#!/usr/bin/perl
#
# A perl script that "babbles" to create new words
# It requires an input file of legal onsets, nuclei, and codas (see Sample.inv for an example)
# Written by Adam Albright (albright@ucsc.edu); last modified 2/16/03

use Math::Round;

print "Enter file of onsets, nuclei, and codas: ";
$inventory_file = <STDIN>;
chomp($inventory_file);
open (INV_FILE, "<$inventory_file") or die "Warning: can't open inventory file: $!\n";

print "Enter number of words desired: ";
$target_words = <STDIN>;
chomp($target_words);
#robustness: should check to make sure it is an integer

print "Enter maximum number of syllables: ";
$max_syllables = <STDIN>;
chomp($target_words);
#robustness: should check to make sure it is an integer

print "Save as file: ";
$outputfile = <STDIN>;
chomp($outputfile);
open (OUTPUTFILE, ">$outputfile") or die "Warning: Can't open output file: $!\n";

# first read in the inventory file
# the structure of an inventory file is onsets, nuclei, and then codas.
# they should be separated by labels ("Onsets", "Nuclei", "Codas") on individual lines,
# and each possible phoneme should be on its own line.
$line = <INV_FILE>;
chomp($line);
if ($line eq "Onsets") { # ok, can skip this line and move on to the next
    $line = <INV_FILE>;
    chomp($line);
    my $i = 0;
    print "Onsets:";
    while ($line ne "Nuclei") {
	$onsets[$i] = $line;
	$i++;
	print " $line";
	$line = <INV_FILE>;
	chomp($line);
    }
    # now we've hit nuclei; can skip this line and move on to the next
    $line = <INV_FILE>;
    chomp($line);
    $i = 0;
    print "\nNuclei: ";
    while ($line ne "Codas" and $line ne "") {
	$nuclei[$i] = $line;
	$i++;
	print " $line";
	$line = <INV_FILE>;
	chomp($line);
    }
    # now we've hit codas; can skip this line and move on to the next
    $line = <INV_FILE>;
    chomp($line);
    $i = 0;
    print "\nCodas: ";
    while ($line ne "") {
	$codas[$i] = $line;
	$i++;
#	print " $line";
	$line = <INV_FILE>;
	chomp($line);
    }
    print "\nDone reading in inventory\n";
}
else {
    print "Warning: inventory file does not appear to be valid (Must begin with onsets)\n";
}

# now generate random syllables
$current_words = 0;
while ($current_words < $target_words) {
    # start with the first onset + nucleus
    $candidate_word = "";
    $candidate_word .= @onsets[round(rand @onsets - 1)];
    $candidate_word .= @nuclei[round(rand @nuclei - 1)];
    $candidate_word .= @codas[round(rand @nuclei - 1)];

    # make monosyllables most likely; go on to another syllable with only 1/3 probability
    if (rand > .66 and $max_syllables >= 2){
	$candidate_word .= @onsets[round(rand @onsets - 1)];
	$candidate_word .= @nuclei[round(rand @nuclei - 1)];
	$candidate_word .= @codas[round(rand @nuclei - 1)];
	if (rand > .66 and $max_syllables >= 3){
	    $candidate_word .= @onsets[round(rand @onsets - 1)];
	    $candidate_word .= @nuclei[round(rand @nuclei - 1)];
	    $candidate_word .= @codas[round(rand @nuclei - 1)];
	    if (rand > .66 and $max_syllables >= 4){
		$candidate_word .= @onsets[round(rand @onsets - 1)];
		$candidate_word .= @nuclei[round(rand @nuclei - 1)];	    
		$candidate_word .= @codas[round(rand @nuclei - 1)];
	    }			
	}	
    }
      
#    print "$candidate_word\n";
    if ($known_words{$candidate_word} eq "") {
	# a new word
	$current_words++;
	$known_words{$candidate_word} = 1;		
    } else {
	$known_words{$candidate_word}++;
#	print "Duplicate word: $candidate_word (time $known_words{$candidate_word})\n";		
    }
    
}

# now have generated the target number of words
foreach $word (keys %known_words) {
    printf OUTPUTFILE "$word\t$known_words{$word}\n";        
}

print "Done generating novel words\n";
close OUTPUTFILE;