xref: /webtrees/app/Http/Middleware/BadBotBlocker.php (revision 748dbe155a6d19d66918ad136947fa23ee8f8469) 
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        'netEstate NE',
73        'panscient',
74        'PetalBot',
75        'proximic',
76        'SemrushBot',
77        'SEOkicks',
78        'SiteKiosk',
79        'Turnitin',
80        'XoviBot',
81        'ZoominfoBot',
82    ];
83
84    /**
85     * Some search engines use reverse/forward DNS to verify the IP address.
86     *
87     * @see https://developer.amazon.com/support/amazonbot
88     * @see https://support.google.com/webmasters/answer/80553?hl=en
89     * @see https://www.bing.com/webmaster/help/which-crawlers-does-bing-use-8c184ec0
90     * @see https://www.bing.com/webmaster/help/how-to-verify-bingbot-3905dc26
91     * @see https://yandex.com/support/webmaster/robot-workings/check-yandex-robots.html
92     */
93    private const ROBOT_REV_FWD_DNS = [
94        'Amazonbot'   => ['.crawl.amazon.com'],
95        'bingbot'     => ['.search.msn.com'],
96        'BingPreview' => ['.search.msn.com'],
97        'Google'      => ['.google.com', '.googlebot.com'],
98        'Mail.RU_Bot' => ['.mail.ru'],
99        'msnbot'      => ['.search.msn.com'],
100        'Qwantify'    => ['.search.qwant.com'],
101        'Sogou'       => ['.crawl.sogou.com'],
102        'Yahoo'       => ['.crawl.yahoo.net'],
103        'Yandex'      => ['.yandex.ru', '.yandex.net', '.yandex.com'],
104    ];
105
106    /**
107     * Some search engines only use reverse DNS to verify the IP address.
108     *
109     * @see https://help.baidu.com/question?prod_id=99&class=0&id=3001
110     * @see https://napoveda.seznam.cz/en/full-text-search/seznambot-crawler
111     * @see https://www.ionos.de/terms-gtc/faq-crawler
112     */
113    private const ROBOT_REV_ONLY_DNS = [
114        'Baiduspider' => ['.baidu.com', '.baidu.jp'],
115        'FreshBot'    => ['.seznam.cz'],
116        'IonCrawl'    => ['.1und1.org'],
117        'Neevabot'    => ['.neeva.com'],
118    ];
119
120    /**
121     * Some search engines operate from designated IP addresses.
122     *
123     * @see https://www.apple.com/go/applebot
124     * @see https://help.duckduckgo.com/duckduckgo-help-pages/results/duckduckbot
125     */
126    private const ROBOT_IPS = [
127        'AppleBot'    => [
128            '17.0.0.0/8',
129        ],
130        'Ask Jeeves'  => [
131            '65.214.45.143',
132            '65.214.45.148',
133            '66.235.124.192',
134            '66.235.124.7',
135            '66.235.124.101',
136            '66.235.124.193',
137            '66.235.124.73',
138            '66.235.124.196',
139            '66.235.124.74',
140            '63.123.238.8',
141            '202.143.148.61',
142        ],
143        'DuckDuckBot' => [
144            '23.21.227.69',
145            '50.16.241.113',
146            '50.16.241.114',
147            '50.16.241.117',
148            '50.16.247.234',
149            '52.204.97.54',
150            '52.5.190.19',
151            '54.197.234.188',
152            '54.208.100.253',
153            '54.208.102.37',
154            '107.21.1.8',
155        ],
156    ];
157
158    /**
159     * Some search engines operate from within a designated autonomous system.
