xref: /webtrees/app/Http/Middleware/BadBotBlocker.php (revision e99ac601cee7ced57ac8545ceba47ece1b15d883) 
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        'DotBot',
67        'Grapeshot',
68        'ia_archiver',
69        'Linguee',
70        'MJ12bot',
71        'panscient',
72        'PetalBot',
73        'proximic',
74        'SemrushBot',
75        'Turnitin',
76        'XoviBot',
77        'ZoominfoBot',
78    ];
79
80    /**
81     * Some search engines use reverse/forward DNS to verify the IP address.
82     *
83     * @see https://support.google.com/webmasters/answer/80553?hl=en
84     * @see https://www.bing.com/webmaster/help/which-crawlers-does-bing-use-8c184ec0
85     * @see https://www.bing.com/webmaster/help/how-to-verify-bingbot-3905dc26
86     * @see https://yandex.com/support/webmaster/robot-workings/check-yandex-robots.html
87     */
88    private const ROBOT_REV_FWD_DNS = [
89        'bingbot'     => ['.search.msn.com'],
90        'BingPreview' => ['.search.msn.com'],
91        'Google'      => ['.google.com', '.googlebot.com'],
92        'Mail.RU_Bot' => ['mail.ru'],
93        'msnbot'      => ['.search.msn.com'],
94        'Qwantify'    => ['.search.qwant.com'],
95        'Sogou'       => ['.crawl.sogou.com'],
96        'Yahoo'       => ['.crawl.yahoo.net'],
97        'Yandex'      => ['.yandex.ru', '.yandex.net', '.yandex.com'],
98    ];
99
100    /**
101     * Some search engines only use reverse DNS to verify the IP address.
102     *
103     * @see https://help.baidu.com/question?prod_id=99&class=0&id=3001
104     * @see https://napoveda.seznam.cz/en/full-text-search/seznambot-crawler
105     */
106    private const ROBOT_REV_ONLY_DNS = [
107        'Baiduspider' => ['.baidu.com', '.baidu.jp'],
108        'FreshBot'    => ['.seznam.cz'],
109        'Seznam'      => ['.seznam.cz'],
110    ];
111
112    /**
113     * Some search engines operate from designated IP addresses.
114     *
115     * @see https://www.apple.com/go/applebot
116     * @see https://help.duckduckgo.com/duckduckgo-help-pages/results/duckduckbot
117     */
118    private const ROBOT_IPS = [
119        'AppleBot'    => [
120            '17.0.0.0/8',
121        ],
122        'Ask Jeeves'  => [
123            '65.214.45.143',
124            '65.214.45.148',
125            '66.235.124.192',
126            '66.235.124.7',
127            '66.235.124.101',
128            '66.235.124.193',
129            '66.235.124.73',
130            '66.235.124.196',
131            '66.235.124.74',
132            '63.123.238.8',
133            '202.143.148.61',
134        ],
135        'DuckDuckBot' => [
136            '23.21.227.69',
137            '50.16.241.113',
138            '50.16.241.114',
139            '50.16.241.117',
140            '50.16.247.234',
141            '52.204.97.54',
142            '52.5.190.19',
143            '54.197.234.188',
144            '54.208.100.253',
145            '54.208.102.37',
146            '107.21.1.8',
147        ],
148    ];
149
150    /**
151     * Some search engines operate from within a designated autonomous system.
