xref: /webtrees/app/Http/Middleware/BadBotBlocker.php (revision e47c3c91014b6760740554027de9a9a8acfc764e) 
1<?php
2
3/**
4 * webtrees: online genealogy
5 * Copyright (C) 2023 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 GuzzleHttp\Client;
26use GuzzleHttp\Exception\GuzzleException;
27use Iodev\Whois\Loaders\CurlLoader;
28use Iodev\Whois\Modules\Asn\AsnRouteInfo;
29use Iodev\Whois\Whois;
30use IPLib\Address\AddressInterface;
31use IPLib\Factory as IPFactory;
32use IPLib\Range\RangeInterface;
33use Psr\Http\Message\ResponseInterface;
34use Psr\Http\Message\ServerRequestInterface;
35use Psr\Http\Server\MiddlewareInterface;
36use Psr\Http\Server\RequestHandlerInterface;
37use Throwable;
38
39use function array_filter;
40use function array_map;
41use function assert;
42use function gethostbyaddr;
43use function gethostbyname;
44use function preg_match_all;
45use function random_int;
46use function response;
47use function str_contains;
48use function str_ends_with;
49
50/**
51 * Middleware to block bad robots before they waste our valuable CPU cycles.
52 */
53class BadBotBlocker implements MiddlewareInterface
54{
55    private const REGEX_OCTET = '(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)';
56    private const REGEX_IPV4  = '/\\b' . self::REGEX_OCTET . '(?:\\.' . self::REGEX_OCTET . '){3}\\b/';
57
58    // Cache whois requests.  Try to avoid all caches expiring at the same time.
59    private const WHOIS_TTL_MIN = 28 * 86400;
60    private const WHOIS_TTL_MAX = 35 * 86400;
61    private const WHOIS_TIMEOUT = 5;
62
63    // Bad robots - SEO optimisers, advertisers, etc.  This list is shared with robots.txt.
64    public const BAD_ROBOTS = [
65        'admantx',
66        'Adsbot',
67        'AhrefsBot',
68        'Amazonbot', // Until it understands crawl-delay and noindex / nofollow
69        'AspiegelBot',
70        'Awario', // Brand management
71        'Barkrowler',
72        'BLEXBot',
73        'DataForSEO',
74        'DataForSeoBot', // https://dataforseo.com/dataforseo-bot
75        'DotBot',
76        'Grapeshot',
77        'Honolulu-bot', // Aggressive crawer, no info available
78        'ia_archiver',
79        'linabot', // Aggressive crawer, no info available
80        'Linguee',
81        'MegaIndex.ru',
82        'MJ12bot',
83        'netEstate NE',
84        'panscient',
85        'PetalBot',
86        'proximic',
87        'SeekportBot', // Pretends to be a search engine - but isn't
88        'SemrushBot',
89        'serpstatbot',
90        'SEOkicks',
91        'SiteKiosk',
92        'Turnitin',
93        'wp_is_mobile', // Nothing to do with wordpress
94        'XoviBot',
95        'ZoominfoBot',
96    ];
97
98    /**
99     * Some search engines use reverse/forward DNS to verify the IP address.
100     *
101     * @see https://developer.amazon.com/support/amazonbot
102     * @see https://support.google.com/webmasters/answer/80553?hl=en
103     * @see https://www.bing.com/webmaster/help/which-crawlers-does-bing-use-8c184ec0
104     * @see https://www.bing.com/webmaster/help/how-to-verify-bingbot-3905dc26
105     * @see https://yandex.com/support/webmaster/robot-workings/check-yandex-robots.html
106     * @see https://www.mojeek.com/bot.html
107     * @see https://support.apple.com/en-gb/HT204683
108     */
109    private const ROBOT_REV_FWD_DNS = [
110        'Amazonbot'        => ['.crawl.amazon.com'],
111        'Applebot'         => ['.applebot.apple.com'],
112        'BingPreview'      => ['.search.msn.com'],
113        'Google'           => ['.google.com', '.googlebot.com'],
114        'Mail.RU_Bot'      => ['.mail.ru'],
115        'MicrosoftPreview' => ['.search.msn.com'],
116        'MojeekBot'        => ['.mojeek.com'],
117        'Qwantify'         => ['.search.qwant.com'],
118        'Sogou'            => ['.crawl.sogou.com'],
119        'Yahoo'            => ['.crawl.yahoo.net'],
120        'Yandex'           => ['.yandex.ru', '.yandex.net', '.yandex.com'],
121        'bingbot'          => ['.search.msn.com'],
122        'msnbot'           => ['.search.msn.com'],
123    ];
124
125    /**
126     * Some search engines only use reverse DNS to verify the IP address.