160     *
161     * @see https://developers.facebook.com/docs/sharing/webmasters/crawler
162     * @see https://www.facebook.com/peering/
163     */
164    private const ROBOT_ASNS = [
165        'facebook' => ['AS32934', 'AS63293'],
166        'twitter'  => ['AS13414'],
167    ];
168
169    /**
170     * @param ServerRequestInterface  $request
171     * @param RequestHandlerInterface $handler
172     *
173     * @return ResponseInterface
174     */
175    public function process(ServerRequestInterface $request, RequestHandlerInterface $handler): ResponseInterface
176    {
177        $ua      = Validator::serverParams($request)->string('HTTP_USER_AGENT', '');
178        $ip      = Validator::attributes($request)->string('client-ip');
179        $address = IPFactory::parseAddressString($ip);
180        assert($address instanceof AddressInterface);
181
182        foreach (self::BAD_ROBOTS as $robot) {
183            if (str_contains($ua, $robot)) {
184                return $this->response();
185            }
186        }
187
188        foreach (self::ROBOT_REV_FWD_DNS as $robot => $valid_domains) {
189            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, false)) {
190                return $this->response();
191            }
192        }
193
194        foreach (self::ROBOT_REV_ONLY_DNS as $robot => $valid_domains) {
195            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, true)) {
196                return $this->response();
197            }
198        }
199
200        foreach (self::ROBOT_IPS as $robot => $valid_ips) {
201            if (str_contains($ua, $robot)) {
202                foreach ($valid_ips as $ip) {
203                    $range = IPFactory::parseRangeString($ip);
204
205                    if ($range instanceof RangeInterface && $range->contains($address)) {
206                        continue 2;
207                    }
208                }
209
210                return $this->response();
211            }
212        }
213
214        foreach (self::ROBOT_ASNS as $robot => $asns) {
215            foreach ($asns as $asn) {
216                if (str_contains($ua, $robot)) {
217                    foreach ($this->fetchIpRangesForAsn($asn) as $range) {
218                        if ($range->contains($address)) {
219                            continue 2;
220                        }
221                    }
222
223                    return $this->response();
224                }
225            }
226        }
227
228        // Allow sites to block access from entire networks.
229        $block_asn = Validator::attributes($request)->string('block_asn', '');
230        preg_match_all('/(AS\d+)/', $block_asn, $matches);
231
232        foreach ($matches[1] as $asn) {
233            foreach ($this->fetchIpRangesForAsn($asn) as $range) {
234                if ($range->contains($address)) {
235                    return $this->response();
236                }
237            }
238        }
239
240        return $handler->handle($request);
241    }
242
243    /**
244     * Check that an IP address belongs to a robot operator using a forward/reverse DNS lookup.
245     *
246     * @param string        $ip
247     * @param array<string> $valid_domains
248     * @param bool          $reverse_only
249     *
250     * @return bool
251     */
252    private function checkRobotDNS(string $ip, array $valid_domains, bool $reverse_only): bool
253    {
254        $host = gethostbyaddr($ip);
255
256        if ($host === false) {
257            return false;
258        }
259
260        foreach ($valid_domains as $domain) {
261            if (str_ends_with($host, $domain)) {
262                return $reverse_only || $ip === gethostbyname($host);
263            }
264        }
265
266        return false;
267    }
268
269    /**
270     * Perform a whois search for an ASN.
271     *
272     * @param string $asn - The autonomous system number to query
273     *
274     * @return array<RangeInterface>
275     */
276    private function fetchIpRangesForAsn(string $asn): array
277    {
278        return Registry::cache()->file()->remember('whois-asn-' . $asn, static function () use ($asn): array {
279            $mapper = static fn (AsnRouteInfo $route_info): ?RangeInterface => IPFactory::parseRangeString($route_info->route ?: $route_info->route6);
280
281            try {
282                $loader = new CurlLoader(self::WHOIS_TIMEOUT);
283                $whois  = new Whois($loader);
284                $info   = $whois->loadAsnInfo($asn);
285                $routes = $info->routes;
286                $ranges = array_map($mapper, $routes);
287
288                return array_filter($ranges);
289            } catch (Throwable $ex) {
290                return [];
291            }
292        }, random_int(self::WHOIS_TTL_MIN, self::WHOIS_TTL_MAX));
293    }
294
295    /**
296     * @return ResponseInterface
297     */
298    private function response(): ResponseInterface
299    {
300        return response('Not acceptable', StatusCodeInterface::STATUS_NOT_ACCEPTABLE);
301    }
302}
303