|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
declare(strict_types = 1); |
|
4
|
|
|
|
|
5
|
|
|
namespace App\Helper; |
|
6
|
|
|
|
|
7
|
|
|
use App\Model\Project; |
|
8
|
|
|
use DateInterval; |
|
9
|
|
|
use MediaWiki\OAuthClient\Client; |
|
10
|
|
|
use Psr\Cache\CacheItemPoolInterface; |
|
11
|
|
|
use Symfony\Component\HttpFoundation\RequestStack; |
|
12
|
|
|
use Symfony\Component\HttpFoundation\Session\SessionInterface; |
|
13
|
|
|
|
|
14
|
|
|
/** |
|
15
|
|
|
* Helper class for fetching semi-automated definitions. |
|
16
|
|
|
*/ |
|
17
|
|
|
class AutomatedEditsHelper |
|
18
|
|
|
{ |
|
19
|
|
|
protected CacheItemPoolInterface $cache; |
|
20
|
|
|
protected SessionInterface $session; |
|
21
|
|
|
|
|
22
|
|
|
/** @var array The list of tools that are considered reverting. */ |
|
23
|
|
|
protected array $revertTools = []; |
|
24
|
|
|
|
|
25
|
|
|
/** @var array The list of tool names and their regexes/tags. */ |
|
26
|
|
|
protected array $tools = []; |
|
27
|
|
|
|
|
28
|
|
|
/** |
|
29
|
|
|
* AutomatedEditsHelper constructor. |
|
30
|
|
|
* @param RequestStack $requestStack |
|
31
|
|
|
* @param CacheItemPoolInterface $cache |
|
32
|
|
|
*/ |
|
33
|
|
|
public function __construct(RequestStack $requestStack, CacheItemPoolInterface $cache) |
|
34
|
|
|
{ |
|
35
|
|
|
$this->session = $requestStack->getSession(); |
|
36
|
|
|
$this->cache = $cache; |
|
37
|
|
|
} |
|
38
|
|
|
|
|
39
|
|
|
/** |
|
40
|
|
|
* Get the tool that matched the given edit summary. |
|
41
|
|
|
* This only works for tools defined with regular expressions, not tags. |
|
42
|
|
|
* @param string $summary Edit summary |
|
43
|
|
|
* @param Project $project |
|
44
|
|
|
* @return string[]|null Tool entry including key for 'name', or false if nothing was found |
|
45
|
|
|
*/ |
|
46
|
|
|
public function getTool(string $summary, Project $project): ?array |
|
47
|
|
|
{ |
|
48
|
|
|
foreach ($this->getTools($project) as $tool => $values) { |
|
49
|
|
|
if (isset($values['regex']) && preg_match('/'.$values['regex'].'/', $summary)) { |
|
50
|
|
|
return array_merge([ |
|
51
|
|
|
'name' => $tool, |
|
52
|
|
|
], $values); |
|
53
|
|
|
} |
|
54
|
|
|
} |
|
55
|
|
|
|
|
56
|
|
|
return null; |
|
57
|
|
|
} |
|
58
|
|
|
|
|
59
|
|
|
/** |
|
60
|
|
|
* Was the edit (semi-)automated, based on the edit summary? |
|
61
|
|
|
* This only works for tools defined with regular expressions, not tags. |
|
62
|
|
|
* @param string $summary Edit summary |
|
63
|
|
|
* @param Project $project |
|
64
|
|
|
* @return bool |
|
65
|
|
|
*/ |
|
66
|
|
|
public function isAutomated(string $summary, Project $project): bool |
|
67
|
|
|
{ |
|
68
|
|
|
return (bool)$this->getTool($summary, $project); |
|
69
|
|
|
} |
|
70
|
|
|
|
|
71
|
|
|
/** |
|
72
|
|
|
* Fetch the config from https://meta.wikimedia.org/wiki/MediaWiki:XTools-AutoEdits.json |
