#!/usr/bin/perl

open EN_OUT, '>', "en_output.txt";
open DE_OUT,  '>', "de_output.txt";

%En = ();
%De = ();

##########

sub new_unit {
  if (/<(CHAPTER|SPEAKER|P>)/) { return 1; }
  else { return 0; }
}

##########

# 1) Processing the English corpus

# If you've mapped a network drive, use this line:
chdir "/Volumes/Corpora/multilingual/europarl/aligned/en-de/en/";
# If you've logged in to jones, comment out the previous line and
# uncomment this line:
#chdir "/Volumes/Data/Corpora/multilingual/europarl/aligned/en-de/en/";

##########

@all_files = <*>;

# loop over all files in the directory
foreach $filename (@all_files) {
  # for time reasons, we'll only look at a subset of the corpora
  if ($filename =~ /ep-00/) {
    # open each file, one at a time
    print "Now processing: $filename (English)\n";
    open FILE, '<', $filename;

    # initialize variables
    $unitnum = 1;
    $found_one = 0;
    $previous_lines = "";

    # read each line in the file
    while (<FILE>) {
      # new_unit indicates that we're starting a new unit of text
      if (&new_unit($_)) {

	# The criterion for adding to the En hash is that we've found
	# the word we're looking for
	if ($found_one) {
	  # make the original word more noticeable
	  $previous_lines =~ s{\b(one)\b}{<ONE>$1</ONE>}ig;

	  # store this position for later
	  $key = $filename.'='.$unitnum;
	  $En{$key} = $previous_lines;
	}

	# re-initialize variables
	$previous_lines = $_; # note that we include the structural mark-up
	$found_one = 0;
	$unitnum++;
      }
      # Store all the previous lines
      else {
	$previous_lines .= $_;
      }
      
      # Test whether 'one' appears in this unit: if so, set $found_one
      # to be some non-zero value
      if (/\bone\b/i) {
	$found_one = $unitnum;
      }
    }
    close FILE;
  }
}

##########

# 2) Processing the German corpus

# If you've mapped a network drive, use this line:
chdir "/Volumes/Corpora/multilingual/europarl/aligned/en-de/de/";
# If you've logged in to jones, comment out the previous line and
# uncomment this line:
#chdir "/Volumes/Data/Corpora/multilingual/europarl/aligned/en-de/de/";

##########

@all_files = <*>;

# loop over all files in the directory
foreach $filename (@all_files) {
  # for time reasons, we'll only look at a subset of the corpora
  if ($filename =~ /ep-00/) {
    # open each file, one at a time
    print "Now processing: $filename (German)\n";
    open FILE, '<', $filename;

    # initialize variables
    $unitnum = 1;
    $previous_lines = $_;

    # read each line in the file
    while (<FILE>) {
      # new_unit indicates that we're starting a new unit of text
      if (&new_unit($_)) {
	$key = $filename.'='.$unitnum;
	# The criterion for adding to the De hash is that the
	# corresponding English unit had the word we're looking for
	if (exists $En{$key}) {
	  $De{$key} = $previous_lines;
	}

	# re-initialize variables
	$previous_lines = $_; # note that we include the structural mark-up
	$unitnum++;
      }
      # Store all the previous lines
      else {
	$previous_lines .= $_;
      }
    }
    close FILE;
  }
}

@en_keys = keys %En;
foreach $key (@en_keys) {
  if (exists $De{$key}) {
    select EN_OUT;
    print "$key\n$En{$key}\n";

    select DE_OUT;
    print "$key\n$De{$key}\n";
  }
}

close EN_OUT;
close DE_OUT;