xref: /webtrees/app/Http/Middleware/BadBotBlocker.php (revision b63ebfff10c4f748c10374a14a84550d3b653912) 
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     */
107    private const ROBOT_REV_ONLY_DNS = [
108        'Baiduspider' => ['.baidu.com', '.baidu.jp'],
109        'FreshBot'    => ['.seznam.cz'],
110        'Seznam'      => ['.seznam.cz'],
111    ];
112
113    /**
114     * Some search engines operate from designated IP addresses.
115     *
116     * @see https://www.apple.com/go/applebot
117     * @see https://help.duckduckgo.com/duckduckgo-help-pages/results/duckduckbot
118     */
119    private const ROBOT_IPS = [
120        'AppleBot'    => [
121            '17.0.0.0/8',
122        ],
123        'Ask Jeeves'  => [
124            '65.214.45.143',
125            '65.214.45.148',
126            '66.235.124.192',
127            '66.235.124.7',
128            '66.235.124.101',
129            '66.235.124.193',
130            '66.235.124.73',
131            '66.235.124.196',
132            '66.235.124.74',
133            '63.123.238.8',
134            '202.143.148.61',
135        ],
136        'DuckDuckBot' => [
137            '23.21.227.69',
138            '50.16.241.113',
139            '50.16.241.114',
140            '50.16.241.117',
141            '50.16.247.234',
142            '52.204.97.54',
143            '52.5.190.19',
144            '54.197.234.188',
145            '54.208.100.253',
146            '54.208.102.37',
147            '107.21.1.8',
148        ],
149    ];
150
151    /**
152     * Some search engines operate from within a designated autonomous system.
153     *
154     * @see https://developers.facebook.com/docs/sharing/webmasters/crawler
155     * @see https://www.facebook.com/peering/
156     */
157    private const ROBOT_ASNS = [
158        'facebook' => ['AS32934', 'AS63293'],
159        'twitter'  => ['AS13414'],
160    ];
161
162    /**
163     * @param ServerRequestInterface  $request
164     * @param RequestHandlerInterface $handler
165     *
166     * @return ResponseInterface
167     */
168    public function process(ServerRequestInterface $request, RequestHandlerInterface $handler): ResponseInterface
169    {
170        $ua      = Validator::serverParams($request)->string('HTTP_USER_AGENT', '');
171        $ip      = Validator::attributes($request)->string('client-ip');
172        $address = IPFactory::parseAddressString($ip);
173        assert($address instanceof AddressInterface);
174
175        foreach (self::BAD_ROBOTS as $robot) {
176            if (str_contains($ua, $robot)) {
177                return $this->response();
178            }
179        }
180
181        foreach (self::ROBOT_REV_FWD_DNS as $robot => $valid_domains) {
182            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, false)) {
183                return $this->response();
184            }
185        }
186
187        foreach (self::ROBOT_REV_ONLY_DNS as $robot => $valid_domains) {
188            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, true)) {
189                return $this->response();
190            }
191        }
192
193        foreach (self::ROBOT_IPS as $robot => $valid_ips) {
194            if (str_contains($ua, $robot)) {
195                foreach ($valid_ips as $ip) {
196                    $range = IPFactory::parseRangeString($ip);
197
198                    if ($range instanceof RangeInterface && $range->contains($address)) {
199                        continue 2;
200                    }
201                }
202
203                return $this->response();
204            }
205        }
206
207        foreach (self::ROBOT_ASNS as $robot => $asns) {
208            foreach ($asns as $asn) {
209                if (str_contains($ua, $robot)) {
210                    foreach ($this->fetchIpRangesForAsn($asn) as $range) {
211                        if ($range->contains($address)) {
212                            continue 2;
213                        }
214                    }
215
216                    return $this->response();
217                }
218            }
219        }
220
221        // Allow sites to block access from entire networks.
222        $block_asn = Validator::attributes($request)->string('block_asn', '');
223        preg_match_all('/(AS\d+)/', $block_asn, $matches);
224
225        foreach ($matches[1] as $asn) {
226            foreach ($this->fetchIpRangesForAsn($asn) as $range) {
227                if ($range->contains($address)) {
228                    return $this->response();
229                }
230            }
231        }
232
233        return $handler->handle($request);
234    }
235
236    /**
237     * Check that an IP address belongs to a robot operator using a forward/reverse DNS lookup.
238     *
239     * @param string        $ip
240     * @param array<string> $valid_domains
241     * @param bool          $reverse_only
242     *
243     * @return bool
244     */
245    private function checkRobotDNS(string $ip, array $valid_domains, bool $reverse_only): bool
246    {
247        $host = gethostbyaddr($ip);
248
249        if ($host === false) {
250            return false;
251        }
252
253        foreach ($valid_domains as $domain) {
254            if (str_ends_with($host, $domain)) {
255                return $reverse_only || $ip === gethostbyname($host);
256            }
257        }
258
259        return false;
260    }
261
262    /**
263     * Perform a whois search for an ASN.
264     *
265     * @param string $asn - The autonomous system number to query
266     *
267     * @return array<RangeInterface>
268     */
269    private function fetchIpRangesForAsn(string $asn): array
270    {
271        return Registry::cache()->file()->remember('whois-asn-' . $asn, static function () use ($asn): array {
272            $mapper = static fn (AsnRouteInfo $route_info): ?RangeInterface => IPFactory::parseRangeString($route_info->route ?: $route_info->route6);
273
274            try {
275                $loader = new CurlLoader(self::WHOIS_TIMEOUT);
276                $whois  = new Whois($loader);
277                $info   = $whois->loadAsnInfo($asn);
278                $routes = $info->routes;
279                $ranges = array_map($mapper, $routes);
280
281                return array_filter($ranges);
282            } catch (Throwable $ex) {
283                return [];
284            }
285        }, random_int(self::WHOIS_TTL_MIN, self::WHOIS_TTL_MAX));
286    }
287
288    /**
289     * @return ResponseInterface
290     */
291    private function response(): ResponseInterface
292    {
293        return response('Not acceptable', StatusCodeInterface::STATUS_NOT_ACCEPTABLE);
294    }
295}
296