|
73
|
|
|
* @param bool $useSandbox Use the sandbox version of the config, located at MediaWiki:XTools-AutoEdits.json/sandbox |
|
74
|
|
|
* @return array |
|
75
|
|
|
*/ |
|
76
|
|
|
public function getConfig(bool $useSandbox = false): array |
|
77
|
|
|
{ |
|
78
|
|
|
$cacheKey = 'autoedits_config'; |
|
79
|
|
|
if (!$useSandbox && $this->cache->hasItem($cacheKey)) { |
|
80
|
|
|
return $this->cache->getItem($cacheKey)->get(); |
|
81
|
|
|
} |
|
82
|
|
|
|
|
83
|
|
|
$uri = 'https://meta.wikimedia.org/w/api.php?' . http_build_query([ |
|
84
|
|
|
'action' => 'query', |
|
85
|
|
|
'prop' => 'revisions', |
|
86
|
|
|
'rvprop' => 'content', |
|
87
|
|
|
'rvslots' => 'main', |
|
88
|
|
|
'format' => 'json', |
|
89
|
|
|
'formatversion' => 2, |
|
90
|
|
|
'titles' => 'MediaWiki:XTools-AutoEdits.json' . ($useSandbox ? '/sandbox' : ''), |
|
91
|
|
|
]); |
|
92
|
|
|
|
|
93
|
|
|
if ($useSandbox && $this->session->get('logged_in_user')) { |
|
94
|
|
|
// Request via OAuth to get around server-side caching. |
|
95
|
|
|
/** @var Client $client */ |
|
96
|
|
|
$client = $this->session->get('oauth_client'); |
|
97
|
|
|
$resp = json_decode($client->makeOAuthCall( |
|
98
|
|
|
$this->session->get('oauth_access_token'), |
|
99
|
|
|
$uri |
|
100
|
|
|
)); |
|
101
|
|
|
} else { |
|
102
|
|
|
$resp = json_decode(file_get_contents($uri)); |
|
103
|
|
|
} |
|
104
|
|
|
|
|
105
|
|
|
$ret = json_decode($resp->query->pages[0]->revisions[0]->slots->main->content, true); |
|
106
|
|
|
|
|
107
|
|
|
if (!$useSandbox) { |
|
108
|
|
|
$cacheItem = $this->cache |
|
109
|
|
|
->getItem($cacheKey) |
|
110
|
|
|
->set($ret) |
|
111
|
|
|
->expiresAfter(new DateInterval('PT20M')); |
|
112
|
|
|
$this->cache->save($cacheItem); |
|
113
|
|
|
} |
|
114
|
|
|
|
|
115
|
|
|
return $ret; |
|
116
|
|
|
} |
|
117
|
|
|
|
|
118
|
|
|
/** |
|
119
|
|
|
* Get list of automated tools and their associated info for the given project. |
|
120
|
|
|
* This defaults to the DEFAULT_PROJECT if entries for the given project are not found. |
|
121
|
|
|
* @param Project $project |
|
122
|
|
|
* @param bool $useSandbox Whether to use the /sandbox version for testing (also bypasses caching). |
|
123
|
|
|
* @return array Each tool with the tool name as the key and 'link', 'regex' and/or 'tag' as the subarray keys. |
|
124
|
|
|
*/ |
|
125
|
|
|
public function getTools(Project $project, bool $useSandbox = false): array |
|
126
|
|
|
{ |
|
127
|
|
|
$projectDomain = $project->getDomain(); |
|
128
|
|
|
|
|
129
|
|
|
if (isset($this->tools[$projectDomain])) { |
|
130
|
|
|
return $this->tools[$projectDomain]; |
|
131
|
|
|
} |
|
132
|
|
|
|
|
133
|
|
|
// Load the semi-automated edit types. |
|
134
|
|
|
$tools = $this->getConfig($useSandbox); |
|
135
|
|
|
|
|
136
|
|
|
if (isset($tools[$projectDomain])) { |
|
