Getopt::Std;
$right = 0;
getopts('s');
while (<>) {
    &prefix;
    if (/^$/) { 
	print $_;
#	printf STDERR "\ntwolout: '$token'";
	next;
    }
    if (/^#r=(.+)#t=(.+)$/) {
	$r = $1; $total = $2;
#	printf STDERR "\ntwolcount: $r $total";
	next;
    }
    if (/^[(]?["][<][>]["]$/) { # empty lines
	$right = 0;
#	printf STDERR "\ntwolout: '\n'";
#	print "\n";
	next;
    }
    if (/^[(]?["][<]([0-9]+[^ ]*)[>]["]$/) { # numbers
	$left = $right; $right++; 
	$token = $1; $base = $token; 
	$token =~ s/^[\$][\\]//;
	$edge = "[<l> = $left <r> = $right ";
	$tags = "NUM NOM";
	$p = sprintf "%0.1e", 1;
	$reading = $edge . "<p> = $p <e> = [<s> = \"$token \" <base> = '$base' '$base' ";
	$reading .= "<ss> = \"$token \" <bs> = \"$base\" " if ($opt_s); 
	$reading .= "$tags]] ";
	printf $reading;
#	printf STDERR "\ntwolout: '$reading'";
        next;
    }
    s/\\!/!/g;
    if (/^[(]?["][<](.)[>]["]$/) { # punctuation
	$left = $right; $right++; 
	$token = $1; $base = $token; 
	$token =~ s/^[\$][\\]//;
	$edge = "[<l> = $left <r> = $right ";
	$tags = "PUNC";
	$p = sprintf "%0.1e", 1;
	$reading = $edge . "<p> = $p <e> = [<s> = \"$token \" <base> = '$base' '$base' ";
	$reading .= "<ss> = \"$token \" <bs> = \"$base\" " if ($opt_s); 
	$reading .= "$tags]] ";
	printf $reading;
#	printf STDERR "\ntwolout: '$reading'";
	$right = 0;
        next;
    }
    if (/^[(]?["][<](.+)[>]["]$/) { # tokens 
	$left = $right; $right++; 
	$token = $1;
	$token =~ s/^[\$][\\]//;
	$edge = "[<l> = $left <r> = $right ";
    }
    if (/^\t[(]?["](.+)["]  /) { # readings
	$base = $1; $tags = $';
	$base =~ s/^[\$][\\]//;
	chop $tags;
	for ($tags) {
	    s/[ )]*$//;
	    @array = split(/[ =-]/);
	    $length = $#array+1;
	    s/.* = //;
        }
#	printf STDERR "\ntwolout length: $length";
	$p = (($total-$length) || 1)/((($r-1)*$total) || 1);
	$p = sprintf "%0.1e", $p;
#	    printf STDERR "\ntwolout p: $p = (($total-$length) || 1)/((($r-1)*$total) || 1)\n";
	$reading = $edge . "<p> = $p <e> = [<s> = \"$token \" <base> = '$base' '$base' ";
	$reading .= "<ss> = \"$token \" <bs> = \"$base\" " if ($opt_s); 
	$reading .= "$tags]] ";
	&postfix;
#	printf STDERR "\ntwolout: '$reading'";
	printf $reading;
    }
#    if ($tags =~ /PUNC/) { 
#	$right = 0;
#    }
    $tags = "";
}
#    printf "\n\n";

sub prefix {
	s/\"seurata\"  PCP1 ACT/"seuraava"  A/; # note TWO spaces!
	s/(?<=\"mikä\"  )Q/FAIL/;
	s/(?<=\"kiitos\"  )N NOM SG/INTJ/;
	s/(?<=\"koska\"  )SUB C/INTERR ADV/;
	s/(?<=\"entä\"  INTERR )ADV/CF/;
	s/(?<=\"jälkeen\"  )ADV ILL/FAIL/;
}

sub postfix {
    for ($reading) {
	s/(?<='minä' PERS PRON)(.*) SG/$1 SG1/; # person marking
	s/(?<='sinä' PERS PRON)(.*) SG/$1 SG2/;
	s/(?<='me' PERS PRON)(.*) PL/$1 PL1/;
	s/(?<='te' PERS PRON)(.*) PL/$1 PL2/;
	s/(?<!PERS PRON) NOM PL(?=\W)/ NOM_PL/; # these are portmanteaus
	s/(?<!PERS PRON) NOM SG(?=\W)/ NOM_SG/;
	s/(?<!PERS PRON) GEN SG(?=\W)/ GEN_SG/;
	s/ PROP N/ PROP_N/; 
	s/NEGV V/NEGV_V/; 
	s/COP V/COP_V/; 
    }
}
# possible tagging penalties:
# length
# compounding 
# derivation penalty