#!/usr/pubsw/bin/perl

# extractbody ELEMENT
# where element has the real text in an SGML/XML(-like) file, assuming that
# the ELEMENT start and end tags are on a line by themselves
# (this works for a lot of the LDC newswire corpora).
# By default, we print the element, so that articles are delimited.
# The -n flag means not to print the element tags.
# Christopher Manning 1999-2001

$extracting = 0;
$printbodytag = 1;

if ($#ARGV < 0) {
  die "usage: extractbody [-n] bodytag\n";
}
$word = shift(@ARGV);
while ($word =~ /^-[a-z]/) {
  if ($word eq "-n") {	# don't print tag body
    $printbodytag = 0;
  } else {
    print "Unrecognized option: $word\n";
  }
  $word = shift(@ARGV);
} 

while (<>) {
  if (/<$word>/) {
    $extracting = 1;
    if ($printbodytag) {
      print;
    }
  } elsif (/<\/$word>/) {
    $extracting = 0;
    if ($printbodytag) {
      print;
    }
  } elsif ($extracting) {
    print;
  }
}