137
|
|
|
$localRules = $tools[$projectDomain]; |
|
138
|
|
|
} else { |
|
139
|
|
|
$localRules = []; |
|
140
|
|
|
} |
|
141
|
|
|
|
|
142
|
|
|
$langRules = $tools[$project->getLang()] ?? []; |
|
143
|
|
|
|
|
144
|
|
|
// Per-wiki rules have priority, followed by language-specific and global. |
|
145
|
|
|
$globalWithLangRules = $this->mergeRules($tools['global'], $langRules); |
|
146
|
|
|
|
|
147
|
|
|
$this->tools[$projectDomain] = $this->mergeRules( |
|
148
|
|
|
$globalWithLangRules, |
|
149
|
|
|
$localRules |
|
150
|
|
|
); |
|
151
|
|
|
|
|
152
|
|
|
// Once last walk through for some tidying up and validation. |
|
153
|
|
|
$invalid = []; |
|
154
|
|
|
array_walk($this->tools[$projectDomain], function (&$data, $tool) use (&$invalid): void { |
|
155
|
|
|
// Populate the 'label' with the tool name, if a label doesn't already exist. |
|
156
|
|
|
$data['label'] = $data['label'] ?? $tool; |
|
157
|
|
|
|
|
158
|
|
|
// 'namespaces' should be an array of ints. |
|
159
|
|
|
$data['namespaces'] = $data['namespaces'] ?? []; |
|
160
|
|
|
if (isset($data['namespace'])) { |
|
161
|
|
|
$data['namespaces'][] = $data['namespace']; |
|
162
|
|
|
unset($data['namespace']); |
|
163
|
|
|
} |
|
164
|
|
|
|
|
165
|
|
|
// 'tags' should be an array of strings. |
|
166
|
|
|
$data['tags'] = $data['tags'] ?? []; |
|
167
|
|
|
if (isset($data['tag'])) { |
|
168
|
|
|
$data['tags'][] = $data['tag']; |
|
169
|
|
|
unset($data['tag']); |
|
170
|
|
|
} |
|
171
|
|
|
|
|
172
|
|
|
// If neither a tag or regex is given, it's invalid. |
|
173
|
|
|
if (empty($data['tags']) && empty($data['regex'])) { |
|
174
|
|
|
$invalid[] = $tool; |
|
175
|
|
|
} |
|
176
|
|
|
}); |
|
177
|
|
|
|
|
178
|
|
|
uksort($this->tools[$projectDomain], 'strcasecmp'); |
|
179
|
|
|
|
|
180
|
|
|
if ($invalid) { |
|
181
|
|
|
$this->tools[$projectDomain]['invalid'] = $invalid; |
|
182
|
|
|
} |
|
183
|
|
|
|
|
184
|
|
|
return $this->tools[$projectDomain]; |
|
185
|
|
|
} |
|
186
|
|
|
|
|
187
|
|
|
/** |
|
188
|
|
|
* Merges the given rule sets, giving priority to the local set. Regex is concatenated, not overridden. |
|
189
|
|
|
* @param string[] $globalRules The global rule set. |
|
190
|
|
|
* @param string[] $localRules The rule set for the local wiki. |
|
191
|
|
|
* @return string[] Merged rules. |
|
192
|
|
|
*/ |
|
193
|
|
|
private function mergeRules(array $globalRules, array $localRules): array |
|
194
|
|
|
{ |
|
195
|
|
|
// Initial set, including just the global rules. |
|
196
|
|
|
$tools = $globalRules; |
|
197
|
|
|
|
|
198
|
|
|
// Loop through local rules and override/merge as necessary. |
|
199
|
|
|
foreach ($localRules as $tool => $rules) { |
|
200
|
|
|
$newRules = $rules; |
|
201
|
|
|
|
|
202
|
|
|
if (isset($globalRules[$tool])) { |
|
203
|
|
|
// Order within array_merge is important, so that local rules get priority. |
