in reply to Swimsuits2008
in thread Swimsuits2004

This script is a bit messy, but it seems to be grabbing all of the core images. No video downloads yet... I'm not that smart. 
#!/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/2009_swims
+uit/models/";

for ($all_model_index =~ m{"/2009[^"]+/index2\.html"}g) {
  next if /(\/2009_swimsuit\/(models|dancers|tennis|onlocation)\/[-\w]
++\/)/;
  print "$_\n";
}
# exit 0;

while ($all_model_index =~ /(\/2009_swimsuit\/(models|dancers|tennis|o
+nlocation)\/[-\w]+\/)/g) {
  doit("$1");
}

doit("/2009_swimsuit/painting/$_/")
  for qw(brooklyn-decker
     julie-henderson
     irina-shayk
     jessica-white);

my %done;

sub doit {
  my $base = shift;
  return if $done{$base}++;

  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 =~ m{\"(http://i.cdn.turner.com/si/pr/subs/swims
+uit/images/)([\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";
    }
  }
}

[download]

-- Randal L. Schwartz, Perl hacker

Replies are listed 'Best First'.
Swimsuits2010
by merlyn (Sage) on Feb 09, 2010 at 18:54 UTC
    Updated for 2010: 
    #!/usr/bin/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/2010_swims
+uit/models/";

while ($all_model_index =~ m{"(/2010[^"]+/)index2\.html"}g) {
  doit("$1");
}
exit 0;

my %done;

sub doit {
  my $base = shift;
  return if $done{$base}++;

  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 =~ m{\"(http://i.cdn.turner.com/si/pr/subs/swims
+uit/images/)([\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";
    }
  }
}

    [download]

    -- Randal L. Schwartz, Perl hacker

    The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119.

[reply]
[d/l]
      And for 2011: 
      #!/web/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/2011_swims
+uit/models/";

while ($all_model_index =~ m{"(/2011[^"]+/)"}g) {
  doit("$1");
}
exit 0;

my %done;

sub doit {
  my $base = shift;
  return if $done{$base}++;

  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 =~ m{\"(http://i.cdn.turner.com/si/pr/subs/swims
+uit/images/)([\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";
    }
  }
}

      [download]

      -- Randal L. Schwartz, Perl hacker

      The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119.

[reply]
[d/l]
        And for 2012: 
        #!/web/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/2012_swims
+uit/models/";

while ($all_model_index =~ m{"(/2012[^"]+/)"}g) {
  doit("$1");
}

doit("/2012_swimsuit/painting/$_/")
  for qw(alex-morgan natalie-coughlin natalie-gulbis);

exit 0;

my %done;

sub doit {
  my $base = shift;
  return if $done{$base}++;

  print "$base =>\n";
  # (print "[TEMP SKIPPED]", "\n"), return unless $base =~ /paint/;
  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 =~ m{(\w+)\.html.*?\"(http://i.cdn.turner.com/si
+/pr/subs/swimsuit/images/.*?[\w.\-]+)_t.jpg\"}g) {
    my $url = "$2.jpg";
    my $file = "RESULTS/$1.jpg";
    if (-e $file) {
      print "$url => $file: ";
      print "skip\n";
    } else {
      print "$url => $file: ";
      print mirror($url, $file), "\n";
    }
  }
}

        [download]

        -- Randal L. Schwartz, Perl hacker

        The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119.

[reply]
[d/l]
        ... RESULTS/11_cintia-dicker_41.jpg ...

        [download]
        Hooray
[reply]
[d/l]

In Section Cool Uses for Perl