gordonbanderson /
silverstripe-elastica
Complex classes like Searchable often do a lot of different things. To break such a class down, we need to identify a cohesive component within that class. A common approach to find such a component is to look for fields/methods that share the same prefixes, or suffixes. You can also have a look at the cohesion graph to spot any un-connected, or weakly-connected components.
Once you have determined the fields that belong together, you can apply the Extract Class refactoring. If the component makes sense as a sub-class, Extract Subclass is also a candidate, and is often faster.
While breaking up the class, it is a good idea to analyze how other classes use Searchable, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 12 | class Searchable extends \DataExtension { |
||
| 13 | |||
| 14 | /** |
||
| 15 | * Counter used to display progress of indexing |
||
| 16 | * @var integer |
||
| 17 | */ |
||
| 18 | public static $index_ctr = 0; |
||
| 19 | |||
| 20 | /** |
||
| 21 | * Everytime progressInterval divides $index_ctr exactly display progress |
||
| 22 | * @var integer |
||
| 23 | */ |
||
| 24 | private static $progressInterval = 0; |
||
| 25 | |||
| 26 | public static $mappings = array( |
||
| 27 | 'Boolean' => 'boolean', |
||
| 28 | 'Decimal' => 'double', |
||
| 29 | 'Currency' => 'double', |
||
| 30 | 'Double' => 'double', |
||
| 31 | 'Enum' => 'string', |
||
| 32 | 'Float' => 'float', |
||
| 33 | 'HTMLText' => 'string', |
||
| 34 | 'HTMLVarchar' => 'string', |
||
| 35 | 'Int' => 'integer', |
||
| 36 | 'Text' => 'string', |
||
| 37 | 'VarChar' => 'string', |
||
| 38 | 'Varchar' => 'string', |
||
| 39 | 'Year' => 'integer', |
||
| 40 | 'Percentage' => 'double', |
||
| 41 | 'Time' => 'date', |
||
| 42 | |||
| 43 | // The 2 different date types will be stored with different formats |
||
| 44 | 'Date' => 'date', |
||
| 45 | 'SS_Datetime' => 'date', |
||
| 46 | 'Datetime' => 'date', |
||
| 47 | 'DBLocale' => 'string' |
||
| 48 | ); |
||
| 49 | |||
| 50 | |||
| 51 | /** |
||
| 52 | * @var ElasticaService associated elastica search service |
||
| 53 | */ |
||
| 54 | protected $service; |
||
| 55 | |||
| 56 | |||
| 57 | /** |
||
| 58 | * Array of fields that need HTML parsed |
||
| 59 | * @var array |
||
| 60 | */ |
||
| 61 | protected $html_fields = array(); |
||
| 62 | |||
| 63 | /** |
||
| 64 | * Store a mapping of relationship name to result type |
||
| 65 | */ |
||
| 66 | protected $relationship_methods = array(); |
||
| 67 | |||
| 68 | |||
| 69 | /** |
||
| 70 | * If importing a large number of items from a fixtures file, or indeed some other source, then |
||
| 71 | * it is quicker to set a flag of value IndexingOff => false. This has the effect of ensuring |
||
| 72 | * no indexing happens, a request is normally made per fixture when loading. One can then run |
||
| 73 | * the reindexing teask to bulk index in one HTTP POST request to Elasticsearch |
||
| 74 | * |
||
| 75 | * @var boolean |
||
| 76 | */ |
||
| 77 | private static $IndexingOff = false; |
||
| 78 | |||
| 79 | |||
| 80 | /** |
||
| 81 | * @see getElasticaResult |
||
| 82 | * @var \Elastica\Result |
||
| 83 | */ |
||
| 84 | protected $elastica_result; |
||
| 85 | |||
| 86 | public function __construct(ElasticaService $service) { |
||
| 90 | |||
| 91 | |||
| 92 | /** |
||
| 93 | * Get the elasticsearch type name |
||
| 94 | * |
||
| 95 | * @return string |
||
| 96 | */ |
||
| 97 | public function getElasticaType() { |
||
| 100 | |||
| 101 | |||
| 102 | /** |
||
| 103 | * If the owner is part of a search result |
||
| 104 | * the raw Elastica search result is returned |
||
| 105 | * if set via setElasticaResult |
||
| 106 | * |
||
| 107 | * @return \Elastica\Result |
||
| 108 | */ |
||
| 109 | public function getElasticaResult() { |
||
| 112 | |||
| 113 | |||
| 114 | /** |
||
| 115 | * Set the raw Elastica search result |
||
| 116 | * |
||
| 117 | * @param \Elastica\Result |
||
| 118 | */ |
||
| 119 | public function setElasticaResult(\Elastica\Result $result) { |
||
| 122 | |||
| 123 | |||
| 124 | /** |
||
| 125 | * Gets an array of elastic field definitions. |
||
| 126 | * |
||
| 127 | * @return array |
||
| 128 | */ |
||
| 129 | public function getElasticaFields($storeMethodName = false, $recurse = true) { |
||
| 130 | $db = $this->owner->db(); |
||
| 131 | $fields = $this->getAllSearchableFields(); |
||
| 132 | $result = array(); |
||
| 133 | |||
| 134 | foreach($fields as $name => $params) { |
||
| 135 | $spec = array(); |
||
| 136 | $name = str_replace('()', '', $name); |
||
| 137 | |||
| 138 | if(array_key_exists($name, $db)) { |
||
| 139 | $class = $db[$name]; |
||
