Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
85.95% |
104 / 121 |
|
66.67% |
12 / 18 |
CRAP | |
0.00% |
0 / 1 |
| Connector | |
85.95% |
104 / 121 |
|
66.67% |
12 / 18 |
49.37 | |
0.00% |
0 / 1 |
| __construct | |
71.43% |
5 / 7 |
|
0.00% |
0 / 1 |
2.09 | |||
| getUrl | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getMap | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getUniqueKey | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getLastUrl | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| resetLastUrl | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| retrieve | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
2 | |||
| similar | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
12 | |||
| search | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| terms | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| write | |
85.71% |
6 / 7 |
|
0.00% |
0 / 1 |
3.03 | |||
| query | |
72.73% |
8 / 11 |
|
0.00% |
0 / 1 |
2.08 | |||
| callWithHttpOptions | |
100.00% |
13 / 13 |
|
100.00% |
1 / 1 |
3 | |||
| isRethrowableSolrException | |
100.00% |
2 / 2 |
|
100.00% |
1 / 1 |
3 | |||
| forceToBackendException | |
50.00% |
3 / 6 |
|
0.00% |
0 / 1 |
6.00 | |||
| trySolrUrls | |
100.00% |
20 / 20 |
|
100.00% |
1 / 1 |
10 | |||
| getCore | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| send | |
96.43% |
27 / 28 |
|
0.00% |
0 / 1 |
4 | |||
| 1 | <?php |
| 2 | |
| 3 | /** |
| 4 | * SOLR connector. |
| 5 | * |
| 6 | * PHP version 8 |
| 7 | * |
| 8 | * Copyright (C) Villanova University 2010. |
| 9 | * |
| 10 | * This program is free software; you can redistribute it and/or modify |
| 11 | * it under the terms of the GNU General Public License version 2, |
| 12 | * as published by the Free Software Foundation. |
| 13 | * |
| 14 | * This program is distributed in the hope that it will be useful, |
| 15 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 16 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 17 | * GNU General Public License for more details. |
| 18 | * |
| 19 | * You should have received a copy of the GNU General Public License |
| 20 | * along with this program; if not, write to the Free Software |
| 21 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
| 22 | * |
| 23 | * @category VuFind |
| 24 | * @package Search |
| 25 | * @author Andrew S. Nagy <vufind-tech@lists.sourceforge.net> |
| 26 | * @author David Maus <maus@hab.de> |
| 27 | * @author Demian Katz <demian.katz@villanova.edu> |
| 28 | * @license http://opensource.org/licenses/gpl-2.0.php GNU General Public License |
| 29 | * @link https://vufind.org |
| 30 | */ |
| 31 | |
| 32 | namespace VuFindSearch\Backend\Solr; |
| 33 | |
| 34 | use Laminas\Http\Client\Adapter\Exception\TimeoutException; |
| 35 | use Laminas\Http\Client as HttpClient; |
| 36 | use Laminas\Http\Request; |
| 37 | use Laminas\Uri\Http; |
| 38 | use VuFindSearch\Backend\Exception\BackendException; |
| 39 | use VuFindSearch\Backend\Exception\HttpErrorException; |
| 40 | use VuFindSearch\Backend\Exception\RemoteErrorException; |
| 41 | use VuFindSearch\Backend\Exception\RequestErrorException; |
| 42 | use VuFindSearch\Backend\Solr\Document\DocumentInterface; |
| 43 | use VuFindSearch\Exception\InvalidArgumentException; |
| 44 | use VuFindSearch\ParamBag; |
| 45 | |
| 46 | use function call_user_func_array; |
| 47 | use function count; |
| 48 | use function is_callable; |
| 49 | use function sprintf; |
| 50 | use function strlen; |
| 51 | |
| 52 | /** |
| 53 | * SOLR connector. |
| 54 | * |
| 55 | * @category VuFind |
| 56 | * @package Search |
| 57 | * @author Andrew S. Nagy <vufind-tech@lists.sourceforge.net> |
| 58 | * @author David Maus <maus@hab.de> |
| 59 | * @author Demian Katz <demian.katz@villanova.edu> |
