Google PageRank API. Определяем PR страницы своими силами.

#!/usr/local/bin/env python
# -*- coding: utf-8 -*-
# (C) 2008 Fred Cirera
# ported in Python from the Ruby code by Vsevolod S. Balashov
# fixed error with algorythm by Konstantin S. Dyakoff
# http://madjack.ru/developer/2009/10/google-pagerank-api.html
# http://snippets.dzone.com/posts/show/3284

import urllib2
import re
import time
import sys

from urllib import urlencode
from pprint import pprint

HOST = "toolbarqueries.google.com"

def mix(a, b, c): 
    M = lambda v: v % 0x100000000 # int32 modulo 
    a, b, c = (M(a), M(b), M(c))
    
    a = M((a-b-c) ^ (c >> 13))
    b = M((b-c-a) ^ (a <<  8))
    c = M((c-a-b) ^ (b >> 13))

    a = M((a-b-c) ^ (c >> 12))
    b = M((b-c-a) ^ (a << 16))
    c = M((c-a-b) ^ (b >> 5))

    a = M((a-b-c) ^ (c >>  3))
    b = M((b-c-a) ^ (a << 10))
    c = M((c-a-b) ^ (b >> 15))

    return a, b, c

def checksum(iurl):
    C2I = lambda s: sum(c << 8*i for i, c in enumerate(s[:4]))
    
    a, b, c = 0x9e3779b9, 0x9e3779b9, 0xe6359a60
    lg  = len(iurl)
    k = 0
    
    while k <= lg-12:
        a = a + C2I(iurl[k:k+4])
        b = b + C2I(iurl[k+4:k+8])
        c = c + C2I(iurl[k+8:k+12])
        a, b, c = mix(a, b, c)
        k += 12

    a = a + C2I(iurl[k:k+4])
    b = b + C2I(iurl[k+4:k+8])
    c = c + (C2I(iurl[k+8:])<<8) + lg
    a, b, c = mix(a, b, c)
    return c

def GoogleHash(value):
    I2C = lambda i: [i & 0xff, i >> 8 & 0xff,  i >> 16 & 0xff, i >> 24 & 0xff]
    ch = checksum([ord(c) for c in value])
    ch = ((ch % 0x0d) & 7) | ((ch/7) << 2)
    
    return "6%s" % checksum(sum((I2C(ch-9*i) for i in range(20)), []))


def make_url(host, site_url):
    url = "info:" + site_url
    ghash = GoogleHash(url)
    params = dict(client="navclient-auto", ch="%s" % ghash,
                  ie="UTF-8", oe="UTF-8", features="Rank", q=url)
    return "http://%s/search?%s" % (host, urlencode(params))


# Where the fun begins

if __name__ == "__main__":
    if len(sys.argv) != 2:
        url = 'http://www.google.com/'
    else:
        url = sys.argv[1]

    if not url.startswith('http://'):
        url = 'http://%s' % url

    print make_url(HOST, url)
    req = urllib2.Request(make_url(HOST, url))
    try:
        f = urllib2.urlopen(req)
        response = f.readline()
    except Exception, err:
        print err
        # print err.read()
        sys.exit(1)
        
    try:
        rank = re.match(r'^Rank_\d+:\d+:(\d+)', response.strip()).group(1)
    except AttributeError:
        print "This page is not ranked"
        rank = -1

    print "PagerRank: %d\tURL: %s" % (int(rank), url)

#!/usr/local/bin/perl
# (C) 2008 Fred Cirera
# ported in Perl from the Python code by Konstantin S. Dyakoff
# http://madjack.ru/developer/2009/10/google-pagerank-api.html

package PR;

use Acme::Lambda;
use URI::Escape;
use LWP::UserAgent;

sub mix{
    my $a = shift;
    my $b = shift;
    my $c = shift;
    my $M = lambda {$_ % 0x100000000};

    $a = $M->($a);
    $b = $M->($b);
    $c = $M->($c);
    
    $a = $M->(($a-$b-$c) ^ ($c >> 13));
    $b = $M->(($b-$c-$a) ^ ($a <<  8));
    $c = $M->(($c-$a-$b) ^ ($b >> 13));
    
    $a = $M->(($a-$b-$c) ^ ($c >> 12));
    $b = $M->(($b-$c-$a) ^ ($a << 16));
    $c = $M->(($c-$a-$b) ^ ($b >> 5));
    
    $a = $M->(($a-$b-$c) ^ ($c >>  3));
    $b = $M->(($b-$c-$a) ^ ($a << 10));
    $c = $M->(($c-$a-$b) ^ ($b >> 15));

    return ($a, $b, $c);
}

sub C2I{
    my $len = pop;
    my $from = pop;
    my @arr = @_;
    my $sum = 0;
    
    my $j = 0;
    for my $i ($from..$len){
	$sum += $arr[$i] << (8 * $j++);
    }
    return $sum;
}

sub checksum{
    my @iurl = @_;    
    my ($a, $b, $c) = (0x9e3779b9, 0x9e3779b9, 0xe6359a60);
    
    my $lg  = scalar(@iurl);
    my $k = 0;
    while($k <= $lg-12){
        $a = $a + C2I(@iurl, $k, $k+3);
        $b = $b + C2I(@iurl, $k+4, $k+7);
        $c = $c + C2I(@iurl, $k+8, $k+11);
        ($a, $b, $c) = mix($a, $b, $c);
        $k += 12;
    }
    
    $a = $a + C2I(@iurl, $k, $k+4);
    $b = $b + C2I(@iurl, $k+4, $k+8);
    $c = $c + (C2I(@iurl, $k+8, $k+scalar(@iurl))<<8) + $lg;
    ($a, $b, $c) = mix($a, $b, $c);
    return $c;
}

sub I2C{
    my $i = shift;
    return ($i & 0xff, $i >> 8 & 0xff, $i >> 16 & 0xff, $i >> 24 & 0xff);
}

sub GoogleHash{
    my $value = shift;
    @va = split '', $value;
    for $i (0..scalar(@va)-1){
	$va[$i] = ord($va[$i]);
    }
    $ch = checksum(@va);
    $ch = (($ch % 0x0d) & 7) | (($ch/7) << 2);
    
    my $sum;
    my $j = 0;
    for $i (0..19){
	($sum[$j++], $sum[$j++], $sum[$j++], $sum[$j++]) = I2C($ch-9*$i);
    }
    
    return sprintf("6%s", checksum(@sum));
}

sub MakePRQuery{
    my $url = shift;
    $url = "info:$url";
    return sprintf("http://toolbarqueries.google.com/search?q=%s&ch=%d&features=%s&oe=%s&ie=%s&client=%s", 
	uri_escape($url), GoogleHash($url), "Rank", "UTF-8", "UTF-8", "navclient-auto");
}

sub QueryPR{
    my $url = shift;
    $url = MakePRQuery($url);
    
    my $a = LWP::UserAgent->new(
	agent => "Mozilla/4.0 (compatible; GoogleToolbar 2.0.108-big; Windows XP 5.1)",
    );
    $a->timeout(30);
    $a->env_proxy;
    my $r = $a->get($url);

    if($r->is_success){
	my @data = split ":", $r->content;
	if(scalar(@data) eq 3){
	    return $data[2];
	}
    }
    return -1;
}

1;

#!/usr/local/bin/perl

use PR;

print "PR: " . PR::QueryPR("http://www.madjack.ru") . "\n";