xref: /webtrees/app/Http/Middleware/BadBotBlocker.php (revision 891c4176f7620b7d170f46aef035e8e6b19bd615) 
1<?php
2
3/**
4 * webtrees: online genealogy
5 * Copyright (C) 2022 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 <https://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\Registry;
24use Fisharebest\Webtrees\Validator;
25use Iodev\Whois\Loaders\CurlLoader;
26use Iodev\Whois\Modules\Asn\AsnRouteInfo;
27use Iodev\Whois\Whois;
28use IPLib\Address\AddressInterface;
29use IPLib\Factory as IPFactory;
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 array_filter;
38use function array_map;
39use function assert;
40use function gethostbyaddr;
41use function gethostbyname;
42use function preg_match_all;
43use function random_int;
44use function response;
45use function str_contains;
46use function str_ends_with;
47
48/**
49 * Middleware to block bad robots before they waste our valuable CPU cycles.
50 */
51class BadBotBlocker implements MiddlewareInterface
52{
53    // Cache whois requests.  Try to avoid all caches expiring at the same time.
54    private const WHOIS_TTL_MIN = 28 * 86400;
55    private const WHOIS_TTL_MAX = 35 * 86400;
56    private const WHOIS_TIMEOUT = 5;
57
58    // Bad robots - SEO optimisers, advertisers, etc.  This list is shared with robots.txt.
59    public const BAD_ROBOTS = [
60        'admantx',
61        'Adsbot',
62        'AhrefsBot',
63        'AspiegelBot',
64        'Barkrowler',
65        'BLEXBot',
66        'DataForSEO',
67        'DotBot',
68        'Grapeshot',
69        'ia_archiver',
70        'Linguee',
71        'MJ12bot',
72        'panscient',
73        'PetalBot',
74        'proximic',
75        'SemrushBot',
76        'Turnitin',
77        'XoviBot',
78        'ZoominfoBot',
79    ];
80
81    /**
82     * Some search engines use reverse/forward DNS to verify the IP address.
83     *
84     * @see https://developer.amazon.com/support/amazonbot
85     * @see https://support.google.com/webmasters/answer/80553?hl=en
86     * @see https://www.bing.com/webmaster/help/which-crawlers-does-bing-use-8c184ec0
87     * @see https://www.bing.com/webmaster/help/how-to-verify-bingbot-3905dc26
88     * @see https://yandex.com/support/webmaster/robot-workings/check-yandex-robots.html
89     */
90    private const ROBOT_REV_FWD_DNS = [
91        'Amazonbot'   => ['.crawl.amazon.com'],
92        'bingbot'     => ['.search.msn.com'],
93        'BingPreview' => ['.search.msn.com'],
94        'Google'      => ['.google.com', '.googlebot.com'],
95        'Mail.RU_Bot' => ['mail.ru'],
96        'msnbot'      => ['.search.msn.com'],
97        'Qwantify'    => ['.search.qwant.com'],
98        'Sogou'       => ['.crawl.sogou.com'],
99        'Yahoo'       => ['.crawl.yahoo.net'],
100        'Yandex'      => ['.yandex.ru', '.yandex.net', '.yandex.com'],
101    ];
102
103    /**
104     * Some search engines only use reverse DNS to verify the IP address.
105     *
106     * @see https://help.baidu.com/question?prod_id=99&class=0&id=3001
107     * @see https://napoveda.seznam.cz/en/full-text-search/seznambot-crawler
108     * @see https://www.ionos.de/terms-gtc/faq-crawler
109     */
110    private const ROBOT_REV_ONLY_DNS = [
111        'Baiduspider' => ['.baidu.com', '.baidu.jp'],
112        'FreshBot'    => ['.seznam.cz'],
113        'IonCrawl'    => ['.1und1.org'],
114    ];
115
116    /**
117     * Some search engines operate from designated IP addresses.
118     *
119     * @see https://www.apple.com/go/applebot
120     * @see https://help.duckduckgo.com/duckduckgo-help-pages/results/duckduckbot
121     */
122    private const ROBOT_IPS = [
123        'AppleBot'    => [
124            '17.0.0.0/8',
125        ],
126        'Ask Jeeves'  => [
127            '65.214.45.143',
128            '65.214.45.148',
129            '66.235.124.192',
130            '66.235.124.7',
131            '66.235.124.101',
132            '66.235.124.193',
133            '66.235.124.73',
134            '66.235.124.196',
135            '66.235.124.74',
136            '63.123.238.8',
137            '202.143.148.61',
138        ],
139        'DuckDuckBot' => [
140            '23.21.227.69',
141            '50.16.241.113',
142            '50.16.241.114',
143            '50.16.241.117',
144            '50.16.247.234',
145            '52.204.97.54',
146            '52.5.190.19',
147            '54.197.234.188',
148            '54.208.100.253',
149            '54.208.102.37',
150            '107.21.1.8',
151        ],
152    ];
153
154    /**
155     * Some search engines operate from within a designated autonomous system.
