2 // $Id: spam.php,v 1.197 2007/07/03 14:51:07 henoheno Exp $
3 // Copyright (C) 2006-2007 PukiWiki Developers Team
4 // License: GPL v2 or (at your option) any later version
6 // Functions for Concept-work of spam-uri metrics
8 // (PHP 4 >= 4.3.0): preg_match_all(PREG_OFFSET_CAPTURE): $method['uri_XXX'] related feature
10 if (! defined('SPAM_INI_FILE')) define('SPAM_INI_FILE', 'spam.ini.php');
11 if (! defined('DOMAIN_INI_FILE')) define('DOMAIN_INI_FILE', 'domain.ini.php');
13 // ---------------------
16 // (PHP 4 >= 4.2.0): var_export(): mail-reporting and dump related
17 if (! function_exists('var_export')) {
18 function var_export() {
19 return 'var_export() is not found on this server' . "\n";
23 // (PHP 4 >= 4.2.0): preg_grep() enables invert option
24 function preg_grep_invert($pattern = '//', $input = array())
27 if (! isset($invert)) $invert = defined('PREG_GREP_INVERT');
30 return preg_grep($pattern, $input, PREG_GREP_INVERT);
32 $result = preg_grep($pattern, $input);
34 return array_diff($input, preg_grep($pattern, $input));
42 // ---------------------
45 // Very roughly, shrink the lines of var_export()
46 // NOTE: If the same data exists, it must be corrupted.
47 function var_export_shrink($expression, $return = FALSE, $ignore_numeric_keys = FALSE)
49 $result = var_export($expression, TRUE);
51 $result = preg_replace(
52 // Remove a newline and spaces
53 '# => \n *array \(#', ' => array (',
57 if ($ignore_numeric_keys) {
58 $result =preg_replace(
59 // Remove numeric keys
60 '#^( *)[0-9]+ => #m', '$1',
73 // Reverse $string with specified delimiter
74 function delimiter_reverse($string = 'foo.bar.example.com', $from_delim = '.', $to_delim = '.')
76 if (! is_string($string) || ! is_string($from_delim) || ! is_string($to_delim))
79 // com.example.bar.foo
80 return implode($to_delim, array_reverse(explode($from_delim, $string)));
84 function ksort_by_domain(& $array)
87 foreach(array_keys($array) as $key) {
88 $sort[delimiter_reverse($key)] = $key;
90 ksort($sort, SORT_STRING);
92 foreach($sort as $key) {
93 $result[$key] = & $array[$key];
98 // Roughly strings(1) using PCRE
99 // This function is useful to:
100 // * Reduce the size of data, from removing unprintable binary data
101 // * Detect _bare_strings_ from binary data
103 // http://www.freebsd.org/cgi/man.cgi?query=strings (Man-page of GNU strings)
104 // http://www.pcre.org/pcre.txt
105 // Note: mb_ereg_replace() is one of mbstring extension's functions
106 // and need to init its encoding.
107 function strings($binary = '', $min_len = 4, $ignore_space = FALSE, $multibyte = FALSE)
110 $binary = (is_array($binary) || $binary === TRUE) ? '' : strval($binary);
112 $regex = $ignore_space ?
113 '[^[:graph:] \t\n]+' : // Remove "\0" etc, and readable spaces
114 '[^[:graph:][:space:]]+'; // Preserve readable spaces if possible
116 $binary = $multibyte ?
