BEGIN {
    my $b__dir = (-d '/home/jockmyfr/perl'?'/home/jockmyfr/perl':( getpwuid($>) )[7].'/perl');
    unshift @INC,$b__dir.'5/lib/perl5',$b__dir.'5/lib/perl5/x86_64-linux',map { $b__dir . $_ } @INC;
}
#!usr/perl/bin;

use strict;					#Perl module to restrict unsafe constructs 
use LWP::UserAgent;			#Web user agent class
use URI::URL;				#Uniform Resource Locators
use HTML::Entities;			#Encode or decode strings with HTML entities
use URI::Escape;
use LWP::Simple;
use HTTP::Cookies;
use Cwd;
use DBI;

print "\n\n**********************************************************";
print "\n\n*    DATA EXTRACTION  : http://www.amazon.com/           *";
print "\n\n*                         - GOJAJ Intelligent Design     *";    
print "\n\n*                         - Developed by John Akhile Jr  *";    
print "\n\n**********************************************************\n\n";


#modify your table name here.
my $dbname="kixcomec_scrape";
my $server="localhost";
my $uname="kixcomec_scrape";
my $pwd='3.2Billion$';


mkdir "images_viviennewestwood", 0777 unless -d "images_viviennewestwood";

my $ua=LWP::UserAgent->new;
$ua->agent("Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.0.10) Gecko/2009042316 Firefox/3.0.10");
#$ua->timeout;
$ua->max_redirect('1');

my $cookie_jar = HTTP::Cookies->new(file=>$0."_cookie.txt",autosave => 1,);
$ua->cookie_jar($cookie_jar);


my $dsn="DBI:mysql:$dbname:$server";
my $dbh;
if(!($dbh=DBI->connect($dsn,"$uname","$pwd",{AutoCommit => 1})))
{
	print"\n\nDatabase Connection Failure\n";
	print "try again...\n\n";
	exit;
}
$dbh-> {'LongTruncOk'} = 1;
$dbh-> {'LongReadLen'} = 90000;
$dbh->do("set character set utf8");


my $home_url="http://www.amazon.com/";
&get_cont($home_url,$home_url,'GET');
my $start_url="http://www.amazon.com/s/ref=brand_abn_shoes?ie=UTF8&bbn=672123011&rh=n%3A672123011%2Cp_4%3AVivienne%20Westwood&rnid=672123011";
my $dir = getcwd;
my $count=1;
my $content=&get_cont($start_url,$home_url,'GET');

while($content=~m/<a[^>]*?href\=\"([^>]+?)\"[^>]*>\s*<span\s*class\=\"refinementLink\">\s*([W|M|J|G|B][^>]+?)\s*</igs)
{
	my $gender_url=$1;
	my $gender=$2;
	my $cat=$gender;
	decode_entities($gender);
	$gender_url=URI::URL->new_abs($gender_url,$home_url);
	print "Gender  ::  $gender\n";
	$gender=~s/\'/\'\'/igs;
	my $gender_content=&get_cont($gender_url,$home_url,'GET');
	if($gender_content=~m/<strong>$cat<\/strong>([\w\W]+?)<\/ul>/is)
	{
		my $block_content=$1;
		while($block_content=~m/<a[^>]*?href\=\"([^>]+?)\"[^>]*>\s*<span\s*class\=\"refinementLink\">\s*([^>]+?)\s*</igs)
		{
			my $category_url=$1;
			my $product_category=decode_entities($2);
			$category_url=URI::URL->new_abs($category_url,$home_url);
			print "\tProduct Category  ::  $product_category\n";
			$product_category=~s/\'/\'\'/igs;
			my $category_content=&get_cont($category_url,$home_url,'GET');
			if($category_content=~m/<strong>[^>]*?<\/strong>\s*<\/li>\s*\s*(<li[^>]*?>[\w\W]+?)<\/ul>/is)
			{
				my $category_block=$1;
				while($category_block=~m/<a[^>]*?href\=\"([^>]+?)\"[^>]*>\s*<span\s*class\=\"refinementLink\">\s*([^>]+?)\s*</igs)
				{
					my $sub_category_url=$1;
					my $sub_category=$2;
					decode_entities($sub_category);
					$sub_category_url=URI::URL->new_abs($sub_category_url,$home_url);
					print "\t\tSub Category  ::  $sub_category\n\n";
					$sub_category=~s/\'/\'\'/igs;
					my $sub_category_content=&get_cont($sub_category_url,$home_url,'GET');
					&get_data($sub_category_content,$gender,$product_category,"Shoes>$gender>$product_category>$sub_category",$sub_category);
				}
			}
			else
			{
				&get_data($category_content,$gender,$product_category,"Shoes>$gender>$product_category");
			}
		}
	}
}

