It's by no means fully functional - it only scans one directory deep, and has trouble with title and description tags that span more than one line, so if anyone can offer any insight into how to improve it it'd be most welcome...
#!/usr/bin/perl
### Configuration Options ###
$root_directory="/home/www/html/"; # server path to web site
$output_file="/home/www/html/sitemap.html";
$root_url="http://www.mydomain.com";
# next, a pipe (¦) seperated list of sub directories to ignore
# while creating the sitemap.
# (the '..' directory is the unix 'up a level' directory, we
# don't want to include that..)
$excluded_dirs="..¦images¦bk2¦mailing_list¦includes¦foo";
### End configuration ###
# open a file to write the site map to
open (OF,">$output_file");
# print the HTML header
&header;
# scan one level deep for sub directories
opendir (DIR,"$root_directory");
@sub_dirlist=grep { -d "$root_directory/$_" } readdir(DIR);
closedir (DIR);
# iterate through the list of directories, calling the
# parse_dir subroutine for each one
foreach $sub_dir (@sub_dirlist) {
unless ($excluded_dirs=~/(^¦\¦)$sub_dir(\¦¦$)/) {
&parse_dir($root_url,$root_directory,$sub_dir);}
}
#print out the HTML footer, close output file and quit.
&footer;
close(OF);
exit;
sub parse_dir {
my ($root_url,$root_directory,$sub_dir)=@_;
# The next 3 lines just put the directory name into a
# format suitable for display - could be customised to
# include your own html
if ($sub_dir eq ".") {$sub_dir_formatted="/"}
else {$sub_dir_formatted="/$sub_dir/"}
print OF "<h2>$sub_dir_formatted</h2>\n";
# get a list of all files in this directory
opendir (DIR,"$root_directory/$sub_dir");
@dirlist=readdir(DIR);
closedir (DIR);
foreach $i (@dirlist) {
$title="";
$desc="";
# check if file is html
if ($i=~/\.html$¦\.htm$/) {
# if so, open it and parse for title and description
open (FP, "$root_directory/$sub_dir/$i");
while (!eof(FP)) {
chomp($line=<FP>);
if ($line=~/^.*<title>.*<\/title>.*$/i) {
($title = $line) =~ s/^.*<title>¦<\/title>.*$//gi;}
if ($line=~/^.*<meta name="description" content=".*">$/i) {
($desc = $line) =~ s/^<meta name="description" content="¦">$//gi;
}
}
# print the title, description to output file if we've found them.
# again, your own html could be substituted here
if ($title) {
print OF "<a href=\"$root_url$sub_dir_formatted$i\">$title</a><br>\n";}
else {
print OF "<a href=\"$root_url$sub_dir_formatted$i\">$i</a><br>\n";}
if ($desc){
print OF " $desc<br><br>\n" }
else {print OF "<br>\n";}
close(FP);
}
}
}
sub header {
print OF qq^
<html><head>
<title>foo</title>
<meta name="description" content="bar">
</head><body>
^;
}
sub footer {
print OF qq^
</body></html>
^;
}
1. They always list the URL and nothing else. Sounds like you've got that covered! :)
2. The output is UGLY. If you could put in something in there that allows the user to create an output template, that would be fantastic!
Picky, ain't I? :)
Fair point - I kept it simple and ugly. I sort of had CSS in mind ;)
(added: a simple output is also easier to integrate into framesets / table structured pages etc)
It wouldn't be hard to put a template in there though... will have a think tomorrow
> Picky, ain't I? :)
Ahem...
Very true... so one could just dump the plain-jane output into an HTML page via an SSI call, and then take care of the fonts & other details via CSS? *sigh* There you scripters go... putting all the *work* off on us designers! ;)
Yeah, although the template feature sort of exists already.
You can edit the html in the &header and &footer subroutines to include prettification stuff in the body tag, link to a stylesheet, set up a table structure or whatever. Or remove the <html><head> etc tags completely and pull the file in by SSI as you say.
The directory names, links and descriptions could be handled either with CSS or by editing the html directly in the lines that print them out.
> putting all the *work* off on us designers!
Ack! My cunning ploy discovered ;)
Can't you use HTML::TokeParser? - I think this would solve your problem with title and metas.
Now that looks useful! I've never come across it before. I'd considered using the vanilla HTML::Parser, but this is a very nice interface to it and will definitely find a place in my toolbox :)
Thanks
URL here [theoryx5.uwinnipeg.ca]
I wrote a script which scanned through all the html/shtml files in the site and compared the linked graphics with those in the images directory. It then made a web page listing (and showing) all these orphaned graphics. Useful to me as I hand code HTML these days (not sure if those fancy graphical editors do this kind of thing themselves or not), if anyone else is interested I could dig it out of my laptop and post it here?