117 mb_ereg_replace($regex, "\n", $binary) :
118 preg_replace('/' . $regex . '/s', "\n", $binary);
121 $binary = preg_replace(
136 // The last character seems "\n" or not
137 $br = (! empty($binary) && $binary[strlen($binary) - 1] == "\n") ? "\n" : '';
139 $min_len = min(1024, intval($min_len));
140 $regex = '/^.{' . $min_len . ',}/S';
141 $binary = implode("\n", preg_grep($regex, explode("\n", $binary))) . $br;
148 // ---------------------
151 // Count leaves (A leaf = value that is not an array, or an empty array)
152 function array_count_leaves($array = array(), $count_empty = FALSE)
154 if (! is_array($array) || (empty($array) && $count_empty)) return 1;
158 foreach ($array as $part) {
159 $count += array_count_leaves($part, $count_empty);
164 // An array-leaves to a flat array
165 function array_flat_leaves($array, $unique = TRUE)
167 if (! is_array($array)) return $array;
170 foreach(array_keys($array) as $key) {
171 if (is_array($array[$key])) {
173 foreach(array_flat_leaves($array[$key]) as $_value) {
177 $tmp[] = & $array[$key];
181 return $unique ? array_values(array_unique($tmp)) : $tmp;
184 // $array['something'] => $array['wanted']
185 function array_rename_keys(& $array, $keys = array('from' => 'to'), $force = FALSE, $default = '')
187 if (! is_array($array) || ! is_array($keys)) return FALSE;
189 // Nondestructive test
191 foreach(array_keys($keys) as $from)
192 if (! isset($array[$from]))
195 foreach($keys as $from => $to) {
196 if ($from === $to) continue;
197 if (! $force || isset($array[$from])) {
198 $array[$to] = & $array[$from];
199 unset($array[$from]);
201 $array[$to] = $default;
208 // Remove redundant values from array()
209 function array_unique_recursive($array = array())
211 if (! is_array($array)) return $array;
214 foreach($array as $key => $value){
215 if (is_array($value)) {
216 $array[$key] = array_unique_recursive($value);
218 if (isset($tmp[$value])) {
230 // ---------------------
231 // Part One : Checker
233 // Rough implementation of globbing
235 // USAGE: $regex = '/^' . generate_glob_regex('*.txt', '/') . '$/i';
237 function generate_glob_regex($string = '', $divider = '/')
239 static $from = array(
242 // 22 => '[', // Maybe cause regex compilation error (e.g. '[]')
258 if (! is_string($string)) return '';
260 $string = str_replace($from, $mid, $string); // Hide
261 $string = preg_quote($string, $divider);
262 $string = str_replace($mid, $to, $string); // Unhide
267 // Generate host (FQDN, IPv4, ...) regex
268 // 'localhost' : Matches with 'localhost' only
269 // 'example.org' : Matches with 'example.org' only (See host_normalize() about 'www')
270 // '.example.org' : Matches with ALL FQDN ended with '.example.org'
271 // '*.example.org' : Almost the same of '.example.org' except 'www.example.org'
272 // '10.20.30.40' : Matches with IPv4 address '10.20.30.40' only
273 // [TODO] '192.' : Matches with all IPv4 hosts started with '192.'
274 // TODO: IPv4, CIDR?, IPv6
275 function generate_host_regex($string = '', $divider = '/')
277 if (! is_string($string)) return '';
279 if (mb_strpos($string, '.') === FALSE)
280 return generate_glob_regex($string, $divider);
283 if (is_ip($string)) {
285 return generate_glob_regex($string, $divider);
288 $part = explode('.', $string, 2);
289 if ($part[0] == '') {
290 $part[0] = '(?:.*\.)?'; // And all related FQDN
291 } else if ($part[0] == '*') {
292 $part[0] = '.*\.'; // All subdomains/hosts only
294 return generate_glob_regex($string, $divider);
296 $part[1] = generate_glob_regex($part[1], $divider);
297 return implode('', $part);
301 // Rough hostname checker
303 // TODO: Strict digit, 0x, CIDR, IPv6
304 function is_ip($string = '')
306 if (preg_match('/^' .
307 '(?:[0-9]{1,3}\.){3}[0-9]{1,3}' . '|' .
