xref: /webtrees/app/Http/Middleware/BadBotBlocker.php (revision 5c20d904a5240e21cf5895bd3d43ed728a641d79) 
1<?php
2
3/**
4 * webtrees: online genealogy
5 * Copyright (C) 2019 webtrees development team
6 * This program is free software: you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License as published by
8 * the Free Software Foundation, either version 3 of the License, or
9 * (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
14 * You should have received a copy of the GNU General Public License
15 * along with this program. If not, see <http://www.gnu.org/licenses/>.
16 */
17
18declare(strict_types=1);
19
20namespace Fisharebest\Webtrees\Http\Middleware;
21
22use Fig\Http\Message\StatusCodeInterface;
23use Fisharebest\Webtrees\Cache;
24use Illuminate\Support\Str;
25use Iodev\Whois\Loaders\CurlLoader;
26use Iodev\Whois\Modules\Asn\AsnRouteInfo;
27use Iodev\Whois\Whois;
28use IPLib\Address\AddressInterface;
29use IPLib\Factory;
30use IPLib\Range\RangeInterface;
31use Psr\Http\Message\ResponseInterface;
32use Psr\Http\Message\ServerRequestInterface;
33use Psr\Http\Server\MiddlewareInterface;
34use Psr\Http\Server\RequestHandlerInterface;
35use Throwable;
36
37use function app;
38use function array_map;
39use function assert;
40use function gethostbyaddr;
41use function gethostbyname;
42use function in_array;
43use function response;
44
45/**
46 * Middleware to block bad robots before they waste our valuable CPU cycles.
47 */
48class BadBotBlocker implements MiddlewareInterface
49{
50    // Cache whois requests.  Try to avoid all caches expiring at the same time.
51    private const WHOIS_TTL_MIN = 28 * 86400;
52    private const WHOIS_TTL_MAX = 35 * 86400;
53    private const WHOIS_TIMEOUT = 5;
54
55    // Bad robots - SEO optimisers, advertisers, etc
56    private const BAD_ROBOTS = [
57        'admantx',
58        'AhrefsBot',
59        'AspiegelBot',
60        'DotBot',
61        'Grapeshot',
62        'ia_archiver',
63        'MJ12bot',
64        'panscient',
65        'proximic',
66        'SemrushBot',
67        'XoviBot',
68    ];
69
70    /**
71     * Some search engines use reverse/forward DNS to verify the IP address.
72     *
73     * @see https://support.google.com/webmasters/answer/80553?hl=en
74     * @see https://www.bing.com/webmaster/help/which-crawlers-does-bing-use-8c184ec0
75     * @see https://www.bing.com/webmaster/help/how-to-verify-bingbot-3905dc26
76     * @see https://yandex.com/support/webmaster/robot-workings/check-yandex-robots.html
77     */
78    private const ROBOT_REV_FWD_DNS = [
79        'bingbot'     => ['.search.msn.com'],
80        'BingPreview' => ['.search.msn.com'],
81        'Google'      => ['.google.com', '.googlebot.com'],
82        'msnbot'      => ['.search.msn.com'],
83        'Qwantify'    => ['.search.qwant.com'],
84        'Sogou'       => ['.crawl.sogou.com'],
85        'Yahoo'       => ['.crawl.yahoo.net'],
86        'Yandex'      => ['.yandex.ru', '.yandex.net', '.yandex.com'],
87    ];
88
89    /**
90     * Some search engines only use reverse DNS to verify the IP address.
91     *
92     * @see https://help.baidu.com/question?prod_id=99&class=0&id=3001
93     */
94    private const ROBOT_REV_ONLY_DNS = [
95        'Baidu' => ['.baidu.com', '.baidu.jp'],
96    ];
97
98    /**
99     * Some search engines operate from designated IP addresses.
100     *
101     * @see http://www.apple.com/go/applebot
102     * @see https://help.duckduckgo.com/duckduckgo-help-pages/results/duckduckbot
103     */
104    private const ROBOT_IPS = [
105        'AppleBot'    => [
106            '17.0.0.0/8',
107        ],
108        'Ask Jeeves'  => [
109            '65.214.45.143',
110            '65.214.45.148',
111            '66.235.124.192',
112            '66.235.124.7',
113            '66.235.124.101',
114            '66.235.124.193',
115            '66.235.124.73',
116            '66.235.124.196',
117            '66.235.124.74',
118            '63.123.238.8',
119            '202.143.148.61',
120        ],
121        'DuckDuckBot' => [
122            '23.21.227.69',
123            '50.16.241.113',
124            '50.16.241.114',
125            '50.16.241.117',
126            '50.16.247.234',
127            '52.204.97.54',
128            '52.5.190.19',
129            '54.197.234.188',
130            '54.208.100.253',
131            '54.208.102.37',
132            '107.21.1.8',
133        ],
134    ];
135
136    /**
137     * Some search engines operate from within a designated autonomous system.