152     *
153     * @see https://developers.facebook.com/docs/sharing/webmasters/crawler
154     * @see https://www.facebook.com/peering/
155     */
156    private const ROBOT_ASNS = [
157        'facebook' => ['AS32934', 'AS63293'],
158        'twitter'  => ['AS13414'],
159    ];
160
161    /**
162     * @param ServerRequestInterface  $request
163     * @param RequestHandlerInterface $handler
164     *
165     * @return ResponseInterface
166     */
167    public function process(ServerRequestInterface $request, RequestHandlerInterface $handler): ResponseInterface
168    {
169        $ua      = Validator::serverParams($request)->string('HTTP_USER_AGENT', '');
170        $ip      = Validator::attributes($request)->string('client-ip');
171        $address = IPFactory::parseAddressString($ip);
172        assert($address instanceof AddressInterface);
173
174        foreach (self::BAD_ROBOTS as $robot) {
175            if (str_contains($ua, $robot)) {
176                return $this->response();
177            }
178        }
179
180        foreach (self::ROBOT_REV_FWD_DNS as $robot => $valid_domains) {
181            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, false)) {
182                return $this->response();
183            }
184        }
185
186        foreach (self::ROBOT_REV_ONLY_DNS as $robot => $valid_domains) {
187            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, true)) {
188                return $this->response();
189            }
190        }
191
192        foreach (self::ROBOT_IPS as $robot => $valid_ips) {
193            if (str_contains($ua, $robot)) {
194                foreach ($valid_ips as $ip) {
195                    $range = IPFactory::parseRangeString($ip);
196
197                    if ($range instanceof RangeInterface && $range->contains($address)) {
198                        continue 2;
199                    }
200                }
201
202                return $this->response();
203            }
204        }
205
206        foreach (self::ROBOT_ASNS as $robot => $asns) {
207            foreach ($asns as $asn) {
208                if (str_contains($ua, $robot)) {
209                    foreach ($this->fetchIpRangesForAsn($asn) as $range) {
210                        if ($range->contains($address)) {
211                            continue 2;
212                        }
213                    }
214
215                    return $this->response();
216                }
217            }
218        }
219
220        // Allow sites to block access from entire networks.
221        $block_asn = Validator::attributes($request)->string('block_asn', '');
222        preg_match_all('/(AS\d+)/', $block_asn, $matches);
223
224        foreach ($matches[1] as $asn) {
225            foreach ($this->fetchIpRangesForAsn($asn) as $range) {
226                if ($range->contains($address)) {
227                    return $this->response();
228                }
229            }
230        }
231
232        return $handler->handle($request);
233    }
234
235    /**
236     * Check that an IP address belongs to a robot operator using a forward/reverse DNS lookup.
237     *
238     * @param string        $ip
239     * @param array<string> $valid_domains
240     * @param bool          $reverse_only
241     *
242     * @return bool
243     */
244    private function checkRobotDNS(string $ip, array $valid_domains, bool $reverse_only): bool
245    {
246        $host = gethostbyaddr($ip);
247
248        if ($host === false) {
249            return false;
250        }
251
252        foreach ($valid_domains as $domain) {
253            if (str_ends_with($host, $domain)) {
254                return $reverse_only || $ip === gethostbyname($host);
255            }
256        }
257
258        return false;
259    }
260
261    /**
262     * Perform a whois search for an ASN.
263     *
264     * @param string $asn - The autonomous system number to query
265     *
266     * @return array<RangeInterface>
267     */
268    private function fetchIpRangesForAsn(string $asn): array
269    {
270        return Registry::cache()->file()->remember('whois-asn-' . $asn, static function () use ($asn): array {
271            $mapper = static fn (AsnRouteInfo $route_info): ?RangeInterface => IPFactory::parseRangeString($route_info->route ?: $route_info->route6);
272
273            try {
274                $loader = new CurlLoader(self::WHOIS_TIMEOUT);
275                $whois  = new Whois($loader);
276                $info   = $whois->loadAsnInfo($asn);
277                $routes = $info->routes;
278                $ranges = array_map($mapper, $routes);
279
280                return array_filter($ranges);
281            } catch (Throwable $ex) {
282                return [];
283            }
284        }, random_int(self::WHOIS_TTL_MIN, self::WHOIS_TTL_MAX));
285    }
286
287    /**
288     * @return ResponseInterface
289     */
290    private function response(): ResponseInterface
291    {
292        return response('Not acceptable', StatusCodeInterface::STATUS_NOT_ACCEPTABLE);
293    }
294}
295