Completed
Push — master ( 246f0f...0e35e2 )
by adam
04:02
created

MediawikiApi::newFromPage()   B

Complexity

Conditions 3
Paths 3

Size

Total Lines 32
Code Lines 19

Duplication

Lines 0
Ratio 0 %

Code Coverage

Tests 17
CRAP Score 3.0015

Importance

Changes 2
Bugs 0 Features 0
Metric Value
c 2
b 0
f 0
dl 0
loc 32
ccs 17
cts 18
cp 0.9444
rs 8.8571
cc 3
eloc 19
nc 3
nop 1
crap 3.0015
1
<?php
2
3
namespace Mediawiki\Api;
4
5
use DOMDocument;
6
use DOMXPath;
7
use GuzzleHttp\Client;
8
use GuzzleHttp\ClientInterface;
9
use GuzzleHttp\Exception\RequestException;
10
use GuzzleHttp\Promise\PromiseInterface;
11
use InvalidArgumentException;
12
use Mediawiki\Api\Guzzle\ClientFactory;
13
use Psr\Http\Message\ResponseInterface;
14
use Psr\Log\LoggerAwareInterface;
15
use Psr\Log\LoggerInterface;
16
use Psr\Log\LogLevel;
17
use Psr\Log\NullLogger;
18
use SimpleXMLElement;
19
20
/**
21
 * Main class for this library
22
 *
23
 * @since 0.1
24
 *
25
 * @author Addshore
26
 */
27
class MediawikiApi implements MediawikiApiInterface, LoggerAwareInterface {
28
29
	/**
30
	 * @var ClientInterface|null Should be accessed through getClient
31
	 */
32
	private $client = null;
33
34
	/**
35
	 * @var bool|string
36
	 */
37
	private $isLoggedIn;
38
39
	/**
40
	 * @var MediawikiSession
41
	 */
42
	private $session;
43
44
	/**
45
	 * @var string
46
	 */
47
	private $version;
48
49
	/**
50
	 * @var LoggerInterface
51
	 */
52
	private $logger;
53
54
	/**
55
	 * @var string
56
	 */
57
	private $apiUrl;
58
59
	/**
60
	 * @since 2.0
61
	 *
62
	 * @param string $apiEndpoint e.g. https://en.wikipedia.org/w/api.php
63
	 *
64
	 * @return self returns a MediawikiApi instance using $apiEndpoint
65
	 */
66
	public static function newFromApiEndpoint( $apiEndpoint ) {
67
		return new self( $apiEndpoint );
68
	}
69
70
	/**
71
	 * Create a new MediawikiApi object from a URL to any page in a MediaWiki website.
72
	 *
73
	 * @since 2.0
74
	 * @see https://en.wikipedia.org/wiki/Really_Simple_Discovery
75
	 *
76
	 * @param string $url e.g. https://en.wikipedia.org OR https://de.wikipedia.org/wiki/Berlin
77
	 * @return self returns a MediawikiApi instance using the apiEndpoint provided by the RSD
78
	 *              file accessible on all Mediawiki pages
79
	 * @throws RsdException If the RSD URL could not be found in the page's HTML.
80
	 */
81 3
	public static function newFromPage( $url ) {
82
		// Set up HTTP client and HTML document.
83 3
		$tempClient = new Client( [ 'headers' => [ 'User-Agent' => 'addwiki-mediawiki-client' ] ] );
84 3
		$pageHtml = $tempClient->get( $url )->getBody();
85 3
		$pageDoc = new DOMDocument();
86
87
		// Try to load the HTML (turn off errors temporarily; most don't matter, and if they do get
88
		// in the way of finding the API URL, will be reported in the RsdException below).
89 3
		$internalErrors = libxml_use_internal_errors( true );
90 3
		$pageDoc->loadHTML( $pageHtml );
91 3
		$libXmlErrors = libxml_get_errors();
92 3
		libxml_use_internal_errors( $internalErrors );
93
94
		// Extract the RSD link.
95 3
		$xpath = 'head/link[@type="application/rsd+xml"][@href]';
96 3
		$link = ( new DOMXpath( $pageDoc ) )->query( $xpath );
97 3
		if ( $link->length === 0 ) {
98
			// Format libxml errors for display.
99
			$libXmlErrorStr = array_reduce( $libXmlErrors, function( $prevErr, $err ) {
100
				return $prevErr . ', ' . $err->message . ' (line '.$err->line . ')';
101 1
			} );
102 1
			if ( $libXmlErrorStr ) {
103
				$libXmlErrorStr = "In addition, libxml had the following errors: $libXmlErrorStr";
104
			}
105 1
			throw new RsdException( "Unable to find RSD URL in page: $url $libXmlErrorStr" );