| 60 | * @license http://opensource.org/licenses/gpl-2.0.php GNU General Public License |
| 61 | * @link https://vufind.org |
| 62 | */ |
| 63 | class Connector implements \Laminas\Log\LoggerAwareInterface |
| 64 | { |
| 65 | use \VuFind\Log\LoggerAwareTrait; |
| 66 | use \VuFindSearch\Backend\Feature\ConnectorCacheTrait; |
| 67 | |
| 68 | /** |
| 69 | * Maximum length of a GET url. |
| 70 | * |
| 71 | * Switches to POST if the SOLR target URL exceeds this length. |
| 72 | * |
| 73 | * @see \VuFindSearch\Backend\Solr\Connector::query() |
| 74 | * |
| 75 | * @var int |
| 76 | */ |
| 77 | public const MAX_GET_URL_LENGTH = 2048; |
| 78 | |
| 79 | /** |
| 80 | * HTTP client factory |
| 81 | * |
| 82 | * @var callable |
| 83 | */ |
| 84 | protected $clientFactory; |
| 85 | |
| 86 | /** |
| 87 | * URL or an array of alternative URLs of the SOLR core. |
| 88 | * |
| 89 | * @var string|array |
| 90 | */ |
| 91 | protected $url; |
| 92 | |
| 93 | /** |
| 94 | * Handler map. |
| 95 | * |
| 96 | * @var HandlerMap |
| 97 | */ |
| 98 | protected $map; |
| 99 | |
| 100 | /** |
| 101 | * Solr field used to store unique identifier |
| 102 | * |
| 103 | * @var string |
| 104 | */ |
| 105 | protected $uniqueKey; |
| 106 | |
| 107 | /** |
| 108 | * Url of the last request |
| 109 | * |
| 110 | * @var ?Http |
| 111 | */ |
| 112 | protected $lastUrl = null; |
| 113 | |
| 114 | /** |
| 115 | * Constructor |
| 116 | * |
| 117 | * @param string|array $url SOLR core URL or an array of alternative |
| 118 | * URLs |
| 119 | * @param HandlerMap $map Handler map |
| 120 | * @param callable|HttpClient $cf HTTP client factory or a client to clone |
| 121 | * @param string $uniqueKey Solr field used to store unique |
| 122 | * identifier |
| 123 | */ |
| 124 | public function __construct( |
| 125 | $url, |
| 126 | HandlerMap $map, |
| 127 | $cf, |
| 128 | $uniqueKey = 'id' |
| 129 | ) { |
| 130 | $this->url = $url; |
| 131 | $this->map = $map; |
| 132 | $this->uniqueKey = $uniqueKey; |
| 133 | if ($cf instanceof HttpClient) { |
| 134 | $this->clientFactory = function () use ($cf) { |
| 135 | return clone $cf; |
| 136 | }; |
| 137 | } else { |
| 138 | $this->clientFactory = $cf; |
| 139 | } |
| 140 | } |
| 141 | |
| 142 | /// Public API |
| 143 | |
| 144 | /** |
| 145 | * Get the Solr URL. |
| 146 | * |
| 147 | * @return string |
| 148 | */ |
| 149 | public function getUrl() |
| 150 | { |
| 151 | return $this->url; |
| 152 | } |
| 153 | |
| 154 | /** |
| 155 | * Return handler map. |
| 156 | * |
| 157 | * @return HandlerMap |
| 158 | */ |
| 159 | public function getMap() |
| 160 | { |
| 161 | return $this->map; |
| 162 | } |
| 163 | |
| 164 | /** |
| 165 | * Get unique key. |
| 166 | * |
| 167 | * @return string |
| 168 | */ |
| 169 | public function getUniqueKey() |
| 170 | { |
| 171 | return $this->uniqueKey; |
| 172 | } |
| 173 | |
| 174 | /** |
| 175 | * Get the last request url. |
| 176 | * |
| 177 | * @return ?Http |
| 178 | */ |
| 179 | public function getLastUrl() |
| 180 | { |
| 181 | return $this->lastUrl; |
| 182 | } |
| 183 | |
| 184 | /** |
| 185 | * Clears the last url |
| 186 | * |
| 187 | * @return void |
| 188 | */ |
| 189 | public function resetLastUrl() |
| 190 | { |
| 191 | $this->lastUrl = null; |
| 192 | } |
| 193 | |
| 194 | /** |
| 195 | * Return document specified by id. |
| 196 | * |
| 197 | * @param string $id The document to retrieve from Solr |
| 198 | * @param ParamBag $params Parameters |
| 199 | * |
| 200 | * @return string |
| 201 | */ |
| 202 | public function retrieve($id, ParamBag $params = null) |
| 203 | { |
| 204 | $params = $params ?: new ParamBag(); |
| 205 | $params |
| 206 | ->set('q', sprintf('%s:"%s"', $this->uniqueKey, addcslashes($id, '"'))); |
