xref: /webtrees/app/Http/Middleware/BadBotBlocker.php (revision 73b8c04073f4aae42af235f4485847d0a07c8f8c) 
1<?php
2
3/**
4 * webtrees: online genealogy
5 * Copyright (C) 2020 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 <http://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 Iodev\Whois\Loaders\CurlLoader;
25use Iodev\Whois\Modules\Asn\AsnRouteInfo;
26use Iodev\Whois\Whois;
27use IPLib\Address\AddressInterface;
28use IPLib\Factory as IPFactory;
29use IPLib\Range\RangeInterface;
30use Psr\Http\Message\ResponseInterface;
31use Psr\Http\Message\ServerRequestInterface;
32use Psr\Http\Server\MiddlewareInterface;
33use Psr\Http\Server\RequestHandlerInterface;
34use Throwable;
35
36use function array_map;
37use function assert;
38use function gethostbyaddr;
39use function gethostbyname;
40use function response;
41use function str_contains;
42use function str_ends_with;
43
44/**
45 * Middleware to block bad robots before they waste our valuable CPU cycles.
46 */
47class BadBotBlocker implements MiddlewareInterface
48{
49    // Cache whois requests.  Try to avoid all caches expiring at the same time.
50    private const WHOIS_TTL_MIN = 28 * 86400;
51    private const WHOIS_TTL_MAX = 35 * 86400;
52    private const WHOIS_TIMEOUT = 5;
53
54    // Bad robots - SEO optimisers, advertisers, etc
55    private const BAD_ROBOTS = [
56        'admantx',
57        'AhrefsBot',
58        'AspiegelBot',
59        'DotBot',
60        'Grapeshot',
61        'ia_archiver',
62        'MJ12bot',
63        'panscient',
64        'proximic',
65        'SemrushBot',
66        'XoviBot',
67    ];
68
69    /**
70     * Some search engines use reverse/forward DNS to verify the IP address.
71     *
72     * @see https://support.google.com/webmasters/answer/80553?hl=en
73     * @see https://www.bing.com/webmaster/help/which-crawlers-does-bing-use-8c184ec0
74     * @see https://www.bing.com/webmaster/help/how-to-verify-bingbot-3905dc26
75     * @see https://yandex.com/support/webmaster/robot-workings/check-yandex-robots.html
76     */
77    private const ROBOT_REV_FWD_DNS = [
78        'bingbot'     => ['.search.msn.com'],
79        'BingPreview' => ['.search.msn.com'],
80        'Google'      => ['.google.com', '.googlebot.com'],
81        'msnbot'      => ['.search.msn.com'],
82        'Qwantify'    => ['.search.qwant.com'],
83        'Sogou'       => ['.crawl.sogou.com'],
84        'Yahoo'       => ['.crawl.yahoo.net'],
85        'Yandex'      => ['.yandex.ru', '.yandex.net', '.yandex.com'],
86    ];
87
88    /**
89     * Some search engines only use reverse DNS to verify the IP address.
90     *
91     * @see https://help.baidu.com/question?prod_id=99&class=0&id=3001
92     */
93    private const ROBOT_REV_ONLY_DNS = [
94        'Baiduspider' => ['.baidu.com', '.baidu.jp'],
95    ];
96
97    /**
98     * Some search engines operate from designated IP addresses.
99     *
100     * @see http://www.apple.com/go/applebot
101     * @see https://help.duckduckgo.com/duckduckgo-help-pages/results/duckduckbot
102     */
103    private const ROBOT_IPS = [
104        'AppleBot'    => [
105            '17.0.0.0/8',
106        ],
107        'Ask Jeeves'  => [
108            '65.214.45.143',
109            '65.214.45.148',
110            '66.235.124.192',
111            '66.235.124.7',
112            '66.235.124.101',
113            '66.235.124.193',
114            '66.235.124.73',
115            '66.235.124.196',
116            '66.235.124.74',
117            '63.123.238.8',
118            '202.143.148.61',
119        ],
120        'DuckDuckBot' => [
121            '23.21.227.69',
122            '50.16.241.113',
123            '50.16.241.114',
124            '50.16.241.117',
125            '50.16.247.234',
126            '52.204.97.54',
127            '52.5.190.19',
128            '54.197.234.188',
129            '54.208.100.253',
130            '54.208.102.37',
131            '107.21.1.8',
132        ],
133    ];
134
135    /**
136     * Some search engines operate from within a designated autonomous system.