308 '(?:[0-9]{1,3}\.){1,3}' . '$/',
310 return 4; // Seems IPv4(dot-decimal)
312 return 0; // Seems not IP
316 function get_blocklist($list = '')
320 if ($list === NULL) {
321 $regexes = NULL; // Unset
325 if (! isset($regexes)) {
327 if (file_exists(SPAM_INI_FILE)) {
328 $blocklist = array();
329 include(SPAM_INI_FILE);
330 // $blocklist['badhost'] = array(
331 // '*.blogspot.com', // Blog services's subdomains (only)
332 // 'IANA-examples' => '#^(?:.*\.)?example\.(?:com|net|org)$#',
334 if (isset($blocklist['list'])) {
335 $regexes['list'] = & $blocklist['list'];
338 $blocklist['list'] = array(
343 foreach(array_keys($blocklist['list']) as $_list) {
344 if (! isset($blocklist[$_list])) continue;
345 foreach ($blocklist[$_list] as $key => $value) {
346 if (is_array($value)) {
347 $regexes[$_list][$key] = array();
348 foreach($value as $_key => $_value) {
349 get_blocklist_add($regexes[$_list][$key], $_key, $_value);
352 get_blocklist_add($regexes[$_list], $key, $value);
355 unset($blocklist[$_list]);
361 return $regexes; // ALL
362 } else if (isset($regexes[$list])) {
363 return $regexes[$list];
369 // Subroutine of get_blocklist()
370 function get_blocklist_add(& $array, $key = 0, $value = '*.example.org')
372 if (is_string($key)) {
373 $array[$key] = & $value; // Treat $value as a regex
375 $array[$value] = '/^' . generate_host_regex($value, '/') . '$/i';
379 // Blocklist metrics: Separate $host, to $blocked and not blocked
380 function blocklist_distiller(& $hosts, $keys = array('goodhost', 'badhost'), $asap = FALSE)
382 if (! is_array($hosts)) $hosts = array($hosts);
383 if (! is_array($keys)) $keys = array($keys);
385 $list = get_blocklist('list');
388 foreach($keys as $key){
389 foreach (get_blocklist($key) as $label => $regex) {
390 if (is_array($regex)) {
391 foreach($regex as $_label => $_regex) {
392 $group = preg_grep($_regex, $hosts);
394 $hosts = array_diff($hosts, $group);
395 $blocked[$key][$label][$_label] = $group;
396 if ($asap && $list[$key]) break;
400 $group = preg_grep($regex, $hosts);
402 $hosts = array_diff($hosts, $group);
403 $blocked[$key][$label] = $group;
404 if ($asap && $list[$key]) break;
414 // ---------------------
417 // Default (enabled) methods and thresholds (for content insertion)
418 function check_uri_spam_method($times = 1, $t_area = 0, $rule = TRUE)
420 $times = intval($times);
421 $t_area = intval($t_area);
425 'quantity' => 8 * $times, // Allow N URIs
426 'non_uniqhost' => 3 * $times, // Allow N duped (and normalized) Hosts
427 //'non_uniquri'=> 3 * $times, // Allow N duped (and normalized) URIs
430 'area_anchor' => $t_area, // Using <a href> HTML tag
431 'area_bbcode' => $t_area, // Using [url] or [link] BBCode
432 //'uri_anchor' => $t_area, // URI inside <a href> HTML tag
433 //'uri_bbcode' => $t_area, // URI inside [url] or [link] BBCode
438 //'asap' => TRUE, // Quit or return As Soon As Possible
439 'uniqhost' => TRUE, // Show uniq host (at block notification mail)
440 'badhost' => TRUE, // Check badhost
446 // Remove non-$positive values
447 foreach (array_keys($positive) as $key) {
448 if ($positive[$key] < 0) unset($positive[$key]);
451 return $positive + $bool;
454 // Simple/fast spam check
455 function check_uri_spam($target = '', $method = array())
460 // Theme to do => Dummy, optional value, or optional array()
462 //'uniqhost' => TRUE,
463 //'non_uniqhost'=> 3,
464 //'non_uniquri' => 3,
466 //'area_anchor' => 0,
467 //'area_bbcode' => 0,
472 // Theme => Volume found (int)
