#!/usr/local/bin/perl

#
#	ListHTMLTitles
#	©2001 Joshua S. Jacob
#
#	version:		1.0.1
#
#	written by:		Josh Jacob (josh.jacob@joshjacob.com)
#
#	description:	This Perl script takes as input a directory and will output
#					to a file a list of all HTML files and the titles in the
#					<title></title> tags.
#
#	license:		This software is provided "as is" and without any express or
#					implied warranties, including, without limitation, the implied
#					warranties of merchantability or fitness for a particular
#					purpose. The copyright holder shall not be held liable for
#					any damages whatsoever arising from the use of, misuse of, or
#					inability to use this software.
#

#
# constants
#

$gLogFileName = "ListHTMLTitles.log";

# list of file extenstions to look in
@gFileExtensions = (".html", ".htm", ".asp");

# list of directories to skip - regular expressions
@gSkipDirectories = ('^_(.*)$', '\.(.*)$');

$gRootDirectory = "";
$gDisplayFlat = 0;

#
# function: checkExtension()
#

sub checkExtension()
{
	# get inputs
	my $inPath = @_[0];
	my $inFileName = @_[1];
	my $inLevel = @_[2];
	
	my $extension;
	my $counter;
	my $fileString;
	
	# for each extension in list
	foreach $extension (@gFileExtensions)
	{
		# if the file matches an extension...
		if ("$inPath$inFileName" =~ /$extension$/i)
		{
			if ($gDisplayFlat == 1)
			{
				$_ = $inPath;
				s/$gRootDirectory//ig;
				print LOG "$_";
			}
			else
			{
				# tab in based on level
				for ($counter = 0; $counter < $inLevel; $counter++)
				{
					print LOG "\t";
				}
			}
			# print file name
			print LOG "$inFileName";
			if ($gDisplayFlat == 1)
			{
				print LOG "\t";
			}
			else
			{
				print LOG "\n";
				# tab in based on level
				for ($counter = -1; $counter < $inLevel; $counter++)
				{
					print LOG "\t";
				}
			}
			# open file
			open (HTMLFILE, $inPath . $inFileName);
			# get the whole file as a string
			while (<HTMLFILE>)
			{
				$fileString = $fileString . $_;
			}
			# close the file
			close HTMLFILE;
			# if the title pattern matches in the file...
			if ($fileString =~ /<title>(.*)<\/title>/is)
			{
				# get the matches title
				$_ = $1;
				# replace any newlines
				s/[\n\r\f]+/ /ig;
				# log the title
				print LOG "<title>$_</title>\n";
			}
			# ...else the page doesn't have a title
			else
			{
				# log no title
				print LOG "-- no title found --\n";
			}
		}
	}
}

#
# function: printDir()
#

sub printDir()
{
	# get inputs
	my $inDir = @_[0];
	my $inLevel = @_[1];
	
	my @myDir;
	my $myFile;
	my $counter;
	my $skipDir;

	# if the directory can be opened for reading
	if (opendir(DIR, $inDir))
	{
		# get all the files
		@myDir = readdir DIR;
		
		# close the directory
		close DIR;
		
		# for each file in the file list
		foreach $myFile (@myDir)
		{
			# if it's not an icon and not a hidden file...
			if (($myFile ne "Icon\n") && ($myFile !~ /^\./))
			{
				# if it's a directory...
				if (-d "$inDir$myFile")
				{
					$skipDir = 0;
					foreach $dir (@gSkipDirectories)
					{
						if ("$myFile" =~ /$dir/i)
						{
							$skipDir = 1;
						}
					}
					if ($skipDir == 0)
					{
						if ($gDisplayFlat != 1)
						{
							# tab in based on level
							for ($counter = 0; $counter < $inLevel; $counter++)
							{
								print LOG "\t";
							}
							# print file name
							print LOG "$myFile:\n";
						}
						# recurse
						&printDir("$inDir$myFile:", $inLevel + 1);
					}
				}
				# ...else it's a file
				else
				{
					# check the file 
					&checkExtension($inDir, $myFile, $inLevel);
				}
			}
		}
	}
}

#
# main
#

# open log file for writing
open (LOG, ">:$gLogFileName") || print "Can't write log file :(\n\n";

# for each item passed in
foreach $file (@ARGV) 
{
	# if it's not the icon file and not a hidden file...
	if (($file ne "Icon\n") && ($file !~ /^\./))
	{
		# if it's a directory...
		if (-d $file)
    	{
    		# log it as the root directory 
    		print LOG "Root directory -- $file\n\n";
    		
    		# store the root directory
    		$gRootDirectory = $file;
    		
    		# pass the directory to printDir with the initial level or zero
    		&printDir($file, 0);
		}
	}
}

# close log file
close LOG;