| 207 | |
| 208 | $handler = $this->map->getHandler(__FUNCTION__); |
| 209 | $this->map->prepare(__FUNCTION__, $params); |
| 210 | |
| 211 | return $this->query($handler, $params, true); |
| 212 | } |
| 213 | |
| 214 | /** |
| 215 | * Return records similar to a given record specified by id. |
| 216 | * |
| 217 | * Uses MoreLikeThis Request Component or MoreLikeThis Handler |
| 218 | * |
| 219 | * @param string $id ID of given record (not currently used, but |
| 220 | * retained for backward compatibility / extensibility). |
| 221 | * @param ParamBag $params Parameters |
| 222 | * |
| 223 | * @return string |
| 224 | * |
| 225 | * @SuppressWarnings(PHPMD.UnusedFormalParameter) |
| 226 | */ |
| 227 | public function similar($id, ParamBag $params) |
| 228 | { |
| 229 | $handler = $this->map->getHandler(__FUNCTION__); |
| 230 | $this->map->prepare(__FUNCTION__, $params); |
| 231 | |
| 232 | try { |
| 233 | return $this->query($handler, $params, true); |
| 234 | } catch (RequestErrorException $e) { |
| 235 | // If Solr was unable to fetch the record, just act like we have no similar records: |
| 236 | if (str_contains($e->getMessage(), 'Could not fetch document with id')) { |
| 237 | return '{}'; |
| 238 | } |
| 239 | throw $e; |
| 240 | } |
| 241 | } |
| 242 | |
| 243 | /** |
| 244 | * Execute a search. |
| 245 | * |
| 246 | * @param ParamBag $params Parameters |
| 247 | * |
| 248 | * @return string |
| 249 | */ |
| 250 | public function search(ParamBag $params) |
| 251 | { |
| 252 | $handler = $this->map->getHandler(__FUNCTION__); |
| 253 | $this->map->prepare(__FUNCTION__, $params); |
| 254 | return $this->query($handler, $params, true); |
| 255 | } |
| 256 | |
| 257 | /** |
| 258 | * Extract terms from a SOLR index. |
| 259 | * |
| 260 | * @param ParamBag $params Parameters |
| 261 | * |
| 262 | * @return string |
| 263 | */ |
| 264 | public function terms(ParamBag $params) |
| 265 | { |
| 266 | $handler = $this->map->getHandler(__FUNCTION__); |
| 267 | $this->map->prepare(__FUNCTION__, $params); |
| 268 | |
| 269 | return $this->query($handler, $params, true); |
| 270 | } |
| 271 | |
| 272 | /** |
| 273 | * Write to the SOLR index. |
| 274 | * |
| 275 | * @param DocumentInterface $document Document to write |
| 276 | * @param string $handler Update handler |
| 277 | * @param ParamBag $params Update handler parameters |
| 278 | * |
| 279 | * @return string Response body |
| 280 | */ |
| 281 | public function write( |
| 282 | DocumentInterface $document, |
| 283 | $handler = 'update', |
| 284 | ParamBag $params = null |
| 285 | ) { |
| 286 | $params = $params ?: new ParamBag(); |
| 287 | $urlSuffix = "/{$handler}"; |
| 288 | if (count($params) > 0) { |
| 289 | $urlSuffix .= '?' . implode('&', $params->request()); |
| 290 | } |
| 291 | $callback = function ($client) use ($document) { |
| 292 | $client->setEncType($document->getContentType()); |
| 293 | $body = $document->getContent(); |
| 294 | $client->setRawBody($body); |
| 295 | $client->getRequest()->getHeaders() |
| 296 | ->addHeaderLine('Content-Length', strlen($body)); |
| 297 | }; |
| 298 | return $this->trySolrUrls('POST', $urlSuffix, $callback); |
| 299 | } |
| 300 | |
| 301 | /// Internal API |
| 302 | |
| 303 | /** |
| 304 | * Send query to SOLR and return response body. |
| 305 | * |
| 306 | * @param string $handler SOLR request handler to use |
| 307 | * @param ParamBag $params Request parameters |
| 308 | * @param bool $cacheable Whether the query is cacheable |
| 309 | * |
| 310 | * @return string Response body |
| 311 | */ |
| 312 | public function query($handler, ParamBag $params, bool $cacheable = false) |
| 313 | { |
| 314 | $urlSuffix = '/' . $handler; |
| 315 | $paramString = implode('&', $params->request()); |