106
		}
107 2
		$rsdUrl = $link->item( 0 )->attributes->getnamedItem( 'href' )->nodeValue;
108
109
		// Then get the RSD XML, and return the API link.
110 2
		$rsdXml = new SimpleXMLElement( $tempClient->get( $rsdUrl )->getBody() );
111 2
		return self::newFromApiEndpoint( (string)$rsdXml->service->apis->api->attributes()->apiLink );
112
	}
113
114
	/**
115
	 * @param string $apiUrl The API Url
116
	 * @param ClientInterface|null $client Guzzle Client
117
	 * @param MediawikiSession|null $session Inject a custom session here
118
	 */
119 24
	public function __construct( $apiUrl, ClientInterface $client = null,
120
								 MediawikiSession $session = null ) {
121 24
		if ( !is_string( $apiUrl ) ) {
122 4
			throw new InvalidArgumentException( '$apiUrl must be a string' );
123
		}
124 20
		if ( $session === null ) {
125 20
			$session = new MediawikiSession( $this );
126
		}
127
128 20
		$this->apiUrl = $apiUrl;
129 20
		$this->client = $client;
130 20
		$this->session = $session;
131
132 20
		$this->logger = new NullLogger();
133 20
	}
134
135
	/**
136
	 * Get the API URL (the URL to which API requests are sent, usually ending in api.php).
137
	 * This is useful if you've created this object via MediawikiApi::newFromPage().
138
	 *
139
	 * @since 2.3
140
	 *
141
	 * @return string The API URL.
142
	 */
143
	public function getApiUrl() {
144
		return $this->apiUrl;
145
	}
146
147
	/**
148
	 * @return ClientInterface
149
	 */
150 21
	private function getClient() {
151 21
		if ( $this->client === null ) {
152 4
			$clientFactory = new ClientFactory();
153 4
			$clientFactory->setLogger( $this->logger );
154 4
			$this->client = $clientFactory->getClient();
155
		}
156 21
		return $this->client;
157
	}
158
159
	/**
160
	 * Sets a logger instance on the object
161
	 *
162
	 * @since 1.1
163
	 *
164
	 * @param LoggerInterface $logger
165
	 *
166
	 * @return null
167
	 */
168
	public function setLogger( LoggerInterface $logger ) {
169
		$this->logger = $logger;
170
		$this->session->setLogger( $logger );
171
	}
172
173
	/**
174
	 * @since 2.0
175
	 *
176
	 * @param Request $request
177
	 *
178
	 * @return PromiseInterface
179
	 *         Normally promising an array, though can be mixed (json_decode result)
180
	 *         Can throw UsageExceptions or RejectionExceptions
181
	 */
182 1
	public function getRequestAsync( Request $request ) {
183 1
		$promise = $this->getClient()->requestAsync(
184 1
			'GET',
185 1
			$this->apiUrl,
186 1
			$this->getClientRequestOptions( $request, 'query' )
187
		);
188
189
		return $promise->then( function( ResponseInterface $response ) {
190 1
			return call_user_func( [ $this, 'decodeResponse' ], $response );
191 1
		} );
192
	}
193
194
	/**
195
	 * @since 2.0
196
	 *
197
	 * @param Request $request
198
	 *
199
	 * @return PromiseInterface
200
	 *         Normally promising an array, though can be mixed (json_decode result)
201
	 *         Can throw UsageExceptions or RejectionExceptions
202
	 */
203 1
	public function postRequestAsync( Request $request ) {
204 1
		$promise = $this->getClient()->requestAsync(
205 1
			'POST',
206 1
			$this->apiUrl,
207 1
			$this->getClientRequestOptions( $request, $this->getPostRequestEncoding( $request ) )
208
		);
209
210
		return $promise->then( function( ResponseInterface $response ) {
211 1
			return call_user_func( [ $this, 'decodeResponse' ], $response );
212 1
		} );
213
	}
214
215
	/**
216
	 * @since 0.2
217
	 *
218
	 * @param Request $request
219
	 *
220
	 * @return mixed Normally an array
221
	 */
222 9
	public function getRequest( Request $request ) {