127     *
128     * @see https://help.baidu.com/question?prod_id=99&class=0&id=3001
129     * @see https://napoveda.seznam.cz/en/full-text-search/seznambot-crawler
130     * @see https://www.ionos.de/terms-gtc/faq-crawler
131     */
132    private const ROBOT_REV_ONLY_DNS = [
133        'Baiduspider' => ['.baidu.com', '.baidu.jp'],
134        'FreshBot'    => ['.seznam.cz'],
135        'IonCrawl'    => ['.1und1.org'],
136        'Neevabot'    => ['.neeva.com'],
137    ];
138
139    /**
140     * Some search engines operate from designated IP addresses.
141     *
142     * @see https://www.apple.com/go/applebot
143     * @see https://help.duckduckgo.com/duckduckgo-help-pages/results/duckduckbot
144     */
145    private const ROBOT_IPS = [
146        'AppleBot'    => [
147            '17.0.0.0/8',
148        ],
149        'Ask Jeeves'  => [
150            '65.214.45.143',
151            '65.214.45.148',
152            '66.235.124.192',
153            '66.235.124.7',
154            '66.235.124.101',
155            '66.235.124.193',
156            '66.235.124.73',
157            '66.235.124.196',
158            '66.235.124.74',
159            '63.123.238.8',
160            '202.143.148.61',
161        ],
162        'DuckDuckBot' => [
163            '23.21.227.69',
164            '50.16.241.113',
165            '50.16.241.114',
166            '50.16.241.117',
167            '50.16.247.234',
168            '52.204.97.54',
169            '52.5.190.19',
170            '54.197.234.188',
171            '54.208.100.253',
172            '54.208.102.37',
173            '107.21.1.8',
174        ],
175    ];
176
177    /**
178     * Some search engines operate from designated IP addresses.
179     *
180     * @see https://bot.seekport.com/
181     */
182    private const ROBOT_IP_FILES = [
183        'SeekportBot' => 'https://bot.seekport.com/seekportbot_ips.txt',
184    ];
185
186    /**
187     * Some search engines operate from within a designated autonomous system.
188     *
189     * @see https://developers.facebook.com/docs/sharing/webmasters/crawler
190     * @see https://www.facebook.com/peering/
191     */
192    private const ROBOT_ASNS = [
193        'facebook' => ['AS32934', 'AS63293'],
194        'twitter'  => ['AS13414'],
195    ];
196
197    /**
198     * @param ServerRequestInterface  $request
199     * @param RequestHandlerInterface $handler
200     *
201     * @return ResponseInterface
202     */
203    public function process(ServerRequestInterface $request, RequestHandlerInterface $handler): ResponseInterface
204    {
205        $ua      = Validator::serverParams($request)->string('HTTP_USER_AGENT', '');
206        $ip      = Validator::attributes($request)->string('client-ip');
207        $address = IPFactory::parseAddressString($ip);
208        assert($address instanceof AddressInterface);
209
210        foreach (self::BAD_ROBOTS as $robot) {
211            if (str_contains($ua, $robot)) {
212                return $this->response();
213            }
214        }
215
216        foreach (self::ROBOT_REV_FWD_DNS as $robot => $valid_domains) {
217            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, false)) {
218                return $this->response();
219            }
220        }
221
222        foreach (self::ROBOT_REV_ONLY_DNS as $robot => $valid_domains) {
223            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, true)) {
224                return $this->response();
225            }
226        }
227
228        foreach (self::ROBOT_IPS as $robot => $valid_ip_ranges) {
229            if (str_contains($ua, $robot)) {
230                foreach ($valid_ip_ranges as $ip_range) {
231                    $range = IPFactory::parseRangeString($ip_range);
232
233                    if ($range instanceof RangeInterface && $range->contains($address)) {
234                        continue 2;
235                    }
236                }
237
238                return $this->response();
239            }
240        }
241
242        foreach (self::ROBOT_IP_FILES as $robot => $url) {
243            if (str_contains($ua, $robot)) {
244                $valid_ip_ranges = $this->fetchIpRangesForUrl($robot, $url);
245
246                foreach ($valid_ip_ranges as $ip_range) {
247                    $range = IPFactory::parseRangeString($ip_range);
248
249                    if ($range instanceof RangeInterface && $range->contains($address)) {
250                        continue 2;
251                    }
252                }
253
254                return $this->response();
255            }
256        }
257
258        foreach (self::ROBOT_ASNS as $robot => $asns) {
259            foreach ($asns as $asn) {
260                if (str_contains($ua, $robot)) {
261                    foreach ($this->fetchIpRangesForAsn($asn) as $range) {
262                        if ($range->contains($address)) {
263                            continue 2;
264                        }
265                    }
266
267                    return $this->response();
268                }
269            }
270        }
271
272        // Allow sites to block access from entire networks.