475 // Flag. If someting defined here,
476 // one or more spam will be included
481 //'category' => array(
491 $sum = & $progress['sum'];
492 $is_spam = & $progress['is_spam'];
493 $progress['method'] = & $method; // Argument
494 $blocked = & $progress['blocked'];
495 $hosts = & $progress['hosts'];
496 $asap = isset($method['asap']);
499 if (! is_array($method) || empty($method)) {
500 $method = check_uri_spam_method();
502 foreach(array_keys($method) as $key) {
503 if (! isset($sum[$key])) $sum[$key] = 0;
505 if (! isset($sum['quantity'])) $sum['quantity'] = 0;
507 if (is_array($target)) {
508 foreach($target as $str) {
509 if (! is_string($str)) continue;
511 $_progress = check_uri_spam($str, $method); // Recurse
514 $_sum = & $_progress['sum'];
515 foreach (array_keys($_sum) as $key) {
516 if (! isset($sum[$key])) {
517 $sum[$key] = & $_sum[$key];
519 $sum[$key] += $_sum[$key];
524 $_is_spam = & $_progress['is_spam'];
525 foreach (array_keys($_is_spam) as $key) {
526 $is_spam[$key] = TRUE;
529 if ($asap && $is_spam) break;
532 $blocked = array_merge_recursive($blocked, $_progress['blocked']);
533 $hosts = array_merge_recursive($hosts, $_progress['hosts']);
537 $blocked = array_unique_recursive($blocked);
538 $hosts = array_unique_recursive($hosts);
540 // Recount $sum['badhost']
541 $sum['badhost'] = array_count_leaves($blocked);
546 // Area: There's HTML anchor tag
547 if ((! $asap || ! $is_spam) && isset($method['area_anchor'])) {
548 $key = 'area_anchor';
549 $_asap = isset($method['asap']) ? array('asap' => TRUE) : array();
550 $result = area_pickup($target, array($key => TRUE) + $_asap);
552 $sum[$key] = $result[$key];
553 if (isset($method[$key]) && $sum[$key] > $method[$key]) {
554 $is_spam[$key] = TRUE;
559 // Area: There's 'BBCode' linking tag
560 if ((! $asap || ! $is_spam) && isset($method['area_bbcode'])) {
561 $key = 'area_bbcode';
562 $_asap = isset($method['asap']) ? array('asap' => TRUE) : array();
563 $result = area_pickup($target, array($key => TRUE) + $_asap);
565 $sum[$key] = $result[$key];
566 if (isset($method[$key]) && $sum[$key] > $method[$key]) {
567 $is_spam[$key] = TRUE;
573 if ($asap && $is_spam) return $progress;
576 $pickups = uri_pickup_normalize(spam_uri_pickup($target, $method));
579 if (empty($pickups)) return $progress;
581 // URI: Check quantity
582 $sum['quantity'] += count($pickups);
584 if ((! $asap || ! $is_spam) && isset($method['quantity']) &&
585 $sum['quantity'] > $method['quantity']) {
586 $is_spam['quantity'] = TRUE;
589 // URI: used inside HTML anchor tag pair
590 if ((! $asap || ! $is_spam) && isset($method['uri_anchor'])) {
592 foreach($pickups as $pickup) {
593 if (isset($pickup['area'][$key])) {
594 $sum[$key] += $pickup['area'][$key];
595 if(isset($method[$key]) &&
596 $sum[$key] > $method[$key]) {
597 $is_spam[$key] = TRUE;
598 if ($asap && $is_spam) break;
600 if ($asap && $is_spam) break;
605 // URI: used inside 'BBCode' pair
606 if ((! $asap || ! $is_spam) && isset($method['uri_bbcode'])) {
608 foreach($pickups as $pickup) {
609 if (isset($pickup['area'][$key])) {
610 $sum[$key] += $pickup['area'][$key];
611 if(isset($method[$key]) &&
612 $sum[$key] > $method[$key]) {
613 $is_spam[$key] = TRUE;
614 if ($asap && $is_spam) break;
616 if ($asap && $is_spam) break;
621 // URI: Uniqueness (and removing non-uniques)
622 if ((! $asap || ! $is_spam) && isset($method['non_uniquri'])) {
625 foreach (array_keys($pickups) as $key) {
