Graphics Tree Walker

Problem: you have a web site, possibly an internal web site, where you provide graphics and other goodies for web page developers. You have a lot of them, and preparing a catalog of them would be tedious and error-prone. It would also require maintenance every time you added a new graphic. Yucch.

A better solution is to have directories on your web site for the graphics, and build the catalog on the fly. That is what this Perl script does.

The user starts at the virtual root directory, and walks through the directory structure with this script. At each directory, all the graphics in the directory are shown. Also, any subdirectories are show, so that the user may enter them as well.

To use it, set a link to it on an appropriate page, such as:

<li>Walk the directory tree and display all the graphics in each directory with this <a href="/cgi-bin/graphics.pl?">graphics tree program</a>.

You can also set the "root" directory by specifying an initial target directory:

<li>Walk the directory tree and display all the graphics in each directory with this <a href="/cgi-bin/graphics.pl?graphics%2estuff">graphics tree program</a>.

This script runs on Perl 5, although I don't think there is anything to prevent it from running on Perl 4. It runs on Linux running Apache, but it should run on almost any Unix based web server. I currently run it on Redhat 5.2 and Mandrake 6.1. See the notes below for some possible local dependencies.

This program shows the use of regular expressions to scan input and to do text substitution, and it also shows how to walk a data file.

#! /usr/bin/perl

# Time-stamp: <2000-07-21 16:19:04 ccurley>

# graphics.pl: A program to do two things. First, show any .gifs, .pngs or
# .jpegs in the given directory. Second, to show links to lower
# directories, so the user can explore them.

# The intent is to have a live file system and build a catalog in real
# time. This will eliminate the need to manually edit catalog web pages
# for graphics collections.

# In order to print a directory, we have to handle greater thans, less
# thans and certain other characters correctly.

# call it from a web page like this: <a
# href="/cgi-bin/graphics.pl?">graphics tree program</a> to start at the
# virtual root.

# some debug statements are included, but commented out. You may find them
# useful when you install this script.

# version 0.2 Added a regex to filter for user directories. Very site
# specific, of course. 31 December 1997

# 1998 11 15 C^2: added a detector for links, which will follow links to
# directories.

#
# 2000/06/27 DTG (David T-G, http://www.bigfoot.com/~davidtg/): made some changes
#   started conversion to CGI.pm
#   case-ignored the hex unpack
#   case-ignored the jpeg-jpg-gif match
#   converted script /path/name to variable
#   enabled handling of embedded-spaces filenames (should use opendir, tho)
#

# 2000 07 03 C^2: added support for .pngs.

#       Cleaned up (?) the space handling for embedded spaces in file
#       names works. Directories with embedded spaces don't work. Yet.

#	Converted to use openfile, et al.

#	We now print all the directories first, then do the graphics. It's
#	easier to navigate a tree that way.

#	Changed some of the HTML to be a bit more standard and easier on the eye.

# 2000 07 20 C^2: added a filter to be sure a file actually is a graphics
# file before displaying it. N.B: This is probably somewhat system
# dependent, as it uses Unix/Linux' "file" program, and greps on the
# output. The grep option --silent is a gnu style, so gnu grep is assumed.

# Also, got directories and file names with spaces working correctly.

# Added a count of the graphics displayed so I could check it against the
# number of graphics in the directory, to be sure I got them all. You may
# find it useful also.




# To do: 1) convert completely to CGI module so we only need one
# library.


$| = 1;				# force flushes after every write to output.
require "cgi-lib.pl";		# from: http://www.bio.cam.ac.uk/cgi-lib/
use CGI;			# from CPAN mirrors everywhere
use File::Basename;		# For getting the file name from a fully qualified path

# what is this script? Very site-specific.
#$thisscript = "graphics.pl";
$thisscript = basename ($0);

# print the header.

my $cgi = new CGI; print ($cgi->header(-type=>'text/html'));
print "<html>\n<head><title>Graphics Tree Walker</title></head>
<body bgcolor=\"#ffffff\" text=\"#000000\" link=\"#0000ff\" vlink=\"#ff0000\"
alink=\"#00ff00\">\n";

# Now tell the user what we have here.

print"<center><h1>Graphics Tree Walker</h1></center>\n";

print"This is a script to display all graphics in a directory.<p>\n";

print"Use it to walk the directory structure of this web page, and view
every graphic (.jpeg, .png, or .gif) in each directory. Use the
\"Back\" button on your bowser to return, and click on directories as you
wish to descend into the tree.<p>";

print"Files are sorted alphabetically in ASCII sort order.<p>\n";

# start printing debug data
#  print"<pre>\n";

# get the argument, the string that comes after the file name in the
# url. This will be a path offset from the virtual root.