273        $block_asn = Validator::attributes($request)->string('block_asn', '');
274        preg_match_all('/(AS\d+)/', $block_asn, $matches);
275
276        foreach ($matches[1] as $asn) {
277            foreach ($this->fetchIpRangesForAsn($asn) as $range) {
278                if ($range->contains($address)) {
279                    return $this->response();
280                }
281            }
282        }
283
284        return $handler->handle($request);
285    }
286
287    /**
288     * Check that an IP address belongs to a robot operator using a forward/reverse DNS lookup.
289     *
290     * @param string        $ip
291     * @param array<string> $valid_domains
292     * @param bool          $reverse_only
293     *
294     * @return bool
295     */
296    private function checkRobotDNS(string $ip, array $valid_domains, bool $reverse_only): bool
297    {
298        $host = gethostbyaddr($ip);
299
300        if ($host === false) {
301            return false;
302        }
303
304        foreach ($valid_domains as $domain) {
305            if (str_ends_with($host, $domain)) {
306                return $reverse_only || $ip === gethostbyname($host);
307            }
308        }
309
310        return false;
311    }
312
313    /**
314     * Perform a whois search for an ASN.
315     *
316     * @param string $asn - The autonomous system number to query
317     *
318     * @return array<RangeInterface>
319     */
320    private function fetchIpRangesForAsn(string $asn): array
321    {
322        return Registry::cache()->file()->remember('whois-asn-' . $asn, static function () use ($asn): array {
323            $mapper = static fn (AsnRouteInfo $route_info): ?RangeInterface => IPFactory::parseRangeString($route_info->route ?: $route_info->route6);
324
325            try {
326                $loader = new CurlLoader(self::WHOIS_TIMEOUT);
327                $whois  = new Whois($loader);
328                $info   = $whois->loadAsnInfo($asn);
329                $routes = $info->routes;
330                $ranges = array_map($mapper, $routes);
331
332                return array_filter($ranges);
333            } catch (Throwable) {
334                return [];
335            }
336        }, random_int(self::WHOIS_TTL_MIN, self::WHOIS_TTL_MAX));
337    }
338
339    /**
340     * Fetch a list of IP addresses from a remote file.
341     *
342     * @param string $ua
343     * @param string $url
344     *
345     * @return array<string>
346     */
347    private function fetchIpRangesForUrl(string $ua, string $url): array
348    {
349        return Registry::cache()->file()->remember('url-ip-list-' . $ua, static function () use ($url): array {
350            try {
351                $client   = new Client();
352                $response = $client->get($url, ['timeout' => 5]);
353                $contents = $response->getBody()->getContents();
354
355                preg_match_all(self::REGEX_IPV4, $contents, $matches);
356
357                return $matches[0];
358            } catch (GuzzleException) {
359                return [];
360            }
361        }, random_int(self::WHOIS_TTL_MIN, self::WHOIS_TTL_MAX));
362    }
363
364    /**
365     * @return ResponseInterface
366     */
367    private function response(): ResponseInterface
368    {
369        return response('Not acceptable', StatusCodeInterface::STATUS_NOT_ACCEPTABLE);
370    }
371}
372