626 $uris[$key] = uri_pickup_implode($pickups[$key]);
628 $count = count($uris);
629 $uris = array_unique($uris);
630 $sum['non_uniquri'] += $count - count($uris);
631 if ($sum['non_uniquri'] > $method['non_uniquri']) {
632 $is_spam['non_uniquri'] = TRUE;
634 if (! $asap || ! $is_spam) {
635 foreach (array_diff(array_keys($pickups),
636 array_keys($uris)) as $remove) {
637 unset($pickups[$remove]);
644 if ($asap && $is_spam) return $progress;
646 // Host: Uniqueness (uniq / non-uniq)
647 foreach ($pickups as $pickup) $hosts[] = & $pickup['host'];
648 $hosts = array_unique($hosts);
649 $sum['uniqhost'] += count($hosts);
650 if ((! $asap || ! $is_spam) && isset($method['non_uniqhost'])) {
651 $sum['non_uniqhost'] = $sum['quantity'] - $sum['uniqhost'];
652 if ($sum['non_uniqhost'] > $method['non_uniqhost']) {
653 $is_spam['non_uniqhost'] = TRUE;
658 if ($asap && $is_spam) return $progress;
660 // URI: Bad host (Separate good/bad hosts from $hosts)
661 if ((! $asap || ! $is_spam) && isset($method['badhost'])) {
664 $list = get_blocklist('list');
665 $blocked = blocklist_distiller($hosts, array_keys($list), $asap);
666 foreach($list as $key=>$type){
667 if (! $type) unset($blocked[$key]); // Ignore goodhost etc
671 if (! empty($blocked)) $is_spam['badhost'] = TRUE;
677 // ---------------------
680 // Summarize $progress (blocked only)
681 function summarize_spam_progress($progress = array(), $blockedonly = FALSE)
684 $tmp = array_keys($progress['is_spam']);
687 $method = & $progress['method'];
688 if (isset($progress['sum'])) {
689 foreach ($progress['sum'] as $key => $value) {
690 if (isset($method[$key]) && $value) {
691 $tmp[] = $key . '(' . $value . ')';
697 return implode(', ', $tmp);
700 function summarize_detail_badhost($progress = array())
702 if (! isset($progress['blocked']) || empty($progress['blocked'])) return '';
706 foreach($progress['blocked'] as $list => $lvalue) {
707 foreach($lvalue as $group => $gvalue) {
708 $flat = implode(', ', array_flat_leaves($gvalue));
709 if ($flat === $group) {
710 $blocked[$list][] = $flat;
712 $blocked[$list][$group] = $flat;
718 // From: 'A-1' => array('ie.to')
719 // To: 'A-1' => 'ie.to'
720 foreach($blocked as $list => $lvalue) {
721 if (is_array($lvalue) &&
722 count($lvalue) == 1 &&
723 is_numeric(key($lvalue))) {
724 $blocked[$list] = current($lvalue);
728 return var_export_shrink($blocked, TRUE, TRUE);
731 function summarize_detail_newtral($progress = array())
733 if (! isset($progress['hosts']) ||
734 ! is_array($progress['hosts']) ||
735 empty($progress['hosts'])) return '';
737 // Generate a responsible $trie
739 foreach($progress['hosts'] as $value) {
740 // 'A.foo.bar.example.com'
741 $resp = whois_responsibility($value); // 'example.com'
743 // One or more test, or do nothing here
744 $resp = strval($value);
747 $rest = rtrim(substr($value, 0, - strlen($resp)), '.'); // 'A.foo.bar'
749 $trie = array_merge_recursive($trie, array($resp => array($rest => NULL)));
752 // Format: var_export_shrink() -like output
754 ksort_by_domain($trie);
755 foreach(array_keys($trie) as $key) {
756 ksort_by_domain($trie[$key]);
757 if (count($trie[$key]) == 1 && key($trie[$key]) == '') {
758 // Just one 'responsibility.example.com'
759 $result[] = ' \'' . $key . '\',';
761 // One subdomain-or-host, or several ones
763 foreach(array_keys($trie[$key]) as $sub) {
767 $subs[] = $sub . '.' . $key;
770 $result[] = ' \'' . $key . '\' => \'' . implode(', ', $subs) . '\',';
776 implode("\n", $result) . "\n" .