137     *
138     * @see https://developers.facebook.com/docs/sharing/webmasters/crawler
139     */
140    private const ROBOT_ASN = [
141        'facebook' => 'AS32934',
142        'twitter'  => 'AS13414',
143    ];
144
145    /**
146     * @param ServerRequestInterface  $request
147     * @param RequestHandlerInterface $handler
148     *
149     * @return ResponseInterface
150     */
151    public function process(ServerRequestInterface $request, RequestHandlerInterface $handler): ResponseInterface
152    {
153        $ua      = $request->getServerParams()['HTTP_USER_AGENT'] ?? '';
154        $ip      = $request->getAttribute('client-ip');
155        $address = IPFactory::addressFromString($ip);
156        assert($address instanceof AddressInterface);
157
158        foreach (self::BAD_ROBOTS as $robot) {
159            if (str_contains($ua, $robot)) {
160                return $this->response();
161            }
162        }
163
164        foreach (self::ROBOT_REV_FWD_DNS as $robot => $valid_domains) {
165            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, false)) {
166                return $this->response();
167            }
168        }
169
170        foreach (self::ROBOT_REV_ONLY_DNS as $robot => $valid_domains) {
171            if (str_contains($ua, $robot) && !$this->checkRobotDNS($ip, $valid_domains, true)) {
172                return $this->response();
173            }
174        }
175
176        foreach (self::ROBOT_IPS as $robot => $valid_ips) {
177            if (str_contains($ua, $robot)) {
178                foreach ($valid_ips as $ip) {
179                    $range = IPFactory::rangeFromString($ip);
180
181                    if ($range instanceof RangeInterface && $range->contains($address)) {
182                        continue 2;
183                    }
184                }
185
186                return $this->response();
187            }
188        }
189
190        foreach (self::ROBOT_ASN as $robot => $asn) {
191            if (str_contains($ua, $robot)) {
192                foreach ($this->fetchIpRangesForAsn($asn) as $range) {
193                    if ($range->contains($address)) {
194                        continue 2;
195                    }
196                }
197
198                return $this->response();
199            }
200        }
201
202        // Allow sites to block access from entire networks.
203        preg_match_all('/(AS\d+)/', $request->getAttribute('block_asn', ''), $matches);
204        foreach ($matches[1] as $asn) {
205            foreach ($this->fetchIpRangesForAsn($asn) as $range) {
206                if ($range->contains($address)) {
207                    return $this->response();
208                }
209            }
210        }
211
212        return $handler->handle($request);
213    }
214
215    /**
216     * Check that an IP address belongs to a robot operator using a forward/reverse DNS lookup.
217     *
218     * @param string        $ip
219     * @param array<string> $valid_domains
220     * @param bool          $reverse_only
221     *
222     * @return bool
223     */
224    private function checkRobotDNS(string $ip, array $valid_domains, bool $reverse_only): bool
225    {
226        $host = gethostbyaddr($ip);
227
228        if ($host === false) {
229            return false;
230        }
231
232        foreach ($valid_domains as $domain) {
233            if (str_ends_with($host, $domain)) {
234                return $reverse_only || $ip === gethostbyname($host);
235            }
236        }
237
238        return false;
239    }
240
241    /**
242     * Perform a whois search for an ASN.
243     *
244     * @param string $asn - The autonomous system number to query
245     *
246     * @return array<RangeInterface>
247     */
248    private function fetchIpRangesForAsn(string $asn): array
249    {
250        return Registry::cache()->file()->remember('whois-asn-' . $asn, static function () use ($asn): array {
251            try {
252                $loader = new CurlLoader(self::WHOIS_TIMEOUT);
253                $whois  = new Whois($loader);
254                $info   = $whois->loadAsnInfo($asn);
255                $routes = $info->getRoutes();
256                $ranges = array_map(static function (AsnRouteInfo $route_info): ?RangeInterface {
257                    return IPFactory::rangeFromString($route_info->getRoute() ?: $route_info->getRoute6());
258                }, $routes);
259
260                return array_filter($ranges);
261            } catch (Throwable $ex) {
262                return [];
263            }
264        }, random_int(self::WHOIS_TTL_MIN, self::WHOIS_TTL_MAX));
265    }
266
267    /**
268     * @return ResponseInterface
269     */
270    private function response(): ResponseInterface
271    {
272        return response('Not acceptable', StatusCodeInterface::STATUS_NOT_ACCEPTABLE);
273    }
274}
275