| 316 | if (strlen($paramString) > self::MAX_GET_URL_LENGTH) { |
| 317 | $method = Request::METHOD_POST; |
| 318 | $callback = function ($client) use ($paramString) { |
| 319 | $client->setRawBody($paramString); |
| 320 | $client->setEncType(HttpClient::ENC_URLENCODED); |
| 321 | $client->setHeaders(['Content-Length' => strlen($paramString)]); |
| 322 | }; |
| 323 | } else { |
| 324 | $method = Request::METHOD_GET; |
| 325 | $urlSuffix .= '?' . $paramString; |
| 326 | $callback = null; |
| 327 | } |
| 328 | |
| 329 | $this->debug(sprintf('Query %s', $paramString)); |
| 330 | return $this->trySolrUrls($method, $urlSuffix, $callback, $cacheable); |
| 331 | } |
| 332 | |
| 333 | /** |
| 334 | * Call a method with provided options for the HTTP client |
| 335 | * |
| 336 | * @param array $options HTTP client options |
| 337 | * @param string $method Method to call |
| 338 | * @param array ...$args Method parameters |
| 339 | * |
| 340 | * @return mixed |
| 341 | */ |
| 342 | public function callWithHttpOptions( |
| 343 | array $options, |
| 344 | string $method, |
| 345 | ...$args |
| 346 | ) { |
| 347 | $reflectionMethod = new \ReflectionMethod($this, $method); |
| 348 | if (!$reflectionMethod->isPublic()) { |
| 349 | throw new InvalidArgumentException("Method '$method' is not public"); |
| 350 | } |
| 351 | if (empty($options)) { |
| 352 | return call_user_func_array([$this, $method], $args); |
| 353 | } |
| 354 | $originalFactory = $this->clientFactory; |
| 355 | try { |
| 356 | $this->clientFactory = function (string $url) use ( |
| 357 | $originalFactory, |
| 358 | $options |
| 359 | ) { |
| 360 | $client = $originalFactory($url); |
| 361 | $client->setOptions($options); |
| 362 | return $client; |
| 363 | }; |
| 364 | return call_user_func_array([$this, $method], $args); |
| 365 | } finally { |
| 366 | $this->clientFactory = $originalFactory; |
| 367 | } |
| 368 | } |
| 369 | |
| 370 | /** |
| 371 | * Check if an exception from a Solr request should be thrown rather than retried |
| 372 | * |
| 373 | * @param \Exception $ex Exception |
| 374 | * |
| 375 | * @return bool |
| 376 | */ |
| 377 | protected function isRethrowableSolrException($ex) |
| 378 | { |
| 379 | // Solr can return 404 when the instance hasn't completed startup, so allow that to be retried: |
| 380 | return $ex instanceof TimeoutException |
| 381 | || (($ex instanceof RequestErrorException) && $ex->getResponse()->getStatusCode() !== 404); |
| 382 | } |
| 383 | |
| 384 | /** |
| 385 | * If an unexpected exception type was received, wrap it in a generic |
| 386 | * BackendException to standardize upstream handling. |
| 387 | * |
| 388 | * @param \Exception $ex Exception |
| 389 | * |
| 390 | * @return \Exception |
| 391 | */ |
| 392 | protected function forceToBackendException($ex) |
| 393 | { |
| 394 | // Don't wrap specific backend exceptions.... |
| 395 | if ( |
| 396 | $ex instanceof RemoteErrorException |
| 397 | || $ex instanceof RequestErrorException |
| 398 | || $ex instanceof HttpErrorException |
| 399 | ) { |
| 400 | return $ex; |
| 401 | } |
| 402 | return |
| 403 | new BackendException('Problem connecting to Solr.', $ex->getCode(), $ex); |
| 404 | } |
| 405 | |
| 406 | /** |
| 407 | * Try all Solr URLs until we find one that works (or throw an exception). |
| 408 | * |
| 409 | * @param string $method HTTP method to use |
| 410 | * @param string $urlSuffix Suffix to append to all URLs tried |
| 411 | * @param callable $callback Callback to configure client (null for none) |
| 412 | * @param bool $cacheable Whether the request is cacheable |
| 413 | * |
| 414 | * @return string Response body |
| 415 | * |
| 416 | * @throws RemoteErrorException SOLR signaled a server error (HTTP 5xx) |