781 // Check responsibility-root of the FQDN
782 // 'foo.bar.example.com' => 'example.com' (.com has the last whois for it)
783 // 'foo.bar.example.au' => 'example.au' (.au has the last whois for it)
784 // 'foo.bar.example.edu.au' => 'example.edu.au' (.edu.au has the last whois for it)
785 // 'foo.bar.example.act.edu.au' => 'example.act.edu.au' (.act.edu.au has the last whois for it)
786 function whois_responsibility($fqdn = 'foo.bar.example.com', $parent = FALSE, $implicit = TRUE)
790 if ($fqdn === NULL) {
791 $domain = NULL; // Unset
794 if (! is_string($fqdn)) return '';
796 if (is_ip($fqdn)) return $fqdn;
798 if (! isset($domain)) {
800 if (file_exists(DOMAIN_INI_FILE)) {
801 include(DOMAIN_INI_FILE); // Set
806 $dcursor = & $domain;
807 $array = array_reverse(explode('.', $fqdn));
810 if (! isset($array[$i])) break;
811 $acursor = $array[$i];
812 if (is_array($dcursor) && isset($dcursor[$acursor])) {
813 $result[] = & $array[$i];
814 $dcursor = & $dcursor[$acursor];
816 if (! $parent && isset($acursor)) {
817 $result[] = & $array[$i]; // Whois servers must know this subdomain
824 // Implicit responsibility: Top-Level-Domains must not be yours
825 // 'bar.foo.something' => 'foo.something'
826 if ($implicit && count($result) == 1 && count($array) > 1) {
827 $result[] = & $array[1];
830 return $result ? implode('.', array_reverse($result)) : '';
834 // ---------------------
838 function spam_dispose()
841 whois_responsibility(NULL);
844 // Common bahavior for blocking
845 // NOTE: Call this function from various blocking feature, to disgueise the reason 'why blocked'
846 function spam_exit($mode = '', $data = array())
855 echo('<pre>' . "\n");
856 echo htmlspecialchars(var_export($data, TRUE));
857 echo('</pre>' . "\n");
861 if ($exit) exit; // Force exit
865 // ---------------------
869 // Simple/fast spam filter ($target: 'a string' or an array())
870 function pkwk_spamfilter($action, $page, $target = array('title' => ''), $method = array(), $exitmode = '')
872 $progress = check_uri_spam($target, $method);
874 if (empty($progress['is_spam'])) {
878 // TODO: detect encoding from $target for mbstring functions
880 // foreach(array_keys($target) as $key) {
881 // $tmp[strings($key, 0, FALSE, TRUE)] = strings($target[$key], 0, FALSE, TRUE); // Removing "\0" etc
885 pkwk_spamnotify($action, $page, $target, $progress, $method);
886 spam_exit($exitmode, $progress);
890 // ---------------------
893 // Mail to administrator(s)
894 function pkwk_spamnotify($action, $page, $target = array('title' => ''), $progress = array(), $method = array())
896 global $notify, $notify_subject;
898 if (! $notify) return;
900 $asap = isset($method['asap']);
902 $summary['ACTION'] = 'Blocked by: ' . summarize_spam_progress($progress, TRUE);
904 $summary['METRICS'] = summarize_spam_progress($progress);
907 $tmp = summarize_detail_badhost($progress);
908 if ($tmp != '') $summary['DETAIL_BADHOST'] = $tmp;
910 $tmp = summarize_detail_newtral($progress);
911 if (! $asap && $tmp != '') $summary['DETAIL_NEUTRAL_HOST'] = $tmp;
913 $summary['COMMENT'] = $action;
914 $summary['PAGE'] = '[blocked] ' . (is_pagename($page) ? $page : '');
915 $summary['URI'] = get_script_uri() . '?' . rawurlencode($page);
916 $summary['USER_AGENT'] = TRUE;
917 $summary['REMOTE_ADDR'] = TRUE;
918 pkwk_mail_notify($notify_subject, var_export($target, TRUE), $summary, TRUE);