$query = $ENV{'QUERY_STRING'};

# print "Length of \$query is " . length($query) . "\n";
# print "Raw query is $query<p>\n";

# now strip out the hex values (%2f, etc.) and replace them with the ASCII
# characters they indicate.

$query =~ s//pack ("C", hex($1))/ieg;

# we need this variable so we can handle directories with spaces in the
# names. Dang Windows idiots.

$spacecleanquery = $query;
$spacecleanquery =~ s/ /\%20/g;

# print "filtered query is $query\n";

# use some of cgi-lib.pl's routines to debug.

# $result = &ReadParse;
# print("Result of ReadParse is $result\n");

# print "Environment: " . &PrintEnv;

# print "Variables: " . &PrintVariables;

# get the base path of the document directory in.
$path = $ENV{'DOCUMENT_ROOT'};

# now get the command line into the path, so we know where we are.
$path .= $query . "/";

# print("The real path before filtering is: $path\n");


# filter out user directories and provide the correct path. This is *very*
# site specific! This should take a path like
# \home\httpd\html\~username\foo.bar and produce
# \home\username\public_html\foo.bar

$path =~ s/httpd\/html\/\~(\w*)\/(.*)/\1\/public_html\/\2/g;

#  print("The real path after filtering is: $path\n");

#  print("The virtual path is: $query/\n");

# stop printing debug data.
#  print "</pre>\n\n";

opendir (DIRHANDLE, $path) or &CgiDie ("Couldn't open $path<p>\n");

while ( defined ($fname = readdir (DIRHANDLE))) {
    @fnames = (@fnames, $fname);
}

@sorted = sort (@fnames);


# now process directories.

$directories_seen = 0;

foreach $fname (@sorted) {

#      print ("Inside $path is $fname.<p>");

    $fqfname = $path . $fname;
#      print ("$fqfname<br>");

    # Now check for directories
    if(-d $fqfname && $fname ne "." && $fname ne ".." ){

	# If we got here, we have at least one directory. Print out our
	# header, then disable the message.

	if (!$directories_seen) {
	    print ("See the graphics in these Directories: <ul>\n");
	    $directories_seen = 1;
	}

	# figure out the fully qualified path name

#  	print ("Showing $fname now:<br>");
	$url = $fname;
	$url =~ s/([\W_])/sprintf("%%%x", ord($1))/eg;

#      print (" URL is $url.<br>");

	print "<li> <a href=\"$thisscript?$ENV{'QUERY_STRING'}%2f$url\">$fname</a>\n";
    }
}


if ($directories_seen) {
    print ("</ul><p>");
}

# Now images

$ImagesDisplayed = 0;

foreach $fname (@sorted) {

#      print ("Inside $path is $fname.<p>");

    $fqfname = $path . $fname;
#    print ("$fqfname<br>");

    # check for gifs, pngs or jpegs, and show them to the user as we find them.
    if($fname =~/.jpeg$|.jpg$|.gif$|.png$/i &&
       -f $fqfname &&

       # Warning: quite likely system specific! "which file" and "which
       # grep" are your friends.

       !(system ("/usr/bin/file \"$fqfname\" \| /bin/grep --silent \"image data\""))
       ) {

	# figure out the fully qualified path name
	$url = $fname;
	$url =~ s/([\W_])/sprintf("%%%x", ord($1))/eg;

#      print (" URL is $url.<br>");

	print "<img border=1 src=\"$spacecleanquery/$url\" alt=\"$query/$fname\"><br>\n";
	print "The file name for this graphic is: $query/$fname<p><p>";

	$ImagesDisplayed++;
    }
}

if ($ImagesDisplayed) {
    if ($ImagesDisplayed == 1) {
	print ("One graphic is displayed.<p>\n");
    } else {
	print ("$ImagesDisplayed graphics are displayed.<p>\n");
    }
}

print"\nThis is the end of the script.";
print"</body></html>\n";

| << | < | > | >> | Welcome | Software | Communications | Classes | Resume | Sample Code | Thomas Jefferson: Patron Saint of the Internet | Yum Repository Notes | NFS and Firewalls on Fedora Core | Netiquette | NT Emacs Installation | My .emacs File | Notes on OpenSSH | Bare Metal Recovery | Fn | Rms | Dump | Register | Atexit | Graphics Tree Walker | which.nvidia | buildiso | wallpaper2 | gps | Fedora Gpsdrive RPMs | Single Source Frames | Notes | A Bug Notification | Helpful Little Paperclip | Linux on Lenovo R51 | Wyoming Travel

Copyright © 1996 through 2008 by Charles Curley Last Modified: 24 Feb, 2008