Statistician in my garbage...

[larsen]

Mainly it's a HTML::Parser exercise, done during a heavy research moment :)

The idea is very simple: randomly putting together texts and images from your browser's cache, you could get a snapshot of your (or someone else's) behaviours. Like a statistician in your garbage :)

I grabbed the idea from an old DOMUS issue, but I've lost the original URL.

Update: I'm thinking to use it as a permanent installation in an Internet Cafe'. Two monitors: one used by surfer, and one connected to a box that automatically refresh a page generated by this script.

#!/usr/bin/perl

use strict;


# Digs in your browser's cache 
# like a statistician in your trashcan...

package Lurker;

use File::Find;

my $cache = {
    IMAGES => [],
    DOCS => [],
};

sub lurk
{
    my $dir = shift;
    
    print STDERR "Reading cache...";

    find(
     sub 
     {   
         for ( $File::Find::name ) {
         /\.gif$/  ||
         /\.png$/  || 
         /\.jpg$/  &&    push @{ $cache->{ IMAGES }}, $_;
         /\.html$/ &&    push @{ $cache->{ DOCS }}, $_;
         }
     }, $dir 
    );

    print STDERR "OK!\n";
}

sub pick_random
{
    my $what = shift; 
    
    my $n = scalar( @{$cache->{ $what }} );
    
    return ${$cache->{ $what }}[ rand $n ];
}




package My_HTML_Parser;

use base 'HTML::Parser';

sub start
{
    my $self = shift;
    my ($tag, $attr, $attrseq, $origtext) = @_;
    
    my ($orig_src, $new_src);
    
    if ($tag eq 'img') {
    $orig_src = $attr->{'src'};        
    $new_src = Lurker::pick_random( 'IMAGES' );
    $origtext =~ s/$orig_src/$new_src/;
    }
    print $origtext;
}

sub text
{
    my $self = shift;
    my ($text) = @_;
    
    print $text;
}

sub end
{
    my $self = shift;
    my ($tag) = @_;
    
    print "</$tag>";
}



package main;

my $cache_directory = '/home/stefano/.netscape/cache';

Lurker::lurk( $cache_directory );

my $doc = Lurker::pick_random('DOCS');

print STDERR "Now parsing $doc...\n";

my $a = new My_HTML_Parser;
$a->parse_file( $doc );

[merlyn]

You're working too hard on that HTML::Parser part:

use HTML::Parser;
use HTML::Entities;
HTML::Parser->new( 
                  default_h => 
                  [sub { 
                     print shift; 
                   }, "text"], 
                  start_h => 
                  [sub { 
                     local ($text, $tag,$attr) = shift; 
                     if ($tag eq "img") {
                       $attr->{src} = Lurker::pick_random( 'IMAGES' );
                       $text = "<$tag";
                       $text .= " $_=\"" . encode_entities($attr->{$_}
+). "\""
                          for keys %$attr;
                       $text .= ">";
                     }
                     print $text;
                   }, "text,tag,attr"], 
                 )->parse_file($doc);

-- Randal L. Schwartz, Perl hacker