sub get_data
{
	my($product_content,$gender,$product_category,$prod_tag,$sub_category)=@_;
	my($product_name,$product_code,$model,$product_price,$description,$color,$image_file,$image_name,$prod_size,$size_name,$color_name,$prod_color);
	nextpage:
	if($product_content=~m/<div\s*id\=\"atfResults\"[^>]*>([\w\W]+?)<div\s*id\=\"centerBelowStatic\">/is)
	{
		my $block=$1;
		while($block=~m/<div\s*class\=\"image\">\s*<a[^>]*?href\=\"([^>]+?)\"[^>]*>\s*<img[^>]*>/igs)
		{
			my $source_url=$1;
			$source_url=URI::URL->new_abs($source_url,$home_url);
			my $final_content=&get_cont($source_url,$home_url,'GET');
			if($final_content=~m/<h1[^>]*>\s*<[^>]*>\s*([^>]+?)\s*</is)
			{
				$product_name=decode_entities($1);
				print "\n\n$count  ::  Product Name  ::  $product_name\n";
				$product_name=~s/\'/\'\'/igs;
			}
			if($final_content=~m/<input\s*type\=\"hidden\"\s*id\=\"ASIN\"\s*name\=\"ASIN\"\s*value\=\"([^>]+?)\"\s*\/>/is)
			{
				$product_code=$1;
				$product_code=~s/\'/\'\'/igs;
			}
			if($final_content=~m/>\s*Item\s*model\s*number\:\s*<[^>]*>\s*([^>]+?)\s*</is)
			{
				$model=decode_entities($1);
				print "Product Model  ::  $model\n";
				$model=~s/\'/\'\'/igs;
			}
			if($final_content=~m/<ul\s*style\=\"list\-style\-type\:\s*disc\;\"\s*>\s*([\w\W]+?)\s*<\/ul>/is)
			{
				$description=$1;
				$description=~s/<[^>]*?>//igs;
				$description=~s/\s+/ /igs;
				$description=~s/^\s+|\s+$//igs;
				$description=~s/\'/\'\'/igs;
				print "Product Description  ::  $description\n";
			}
			if($final_content=~m/\'color_name\'\s*\:\s*\[\s*([\w\W]+?)\s*\]/is)
			{
				my $color_block=$1;
				while($color_block=~m/\'\s*([^>]+?)\s*\'/igs)
				{
					$color=$1;
					print "Color  ::  $color\n";
					$color=~s/\*+//igs;
					if($final_content=~m/\"$color\"\:\[\{([\w\W]+?)\}\]/is)
					{
						my $image_block=$1;
						my $col=1;
						$image_file="";
						$image_name="";
						while($image_block=~m/\"main\"\:\[\"([^>]+?)\"/igs)
						{
							my $img_src=$1;
							if($img_src=~m/(?:.+\/)([^>]*?\.[a-z]+)/is)
							{
								my $img_fname=$1;
								getstore($img_src,"images_viviennewestwood/$img_fname");
								$img_fname=$dir."/images_viviennewestwood/$img_fname";
								$image_name=$image_name."Product_Image_filename_".$col.",";
								$img_fname=~s/\'/\'\'/igs;
								$image_file=$image_file."\'$img_fname\',";
								$col++;
							}
							undef($img_src);
							last if($col>5);
						}
						undef($image_block);
						$image_file=~s/\,$//igs;
						$image_name=~s/\,$//igs;
					}
					my $col=1;
					$prod_size=",";
					$size_name=",";
					my $asin_list="";
					$product_price=0;
					if($final_content=~m/<select[^>]*?name\=\"childVariationASIN\"[^>]*?>([\w\W]+?)<\/select>/is)
					{
						my $size_block=$1;
						while($size_block=~m/<option\s*value\=\"([^>]+?)\"\s*title\=\"\s*([^>]+?)\s*\s*\:\:\s*$color\"[^>]*?>/igs)
						{
							$asin_list=$asin_list.$1.",";
							my $size=decode_entities($2);
							$size=~s/\'/\'\'/igs;
							$prod_size=$prod_size."\'".$size."\',";
							$size_name=$size_name."Product_Available_Size_".$col.",";
							if($col%10==0)
							{
								$asin_list=~s/\,$//igs;
								my $price_content=&get_cont("http://www.amazon.com/gp/twister/ajax/prefetch/177-1721571-7723263?json=1&sid=177-1721571-7723263&rid=037TS7MMN2XD6J8BEV33&parentAsin=$product_code&qid=&sr=&asinList=$asin_list&productGroupID=shoes_display_on_website&merchantID=&PowerBar=0&pfWrapFeatures=0&rps=0&_=1324294703674",$home_url,'GET');
								while($price_content=~m/<b\s*class\=\"priceLarge\">\$([\d\.]+)\s*</igs)
								{
									my $price=$1;
									$product_price=$price if($price>=$product_price);
								}
								$asin_list="";
							}
							$col++;
							last if($col>20);
						}
						$asin_list=~s/\,$//igs;
						my $price_content=&get_cont("http://www.amazon.com/gp/twister/ajax/prefetch/177-1721571-7723263?json=1&sid=177-1721571-7723263&rid=037TS7MMN2XD6J8BEV33&parentAsin=$product_code&qid=&sr=&asinList=$asin_list&productGroupID=shoes_display_on_website&merchantID=&PowerBar=0&pfWrapFeatures=0&rps=0&_=1324294703674",$home_url,'GET');
						while($price_content=~m/<b\s*class\=\"priceLarge\">\$([\d\.]+)\s*</igs)
						{
							my $price=$1;
							$product_price=$price if($price>=$product_price);
						}
						print "Product Price  ::  $product_price\n";
					}
					$size_name=~s/\,$//igs;
					$prod_size=~s/\,$//igs;
					print "\n\n";
					my $query="insert into Amazon_Vivienne(Gender,Product_Category,Sub_Category,Source_Url,Product_Code,Product_Model,Product_Name,Color,Description,Product_Price,".$image_name."".$size_name.") Values(\'$gender\',\'$product_category\',\'$sub_category\',\'$source_url\',\'$product_code\',\'$model\',\'$product_name\',\'$color\',\'$description\',\'$product_price\',".$image_file."".$prod_size.")";
					$query=~s/\,+/\,/igs;
					&DBProcess($query);
					undef($query);
				}
				undef($color_block);
			}
			$count++;
		}
	}
	if($product_content=~m/<a[^>]*?href\=\"([^>]+?pg_([\d]+)[^>]+?)\"[^>]*?>Next[^>]*?</is)
	{
		my $next_url=$1;
		my $page_num=$2;
		print "\n\t\tGo to Page Number  $page_num...\n\n";
		$next_url=URI::URL->new_abs($next_url,$home_url);
		$product_content=&get_cont($next_url,$home_url,'GET');
		goto nextpage;
	}
}

