#!/usr/bin/env perl use strict; $|++; use LWP::Simple; -d "RESULTS" or mkdir "RESULTS", 0755 or die "cannot mkdir RESULTS: $!"; my $all_model_index = get "http://sportsillustrated.cnn.com/features/2007_swimsuit/models/"; while ($all_model_index =~ /(\/features\/2007_swimsuit\/(models|painting|onlocation)\/[-\w]+\/)/g) { doit("$1"); } doit("/features/2007_swimsuit/beyonce/"); doit("/features/2007_swimsuit/3d/"); sub doit { my $base = shift; print "$base =>\n"; my $model_index = get "http://sportsillustrated.cnn.com/$base/index2.html"; unless ($model_index) { $model_index = get "http://sportsillustrated.cnn.com/$base/"; } while ($model_index =~ /\"(http:\/\/i.a.cnn.net\/si\/features\/2007_swimsuit\/images\/photos\/)([\w.\-]+)t.jpg\"/g) { my $url = "$1$2.jpg"; my $file = "RESULTS/$2.jpg"; if (-e $file) { print "$url => $file: "; print "skip\n"; } else { print "$url => $file: "; print mirror($url, $file), "\n"; } } }