#!/usr/bin/perl

# general use directives
use strict;
use warnings;

# project specific use directives
# this comes with the standard ActiveState
# distribution.  You can also look for
# a newer version with PPM
use Win32::OLE;

my $wd;

# get the document
# use the full path
eval 
{
  $wd = Win32::OLE->GetObject('C:/pathto/document/foo.doc') 
};

die "Unable to load document\n" if $@;

# all of the Word document data members I'm using
# are explained in the MSDN documentation of the 
# external interfaces of a Word Document.
# if you have MSDN, search for "Word OLE".

# get the number of paragraphs
my $paraCount = $wd->{Paragraphs}->Count;

# set the counter
my $foo = 0;
my @words;

while ($foo++ < $paraCount)
{
    push @words, split /\s/, $wd->{Paragraphs}{$foo}{Range}{Text};
}

#clean up at the end
undef $wd;
[download]

After massaging your code a little I was able to see that only some of it pertains to the question at hand. Why not start from scratch? If you have to search for files, use File::Find to get files. Then look at the file extension (because that's how Windows tells files apart) to figure out what to do.

Unless I've misunderstood what your objective is.

"Excellent. Release the hounds." -- Monty Burns.

this is the part to index a pdf file

# Checks if a file is PDF depending on the filename. If so, write it t
+o a
# temporary file and feed it to $PDFTOTEXT, return the output. If it's
+ not
# PDF, return the buffer unmodified.
sub parse_pdf {
  my $buffer = $_[0];
  my $url = $_[1];
  if ($url =~ m/\.pdf$/i && $PDFTOTEXT) {
    my $tmpfile = "$TMP_DIR/temp.pdf";
    # Saving to a temporary file is necessary for http requested PDFs.
+ To
    # keeps things simpler, we also do it for local files from disk.
    open(TMPFILE, ">$tmpfile") or warn "Cannot write '$tmpfile': $!";
    binmode(TMPFILE);
    print TMPFILE ${$buffer};
    close(TMPFILE);
    # filename security check is done in to_be_ignored():
    ${$buffer} = `$PDFTOTEXT "$tmpfile" -` or (warn "Cannot execute '$
+PDFTOTEXT $tmpfile -': $!" and return undef);
    unlink $tmpfile or warn "Cannot remove '$tmpfile: $!'"
  }
}

# Save a term's ID to the database, if it does not yet exist. Return t
+he ID.
sub record_term {
  my $term = $_[0];
  print STDERR "Warning: record_term($term): No term was supplied\n" u
+nless $term;
  if ($terms_db{$term}) {
    return $terms_db{$term};
  } else {
    ++$TN;
    $terms_db{$term} = $TN;
    return $TN;
  }
}

# Is the file listed in @no_index or is it a PDF file with illegal cha
+racters
# in the filename?
# Supported ways to list a file in conf/no_index:
# /home/www/test/index.html (absolute path)
# /test/index.html (path relative to webroot, but with slash)
# test/index.html (path relative to webroot, no slash)
# http://localhost/test/index.html (absolute URL)
sub to_be_ignored {
  my $file = shift;
  # Check @no_index:
  my $file_relative;
  $file_relative = cut_document_root($file);
  foreach my $regexp (@no_index) {
    if( $file_relative =~ m/^\/?$regexp$/ || $file =~ m/^$regexp$/ ) {
      return "listed in no_index.txt";
    }
  }
  # For PDF files check filename for security reasons (it later gets h
+anded to a shell!):
  if( $file =~ m/\.pdf$/i && $PDFTOTEXT ) {
    if( $file !~ m/^[\/\\a-zA-Z0-9_.:+-]*$/ || $file =~ m/\.\./ ) {
      return "Ignoring '$file': illegal characters in filename";
    }
  }
  return undef;
}
[download]

I search some like this but for .doc .ppt and .xls file...

jcwren - 2001/12/20 22:18:00 UTC - added code tags