|
204
|
|
|
$newRules = array_merge($globalRules[$tool], $rules); |
|
|
|
|
|
|
205
|
|
|
} |
|
206
|
|
|
|
|
207
|
|
|
// Regex should be merged, not overridden. |
|
208
|
|
|
if (isset($rules['regex']) && isset($globalRules[$tool]['regex'])) { |
|
209
|
|
|
$newRules['regex'] = implode('|', [ |
|
210
|
|
|
$rules['regex'], |
|
211
|
|
|
$globalRules[$tool]['regex'], |
|
212
|
|
|
]); |
|
213
|
|
|
} |
|
214
|
|
|
|
|
215
|
|
|
$tools[$tool] = $newRules; |
|
216
|
|
|
} |
|
217
|
|
|
|
|
218
|
|
|
return $tools; |
|
219
|
|
|
} |
|
220
|
|
|
|
|
221
|
|
|
/** |
|
222
|
|
|
* Get only tools that are used to revert edits. |
|
223
|
|
|
* Revert detection happens only by testing against a regular expression, and not by checking tags. |
|
224
|
|
|
* @param Project $project |
|
225
|
|
|
* @return string[][] Each tool with the tool name as the key, |
|
226
|
|
|
* and 'link' and 'regex' as the subarray keys. |
|
227
|
|
|
*/ |
|
228
|
|
|
public function getRevertTools(Project $project): array |
|
229
|
|
|
{ |
|
230
|
|
|
$projectDomain = $project->getDomain(); |
|
231
|
|
|
|
|
232
|
|
|
if (isset($this->revertTools[$projectDomain])) { |
|
233
|
|
|
return $this->revertTools[$projectDomain]; |
|
234
|
|
|
} |
|
235
|
|
|
|
|
236
|
|
|
$revertEntries = array_filter( |
|
237
|
|
|
$this->getTools($project), |
|
238
|
|
|
function ($tool) { |
|
239
|
|
|
return isset($tool['revert']) && isset($tool['regex']); |
|
240
|
|
|
} |
|
241
|
|
|
); |
|
242
|
|
|
|
|
243
|
|
|
// If 'revert' is set to `true`, then use 'regex' as the regular expression, |
|
244
|
|
|
// otherwise 'revert' is assumed to be the regex string. |
|
245
|
|
|
$this->revertTools[$projectDomain] = array_map(function ($revertTool) { |
|
246
|
|
|
return [ |
|
247
|
|
|
'link' => $revertTool['link'], |
|
248
|
|
|
'regex' => true === $revertTool['revert'] ? $revertTool['regex'] : $revertTool['revert'], |
|
249
|
|
|
]; |
|
250
|
|
|
}, $revertEntries); |
|
251
|
|
|
|
|
252
|
|
|
return $this->revertTools[$projectDomain]; |
|
253
|
|
|
} |
|
254
|
|
|
|
|
255
|
|
|
/** |
|
256
|
|
|
* Was the edit a revert, based on the edit summary? |
|
257
|
|
|
* This only works for tools defined with regular expressions, not tags. |
|
258
|
|
|
* @param string|null $summary Edit summary. Can be null for instance for suppressed edits. |
|
259
|
|
|
* @param Project $project |
|
260
|
|
|
* @return bool |
|
261
|
|
|
*/ |
|
262
|
|
|
public function isRevert(?string $summary, Project $project): bool |
|
263
|
|
|
{ |
|
264
|
|
|
foreach (array_values($this->getRevertTools($project)) as $values) { |
|
265
|
|
|
if (preg_match('/'.$values['regex'].'/', (string)$summary)) { |
|
266
|
|
|
return true; |
|
267
|
|
|
} |
|
268
|
|
|
} |
|
269
|
|
|
|
|
270
|
|
|
return false; |
|
271
|
|
|
} |
|
272
|
|
|
} |
|
273
|
|
|
|