156     *
157     * @see https://developers.facebook.com/docs/sharing/webmasters/crawler
158     * @see https://www.facebook.com/peering/
159     */
160    private const ROBOT_ASNS = [
161        'facebook' => ['AS32934', 'AS63293'],
162        'twitter'  => ['AS13414'],
163    ];
164
165    /**
166     * @param ServerRequestInterface  $request
167     * @param RequestHandlerInterface $handler
168     *
169     * @return ResponseInterface
170     */
171    public function process(ServerRequestInterface $request, RequestHandlerInterface $handler): ResponseInterface
172    {
173        $ua      = Validator::serverParams($request)->string('HTTP_USER_AGENT', '');
174        $ip      = Validator::attributes($request)->string('client-ip');
175        $address = IPFactory::parseAddressString($ip);
176        assert($address instanceof AddressInterface);
177
178        foreach (self::BAD_ROBOTS as $robot) {
179            if (str_contains($ua, $robot)) {
180                return $this->response();
181            }
182        }
183
184        foreach (self::ROBOT_REV_FWD_DNS as $robot => $valid_domains) {
185            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, false)) {
186                return $this->response();
187            }
188        }
189
190        foreach (self::ROBOT_REV_ONLY_DNS as $robot => $valid_domains) {
191            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, true)) {
192                return $this->response();
193            }
194        }
195
196        foreach (self::ROBOT_IPS as $robot => $valid_ips) {
197            if (str_contains($ua, $robot)) {
198                foreach ($valid_ips as $ip) {
199                    $range = IPFactory::parseRangeString($ip);
200
201                    if ($range instanceof RangeInterface && $range->contains($address)) {
202                        continue 2;
203                    }
204                }
205
206                return $this->response();
207            }
208        }
209
210        foreach (self::ROBOT_ASNS as $robot => $asns) {
211            foreach ($asns as $asn) {
212                if (str_contains($ua, $robot)) {
213                    foreach ($this->fetchIpRangesForAsn($asn) as $range) {
214                        if ($range->contains($address)) {
215                            continue 2;
216                        }
217                    }
218
219                    return $this->response();
220                }
221            }
222        }
223
224        // Allow sites to block access from entire networks.
225        $block_asn = Validator::attributes($request)->string('block_asn', '');
226        preg_match_all('/(AS\d+)/', $block_asn, $matches);
227
228        foreach ($matches[1] as $asn) {
229            foreach ($this->fetchIpRangesForAsn($asn) as $range) {
230                if ($range->contains($address)) {
231                    return $this->response();
232                }
233            }
234        }
235
236        return $handler->handle($request);
237    }
238
239    /**
240     * Check that an IP address belongs to a robot operator using a forward/reverse DNS lookup.
241     *
242     * @param string        $ip
243     * @param array<string> $valid_domains
244     * @param bool          $reverse_only
245     *
246     * @return bool
247     */
248    private function checkRobotDNS(string $ip, array $valid_domains, bool $reverse_only): bool
249    {
250        $host = gethostbyaddr($ip);
251
252        if ($host === false) {
253            return false;
254        }
255
256        foreach ($valid_domains as $domain) {
257            if (str_ends_with($host, $domain)) {
258                return $reverse_only || $ip === gethostbyname($host);
259            }
260        }
261
262        return false;
263    }
264
265    /**
266     * Perform a whois search for an ASN.
267     *
268     * @param string $asn - The autonomous system number to query
269     *
270     * @return array<RangeInterface>
271     */
272    private function fetchIpRangesForAsn(string $asn): array
273    {
274        return Registry::cache()->file()->remember('whois-asn-' . $asn, static function () use ($asn): array {
275            $mapper = static fn (AsnRouteInfo $route_info): ?RangeInterface => IPFactory::parseRangeString($route_info->route ?: $route_info->route6);
276
277            try {
278                $loader = new CurlLoader(self::WHOIS_TIMEOUT);
279                $whois  = new Whois($loader);
280                $info   = $whois->loadAsnInfo($asn);
281                $routes = $info->routes;
282                $ranges = array_map($mapper, $routes);
283
284                return array_filter($ranges);
285            } catch (Throwable $ex) {
286                return [];
287            }
288        }, random_int(self::WHOIS_TTL_MIN, self::WHOIS_TTL_MAX));
289    }
290
291    /**
292     * @return ResponseInterface
293     */
294    private function response(): ResponseInterface
295    {
296        return response('Not acceptable', StatusCodeInterface::STATUS_NOT_ACCEPTABLE);
297    }
298}
299