223 9
		$response = $this->getClient()->request(
224 9
			'GET',
225 9
			$this->apiUrl,
226 9
			$this->getClientRequestOptions( $request, 'query' )
227
		);
228
229 9
		return $this->decodeResponse( $response );
230
	}
231
232
	/**
233
	 * @since 0.2
234
	 *
235
	 * @param Request $request
236
	 *
237
	 * @return mixed Normally an array
238
	 */
239 10
	public function postRequest( Request $request ) {
240 10
		$response = $this->getClient()->request(
241 10
			'POST',
242 10
			$this->apiUrl,
243 10
			$this->getClientRequestOptions( $request, $this->getPostRequestEncoding( $request ) )
244
		);
245
246 10
		return $this->decodeResponse( $response );
247
	}
248
249
	/**
250
	 * @param ResponseInterface $response
251
	 *
252
	 * @return mixed
253
	 * @throws UsageException
254
	 */
255 21
	private function decodeResponse( ResponseInterface $response ) {
256 21
		$resultArray = json_decode( $response->getBody(), true );
257
258 21
		$this->logWarnings( $resultArray );
259 21
		$this->throwUsageExceptions( $resultArray );
260
261 19
		return $resultArray;
262
	}
263
264
	/**
265
	 * @param Request $request
266
	 *
267
	 * @return string
268
	 */
269 9
	private function getPostRequestEncoding( Request $request ) {
270 9
	    foreach ( $request->getParams() as $value ) {
271 9
			if ( is_resource( $value ) ) {
272 9
				return 'multipart';
273
			}
274
		   }
275 8
		return 'form_params';
276
	}
277
278
	/**
279
	 * @param Request $request
280
	 * @param string $paramsKey either 'query' or 'multipart'
281
	 *
282
	 * @throws RequestException
283
	 *
284
	 * @return array as needed by ClientInterface::get and ClientInterface::post
285
	 */
286 21
	private function getClientRequestOptions( Request $request, $paramsKey ) {
287
288 21
		$params = array_merge( $request->getParams(), [ 'format' => 'json' ] );
289 21
		if ( $paramsKey === 'multipart' ) {
290 1
			$params = $this->encodeMultipartParams( $params );
291
		}
292
293
		return [
294 21
			$paramsKey => $params,
295 21
			'headers' => array_merge( $this->getDefaultHeaders(), $request->getHeaders() ),
296
		];
297
	}
298
299
	/**
300
	 * @param array $params
301
	 *
302
	 * @return array
303
	 */
304 1
	private function encodeMultipartParams( $params ) {
305
306 1
		return array_map(
307 1
			function ( $name, $value ) {
308
309
				return [
310 1
					'name' => $name,
311 1
					'contents' => $value,
312
				];
313 1
			},
314
			array_keys( $params ),
315
			$params
316
		);
317
	}
318
319
	/**
320
	 * @return array
321
	 */
322 17
	private function getDefaultHeaders() {
323
		return [
324 17
			'User-Agent' => $this->getUserAgent(),
325
		];
326
	}
327
328 17
	private function getUserAgent() {
329 17
		$loggedIn = $this->isLoggedin();
330 17
		if ( $loggedIn ) {
331
			return 'addwiki-mediawiki-client/' . $loggedIn;
332
		}
333 17
		return 'addwiki-mediawiki-client';
334
	}
335
336
	/**
337
	 * @param $result
338
	 */
339 17
	private function logWarnings( $result ) {
340 17
		if ( is_array( $result ) && array_key_exists( 'warnings', $result ) ) {
341
			foreach ( $result['warnings'] as $module => $warningData ) {
342
				// Accomodate both formatversion=2 and old-style API results
343
				$logPrefix = $module . ': ';
344
				if ( isset( $warningData['*'] ) ) {
345
					$this->logger->warning( $logPrefix . $warningData['*'], [ 'data' => $warningData ] );
346
				} else {
347
					$this->logger->warning( $logPrefix . $warningData['warnings'], [ 'data' => $warningData ] );
348
				}
349
			}
350
		}
351 17
	}
352
353
	/**
354
	 * @param array $result
355
	 *
356
	 * @throws UsageException