138     *
139     * @see https://developers.facebook.com/docs/sharing/webmasters/crawler
140     */
141    private const ROBOT_ASN = [
142        'facebook' => 'AS32934',
143        'twitter'  => 'AS13414',
144    ];
145
146    /**
147     * These ASNs belong to server farms.
148     */
149    private const BLOCK_ASN = [
150        'hetzner'   => 'AS24920',
151        'hostdime'  => 'AS33182',
152        'linode'    => 'AS63949',
153        'ovh'       => 'AS16276',
154        'rackspace' => 'AS15395',
155    ];
156
157    /**
158     * @param ServerRequestInterface  $request
159     * @param RequestHandlerInterface $handler
160     *
161     * @return ResponseInterface
162     */
163    public function process(ServerRequestInterface $request, RequestHandlerInterface $handler): ResponseInterface
164    {
165        $ua      = $request->getServerParams()['HTTP_USER_AGENT'] ?? '';
166        $ip      = $request->getAttribute('client-ip');
167        $address = Factory::addressFromString($ip);
168        assert($address instanceof AddressInterface);
169
170        if (Str::contains($ua, self::BAD_ROBOTS)) {
171            return $this->response();
172        }
173
174        foreach (self::ROBOT_REV_FWD_DNS as $robot => $valid_domains) {
175            if (Str::contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, false)) {
176                return $this->response();
177            }
178        }
179
180        foreach (self::ROBOT_REV_ONLY_DNS as $robot => $valid_domains) {
181            if (Str::contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, true)) {
182                return $this->response();
183            }
184        }
185
186        foreach (self::ROBOT_IPS as $robot => $valid_ips) {
187            if (Str::contains($ua, $robot)) {
188                foreach ($valid_ips as $ip) {
189                    $range = Factory::rangeFromString($ip);
190
191                    if ($range instanceof RangeInterface && $range->contains($address)) {
192                        continue 2;
193                    }
194                }
195
196                return $this->response();
197            }
198        }
199
200        foreach (self::ROBOT_ASN as $robot => $asn) {
201            if (Str::contains($ua, $robot)) {
202                foreach ($this->fetchIpRangesForAsn($asn) as $range) {
203                    if ($range->contains($address)) {
204                        continue 2;
205                    }
206                }
207
208                return $this->response();
209            }
210        }
211
212        // This is potentially controversial, and whois lookups may be slow.
213        //foreach (self::BLOCK_ASN as $host => $asn) {
214        //    foreach ($this->fetchIpRangesForAsn($asn) as $range) {
215        //        if ($range->contains($address)) {
216        //            return $this->response();
217        //        }
218        //    }
219        //}
220
221        return $handler->handle($request);
222    }
223
224    /**
225     * Check that an IP address belongs to a robot operator using a forward/reverse DNS lookup.
226     *
227     * @param string        $ip
228     * @param array<string> $valid_domains
229     * @param bool          $reverse_only
230     *
231     * @return bool
232     */
233    private function checkRobotDNS(string $ip, array $valid_domains, bool $reverse_only): bool
234    {
235        $host = gethostbyaddr($ip);
236
237        if ($host === false || !Str::endsWith($host, $valid_domains)) {
238            return false;
239        }
240
241        return $reverse_only || $ip === gethostbyname($host);
242    }
243
244    /**
245     * Perform a whois search for an ASN.
246     *
247     * @param string $asn - The autonomous system number to query
248     *
249     * @return array<RangeInterface>
250     */
251    private function fetchIpRangesForAsn(string $asn): array
252    {
253        $cache = app('cache.files');
254        assert($cache instanceof Cache);
255
256        return $cache->remember('whois-asn-' . $asn, static function () use ($asn): array {
257            try {
258                $loader = new CurlLoader(self::WHOIS_TIMEOUT);
259                $whois  = new Whois($loader);
260                $info   = $whois->loadAsnInfo($asn);
261                $routes = $info->getRoutes();
262                $ranges = array_map(static function (AsnRouteInfo $route_info): ?RangeInterface {
263                    return Factory::rangeFromString($route_info->getRoute() ?: $route_info->getRoute6());
264                }, $routes);
265
266                return array_filter($ranges);
267            } catch (Throwable $ex) {
268                return [];
269            }
270        }, random_int(self::WHOIS_TTL_MIN, self::WHOIS_TTL_MAX));
271    }
272
273    /**
274     * @return ResponseInterface
275     */
276    private function response(): ResponseInterface
277    {
278        return response('Not acceptable', StatusCodeInterface::STATUS_NOT_ACCEPTABLE);
279    }
280}
281