xref: /webtrees/app/Http/Middleware/BadBotBlocker.php (revision 560f22f73974e95e0d89d39f82e4d007c062443c) 
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://support.google.com/webmasters/answer/80553?hl=en
85     * @see https://www.bing.com/webmaster/help/which-crawlers-does-bing-use-8c184ec0
86     * @see https://www.bing.com/webmaster/help/how-to-verify-bingbot-3905dc26
87     * @see https://yandex.com/support/webmaster/robot-workings/check-yandex-robots.html
88     */
89    private const ROBOT_REV_FWD_DNS = [
90        'bingbot'     => ['.search.msn.com'],
91        'BingPreview' => ['.search.msn.com'],
92        'Google'      => ['.google.com', '.googlebot.com'],
93        'Mail.RU_Bot' => ['mail.ru'],
94        'msnbot'      => ['.search.msn.com'],
95        'Qwantify'    => ['.search.qwant.com'],
96        'Sogou'       => ['.crawl.sogou.com'],
97        'Yahoo'       => ['.crawl.yahoo.net'],
98        'Yandex'      => ['.yandex.ru', '.yandex.net', '.yandex.com'],
99    ];
100
101    /**
102     * Some search engines only use reverse DNS to verify the IP address.
103     *
104     * @see https://help.baidu.com/question?prod_id=99&class=0&id=3001
105     * @see https://napoveda.seznam.cz/en/full-text-search/seznambot-crawler
106     * @see https://www.ionos.de/terms-gtc/faq-crawler
107     */
108    private const ROBOT_REV_ONLY_DNS = [
109        'Baiduspider' => ['.baidu.com', '.baidu.jp'],
110        'FreshBot'    => ['.seznam.cz'],
111        'IonCrawl'    => ['.1und1.org'],
112    ];
113
114    /**
115     * Some search engines operate from designated IP addresses.
116     *
117     * @see https://www.apple.com/go/applebot
118     * @see https://help.duckduckgo.com/duckduckgo-help-pages/results/duckduckbot
119     */
120    private const ROBOT_IPS = [
121        'AppleBot'    => [
122            '17.0.0.0/8',
123        ],
124        'Ask Jeeves'  => [
125            '65.214.45.143',
126            '65.214.45.148',
127            '66.235.124.192',
128            '66.235.124.7',
129            '66.235.124.101',
130            '66.235.124.193',
131            '66.235.124.73',
132            '66.235.124.196',
133            '66.235.124.74',
134            '63.123.238.8',
135            '202.143.148.61',
136        ],
137        'DuckDuckBot' => [
138            '23.21.227.69',
139            '50.16.241.113',
140            '50.16.241.114',
141            '50.16.241.117',
142            '50.16.247.234',
143            '52.204.97.54',
144            '52.5.190.19',
145            '54.197.234.188',
146            '54.208.100.253',
147            '54.208.102.37',
148            '107.21.1.8',
149        ],
150    ];
151
152    /**
153     * Some search engines operate from within a designated autonomous system.
154     *
155     * @see https://developers.facebook.com/docs/sharing/webmasters/crawler
156     * @see https://www.facebook.com/peering/
157     */
158    private const ROBOT_ASNS = [
159        'facebook' => ['AS32934', 'AS63293'],
160        'twitter'  => ['AS13414'],
161    ];
162
163    /**
164     * @param ServerRequestInterface  $request
165     * @param RequestHandlerInterface $handler
166     *
167     * @return ResponseInterface
168     */
169    public function process(ServerRequestInterface $request, RequestHandlerInterface $handler): ResponseInterface
170    {
171        $ua      = Validator::serverParams($request)->string('HTTP_USER_AGENT', '');
172        $ip      = Validator::attributes($request)->string('client-ip');
173        $address = IPFactory::parseAddressString($ip);
174        assert($address instanceof AddressInterface);
175
176        foreach (self::BAD_ROBOTS as $robot) {
177            if (str_contains($ua, $robot)) {
178                return $this->response();
179            }
180        }
181
182        foreach (self::ROBOT_REV_FWD_DNS as $robot => $valid_domains) {
183            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, false)) {
184                return $this->response();
185            }
186        }
187
188        foreach (self::ROBOT_REV_ONLY_DNS as $robot => $valid_domains) {
189            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, true)) {
190                return $this->response();
191            }
192        }
193
194        foreach (self::ROBOT_IPS as $robot => $valid_ips) {
195            if (str_contains($ua, $robot)) {
196                foreach ($valid_ips as $ip) {
197                    $range = IPFactory::parseRangeString($ip);
198
199                    if ($range instanceof RangeInterface && $range->contains($address)) {
200                        continue 2;
201                    }
202                }
203
204                return $this->response();
205            }
206        }
207
208        foreach (self::ROBOT_ASNS as $robot => $asns) {
209            foreach ($asns as $asn) {
210                if (str_contains($ua, $robot)) {
211                    foreach ($this->fetchIpRangesForAsn($asn) as $range) {
212                        if ($range->contains($address)) {
213                            continue 2;
214                        }
215                    }
216
217                    return $this->response();
218                }
219            }
220        }
221
222        // Allow sites to block access from entire networks.
223        $block_asn = Validator::attributes($request)->string('block_asn', '');
224        preg_match_all('/(AS\d+)/', $block_asn, $matches);
225
226        foreach ($matches[1] as $asn) {
227            foreach ($this->fetchIpRangesForAsn($asn) as $range) {
228                if ($range->contains($address)) {
229                    return $this->response();
230                }
231            }
232        }
233
234        return $handler->handle($request);
235    }
236
237    /**
238     * Check that an IP address belongs to a robot operator using a forward/reverse DNS lookup.
239     *
240     * @param string        $ip
241     * @param array<string> $valid_domains
242     * @param bool          $reverse_only
243     *
244     * @return bool
245     */
246    private function checkRobotDNS(string $ip, array $valid_domains, bool $reverse_only): bool
247    {
248        $host = gethostbyaddr($ip);
249
250        if ($host === false) {
251            return false;
252        }
253
254        foreach ($valid_domains as $domain) {
255            if (str_ends_with($host, $domain)) {
256                return $reverse_only || $ip === gethostbyname($host);
257            }
258        }
259
260        return false;
261    }
262
263    /**
264     * Perform a whois search for an ASN.
265     *
266     * @param string $asn - The autonomous system number to query
267     *
268     * @return array<RangeInterface>
269     */
270    private function fetchIpRangesForAsn(string $asn): array
271    {
272        return Registry::cache()->file()->remember('whois-asn-' . $asn, static function () use ($asn): array {
273            $mapper = static fn (AsnRouteInfo $route_info): ?RangeInterface => IPFactory::parseRangeString($route_info->route ?: $route_info->route6);
274
275            try {
276                $loader = new CurlLoader(self::WHOIS_TIMEOUT);
277                $whois  = new Whois($loader);
278                $info   = $whois->loadAsnInfo($asn);
279                $routes = $info->routes;
280                $ranges = array_map($mapper, $routes);
281
282                return array_filter($ranges);
283            } catch (Throwable $ex) {
284                return [];
285            }
286        }, random_int(self::WHOIS_TTL_MIN, self::WHOIS_TTL_MAX));
287    }
288
289    /**
290     * @return ResponseInterface
291     */
292    private function response(): ResponseInterface
293    {
294        return response('Not acceptable', StatusCodeInterface::STATUS_NOT_ACCEPTABLE);
295    }
296}
297