357
	 */
358 17
	private function throwUsageExceptions( $result ) {
359 17
		if ( is_array( $result ) && array_key_exists( 'error', $result ) ) {
360 2
			throw new UsageException(
361 2
				$result['error']['code'],
362 2
				$result['error']['info'],
363
				$result
364
			);
365
		}
366 15
	}
367
368
	/**
369
	 * @since 0.1
370
	 *
371
	 * @return bool|string false or the name of the current user
372
	 */
373 17
	public function isLoggedin() {
374 17
		return $this->isLoggedIn;
375
	}
376
377
	/**
378
	 * @since 0.1
379
	 *
380
	 * @param ApiUser $apiUser
381
	 *
382
	 * @throws UsageException
383
	 * @return bool success
384
	 */
385 2
	public function login( ApiUser $apiUser ) {
386 2
		$this->logger->log( LogLevel::DEBUG, 'Logging in' );
387 2
		$credentials = $this->getLoginParams( $apiUser );
388 2
		$result = $this->postRequest( new SimpleRequest( 'login', $credentials ) );
389 2
		if ( $result['login']['result'] == "NeedToken" ) {
390 2
			$params = array_merge( [ 'lgtoken' => $result['login']['token'] ], $credentials );
391 2
			$result = $this->postRequest( new SimpleRequest( 'login', $params ) );
392
		}
393 2
		if ( $result['login']['result'] == "Success" ) {
394 1
			$this->isLoggedIn = $apiUser->getUsername();
395 1
			return true;
396
		}
397
398 1
		$this->isLoggedIn = false;
399 1
		$this->logger->log( LogLevel::DEBUG, 'Login failed.', $result );
400 1
		$this->throwLoginUsageException( $result );
401
		return false;
402
	}
403
404
	/**
405
	 * @param ApiUser $apiUser
406
	 *
407
	 * @return string[]
408
	 */
409 2
	private function getLoginParams( ApiUser $apiUser ) {
410
		$params = [
411 2
			'lgname' => $apiUser->getUsername(),
412 2
			'lgpassword' => $apiUser->getPassword(),
413
		];
414
415 2
		if ( !is_null( $apiUser->getDomain() ) ) {
416
			$params['lgdomain'] = $apiUser->getDomain();
417
		}
418 2
		return $params;
419
	}
420
421
	/**
422
	 * @param array $result
423
	 *
424
	 * @throws UsageException
425
	 */
426 1
	private function throwLoginUsageException( $result ) {
427 1
		$loginResult = $result['login']['result'];
428
429 1
		throw new UsageException(
430 1
			'login-' . $loginResult,
431 1
			array_key_exists( 'reason', $result['login'] )
432
				? $result['login']['reason']
433 1
				: 'No Reason given',
434
			$result
435
		);
436
	}
437
438
	/**
439
	 * @since 0.1
440
	 *
441
	 * @return bool success
442
	 */
443 2
	public function logout() {
444 2
		$this->logger->log( LogLevel::DEBUG, 'Logging out' );
445 2
		$result = $this->postRequest( new SimpleRequest( 'logout' ) );
446 2
		if ( $result === [] ) {
447 1
			$this->isLoggedIn = false;
448 1
			$this->clearTokens();
449 1
			return true;
450
		}
451 1
		return false;
452
	}
453
454
	/**
455
	 * @since 0.1
456
	 *
457
	 * @param string $type
458
	 *
459
	 * @return string
460
	 */
461 2
	public function getToken( $type = 'csrf' ) {
462 2
		return $this->session->getToken( $type );
463
	}
464
465
	/**
466
	 * @since 0.1
467
	 *
468
	 * Clears all tokens stored by the api
469
	 */
470 1
	public function clearTokens() {
471 1
		$this->session->clearTokens();
472 1
	}
473
474
	/**
475
	 * @return string
476
	 */
477 4
	public function getVersion() {
478 4
		if ( !isset( $this->version ) ) {
479 4
			$result = $this->getRequest( new SimpleRequest( 'query', [
480 4
				'meta' => 'siteinfo',
481
				'continue' => '',
482
			] ) );
483 4
			preg_match(
484 4
				'/\d+(?:\.\d+)+/',
485 4
				$result['query']['general']['generator'],
486
				$versionParts
487
			);
488 4
			$this->version = $versionParts[0];
489
		}
490 4
		return $this->version;
491
	}
492
493
}
494