| 417 | * @throws RequestErrorException SOLR signaled a client error (HTTP 4xx) |
| 418 | */ |
| 419 | protected function trySolrUrls( |
| 420 | $method, |
| 421 | $urlSuffix, |
| 422 | $callback = null, |
| 423 | bool $cacheable = false |
| 424 | ) { |
| 425 | // This exception should never get thrown; it's just a safety in case |
| 426 | // something unanticipated occurs. |
| 427 | $exception = new \Exception('Unexpected exception.'); |
| 428 | |
| 429 | // Loop through all base URLs and try them in turn until one works. |
| 430 | $cacheKey = null; |
| 431 | foreach ((array)$this->url as $base) { |
| 432 | $client = ($this->clientFactory)($base . $urlSuffix); |
| 433 | $client->setMethod($method); |
| 434 | if (is_callable($callback)) { |
| 435 | $callback($client); |
| 436 | } |
| 437 | // Always create the cache key from the first server, and only after any |
| 438 | // callback has been called above. |
| 439 | if ($cacheable && $this->cache && null === $cacheKey) { |
| 440 | $cacheKey = $this->getCacheKey($client); |
| 441 | if ($result = $this->getCachedData($cacheKey)) { |
| 442 | return $result; |
| 443 | } |
| 444 | } |
| 445 | try { |
| 446 | $result = $this->send($client); |
| 447 | if ($cacheKey) { |
| 448 | $this->putCachedData($cacheKey, $result); |
| 449 | } |
| 450 | return $result; |
| 451 | } catch (\Exception $ex) { |
| 452 | if ($this->isRethrowableSolrException($ex)) { |
| 453 | throw $this->forceToBackendException($ex); |
| 454 | } |
| 455 | $exception = $ex; |
| 456 | } |
| 457 | } |
| 458 | |
| 459 | // If we got this far, everything failed -- throw a BackendException with |
| 460 | // the most recent exception caught above set as the previous exception. |
| 461 | throw $this->forceToBackendException($exception); |
| 462 | } |
| 463 | |
| 464 | /** |
| 465 | * Extract the Solr core from the connector's URL. |
| 466 | * |
| 467 | * @return string |
| 468 | */ |
| 469 | public function getCore(): string |
| 470 | { |
| 471 | $url = rtrim($this->getUrl(), '/'); |
| 472 | $parts = explode('/', $url); |
| 473 | return array_pop($parts); |
| 474 | } |
| 475 | |
| 476 | /** |
| 477 | * Send request the SOLR and return the response. |
| 478 | * |
| 479 | * @param HttpClient $client Prepared HTTP client |
| 480 | * |
| 481 | * @return string Response body |
| 482 | * |
| 483 | * @throws RemoteErrorException SOLR signaled a server error (HTTP 5xx) |
| 484 | * @throws RequestErrorException SOLR signaled a client error (HTTP 4xx) |
| 485 | */ |
| 486 | protected function send(HttpClient $client) |
| 487 | { |
| 488 | $this->debug( |
| 489 | sprintf('=> %s %s', $client->getMethod(), $client->getUri()) |
| 490 | ); |
| 491 | |
| 492 | $this->lastUrl = $client->getUri(); |
| 493 | |
| 494 | $time = microtime(true); |
| 495 | $response = $client->send(); |
| 496 | $time = microtime(true) - $time; |
| 497 | |
| 498 | $this->debug( |
| 499 | sprintf( |
| 500 | '<= %s %s', |
| 501 | $response->getStatusCode(), |
| 502 | $response->getReasonPhrase() |
| 503 | ), |
| 504 | ['time' => $time] |
| 505 | ); |
| 506 | |
| 507 | if (!$response->isSuccess()) { |
| 508 | // Return a more detailed error message for a 400 error: |
| 509 | if ($response->getStatusCode() === 400) { |
| 510 | $json = json_decode($response->getBody(), true); |
| 511 | $msgParts = ['400', $response->getReasonPhrase()]; |
| 512 | if ($msg = $json['error']['msg'] ?? '') { |
| 513 | $msgParts[] = $msg; |
| 514 | } |
| 515 | throw new RequestErrorException( |
| 516 | implode(' ', $msgParts), |
| 517 | 400, |
| 518 | $response |
| 519 | ); |
| 520 | } |
| 521 | throw HttpErrorException::createFromResponse($response); |
| 522 | } |
| 523 | return $response->getBody(); |
| 524 | } |
| 525 | } |