sub get_cont
{
	my ($g_url,$home_url,$method,$content)=@_;
	my ($req,$res,$code);
	my $debug=0;
	start1:
	if($method eq 'GET')
	{
		$req= HTTP::Request->new(GET=>$g_url);
		$req->header("Content-Type" => "application/x-www-form-urlencoded");
		$req->header("Accept"=>"text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8");
		$res= $ua->request($req);
		$cookie_jar->extract_cookies($res);
		$cookie_jar->save;
		$cookie_jar->add_cookie_header($req);
		$code=$res->code();
	}
	elsif($method eq 'POST')
	{
		$req = HTTP::Request -> new(POST => "$g_url");
		$req->header("Content-Type" => "application/x-www-form-urlencoded");
		$req->header("Accept"=>"text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8");
		$req->content($content);
		$res= $ua->request($req);
		$cookie_jar->extract_cookies($res);
		$cookie_jar->save;
		$cookie_jar->add_cookie_header($req);
		$code=$res->code();
	}
	if($code=~m/40/is)
	{
		print "\nURL missing .Page not found\n";
		return;
	}
	elsif($code=~m/50/is)
	{
		$req=HTTP::Request->new(GET=>"http://www.cpan.org");
		$req->header("Content-Type"=> "application/x-www-form-urlencoded");
		$res=$ua->request($req);
		my $code1=$res->code();
	 	print "\nCODE1:$code1\n";
		if($code1=~m/50/is)
		{
			print "\nNet Failure";
			sleep(1000);
			goto start1;
		}
		elsif($code1=~m/20/is)
		{
			print "\nURL missing .Page not found\n" if $debug==1;
			return if $debug==1;
			$debug=1;
			sleep(50);
			goto start1;
		}
	}
	elsif($code=~m/30/is)
	{
		my $location=$res->header('location');
		my $loc_url=URI::URL->new_abs($location,$home_url);
		print "\nLOC :: $loc_url\n";
		&get_cont($loc_url,$home_url,'GET','');
	}
	elsif($code=~m/20/is)
	{
		my $g_con=$res->content;
		return $g_con;
	}
}

# Data Base Query Execute
sub DBProcess($)
{
    my $query=shift;
	my $sth=$dbh->prepare($query);
	if($sth->execute())
	{
		return $sth;
		$sth->finish();
	}
	else
	{
		my $err=$DBI::errstr;
		if($err=~m/duplicate/is)
		{
			return;
		}

		elsif($err=~m/syntax/is)
		{
			print "\nsyntax error\n";
			open(FH,">>Amazon_Logfile.txt");
			print FH "\n$query\;";
			close FH;
		}
		else
		{
			open(FH,">>Amazon_Logfile.txt");
			print FH "\n$query\;";
			close FH;
			Connect_Again:
			if($dbh=DBI->connect($dsn,"$uname","$pwd",{AutoCommit => 1}))
			{
				$dbh-> {'LongTruncOk'} = 1;
				$dbh-> {'LongReadLen'} = 1000000;
			}
			else
			{
				sleep 10;
				goto Connect_Again;

			}

		}

	}
}
-----------------------------82368347011361771591525345383
Content-Disposition: form-data; name="user_search"


-----------------------------82368347011361771591525345383
Content-Disposition: form-data; name="user_add"


-----------------------------82368347011361771591525345383
Content-Disposition: form-data; name="user_modify"


-----------------------------82368347011361771591525345383
Content-Disposition: form-data; name="user_delete"


-----------------------------82368347011361771591525345383
Content-Disposition: form-data; name="user"


-----------------------------82368347011361771591525345383
Content-Disposition: form-data; name="user_trans"


-----------------------------82368347011361771591525345383
Content-Disposition: form-data; name="authenticate"

yes
-----------------------------82368347011361771591525345383
Content-Disposition: form-data; name="message"


-----------------------------82368347011361771591525345383
Content-Disposition: form-data; name="file_name_convention"

Zy009876.jpg