| 140 | $this->assignSpecForStandardFieldType($name, $class, $spec); |
||
| 141 | } else { |
||
| 142 | // field name is not in the db, it could be a method |
||
| 143 | $has_lists = $this->getListRelationshipMethods(); |
||
| 144 | $has_ones = $this->owner->has_one(); |
||
| 145 | |||
| 146 | // check has_many and many_many relations |
||
| 147 | if(isset($has_lists[$name])) { |
||
| 148 | // the classes returned by the list method |
||
| 149 | $resultType = $has_lists[$name]; |
||
| 150 | $this->assignSpecForRelationship($name, $resultType, $spec, $storeMethodName, $recurse); |
||
| 151 | } else if(isset($has_ones[$name])) { |
||
| 152 | $resultType = $has_ones[$name]; |
||
| 153 | $this->assignSpecForRelationship($name, $resultType, $spec, $storeMethodName, $recurse); |
||
| 154 | } |
||
| 155 | // otherwise fall back to string - Enum is one such category |
||
| 156 | else { |
||
| 157 | $spec["type"] = "string"; |
||
| 158 | } |
||
| 159 | } |
||
| 160 | |||
| 161 | $this->addIndexedFields($name, $spec); |
||
| 162 | |||
| 163 | $result[$name] = $spec; |
||
| 164 | } |
||
| 165 | |||
| 166 | if($this->owner->hasMethod('updateElasticHTMLFields')) { |
||
| 167 | $this->html_fields = $this->owner->updateElasticHTMLFields($this->html_fields); |
||
| 168 | } |
||
| 169 | |||
| 170 | return $result; |
||
| 171 | } |
||
| 172 | |||
| 173 | |||
| 174 | |||
| 175 | private function addIndexedFields($name, &$spec) { |
||
| 176 | // in the case of a relationship type will not be set |
||
| 177 | if(isset($spec['type'])) { |
||
| 178 | if($spec['type'] == 'string') { |
||
| 179 | $unstemmed = array(); |
||
| 180 | $unstemmed['type'] = "string"; |
||
| 181 | $unstemmed['analyzer'] = "unstemmed"; |
||
| 182 | $unstemmed['term_vector'] = "yes"; |
||
| 183 | $extraFields = array('standard' => $unstemmed); |
||
| 184 | |||
| 185 | $shingles = array(); |
||
| 186 | $shingles['type'] = "string"; |
||
| 187 | $shingles['analyzer'] = "shingles"; |
||
| 188 | $shingles['term_vector'] = "yes"; |
||
| 189 | $extraFields['shingles'] = $shingles; |
||
| 190 | |||
| 191 | //Add autocomplete field if so required |
||
| 192 | $autocomplete = \Config::inst()->get($this->owner->ClassName, 'searchable_autocomplete'); |
||
| 193 | |||
| 194 | if(isset($autocomplete) && in_array($name, $autocomplete)) { |
||
| 195 | $autocompleteField = array(); |
||
| 196 | $autocompleteField['type'] = "string"; |
||
| 197 | $autocompleteField['index_analyzer'] = "autocomplete_index_analyzer"; |
||
| 198 | $autocompleteField['search_analyzer'] = "autocomplete_search_analyzer"; |
||
| 199 | $autocompleteField['term_vector'] = "yes"; |
||
| 200 | $extraFields['autocomplete'] = $autocompleteField; |
||
| 201 | } |
||
| 202 | |||
| 203 | $spec['fields'] = $extraFields; |
||
| 204 | // FIXME - make index/locale specific, get from settings |
||
| 205 | $spec['analyzer'] = 'stemmed'; |
||
| 206 | $spec['term_vector'] = "yes"; |
||
| 207 | } |
||
| 208 | } |
||
| 209 | } |
||
| 210 | |||
| 211 | |||
| 212 | /** |
||
| 213 | * @param string &$name |
||
| 214 | * @param boolean $storeMethodName |
||
| 215 | * @param boolean $recurse |
||
| 216 | */ |
||
| 217 | private function assignSpecForRelationship(&$name, $resultType, &$spec, $storeMethodName, $recurse) { |
||
| 218 | $resultTypeInstance = \Injector::inst()->create($resultType); |
||
| 219 | $resultTypeMapping = array(); |
||
| 220 | // get the fields for the result type, but do not recurse |
||
| 221 | if($recurse) { |
||
| 222 | $resultTypeMapping = $resultTypeInstance->getElasticaFields($storeMethodName, false); |
||
| 223 | } |
||
| 224 | $resultTypeMapping['ID'] = array('type' => 'integer'); |
||
| 225 | if($storeMethodName) { |
||
| 226 | $resultTypeMapping['__method'] = $name; |
||
| 227 | } |
||
| 228 | $spec = array('properties' => $resultTypeMapping); |
||
| 229 | // we now change the name to the result type, not the method name |
||
| 230 | $name = $resultType; |
||
| 231 | } |
||
| 232 | |||
| 233 | |||
| 234 | /** |
||
| 235 | * @param string $name |
||
| 236 | */ |
||
| 237 | private function assignSpecForStandardFieldType($name, $class, &$spec) { |
||
| 238 | if(($pos = strpos($class, '('))) { |
||
| 239 | // Valid in the case of Varchar(255) |
||
| 240 | $class = substr($class, 0, $pos); |
||
| 241 | } |
||
| 242 | |||
| 243 | if(array_key_exists($class, self::$mappings)) { |
||
| 244 | $spec['type'] = self::$mappings[$class]; |
||
| 245 | if($spec['type'] === 'date') { |
||
| 246 | $spec['format'] = $this->getFormatForDate($class); |
||
| 247 | } |
||
| 248 | |||
| 249 | if($class === 'HTMLText' || $class === 'HTMLVarchar') { |
||
| 250 | array_push($this->html_fields, $name); |
||
| 251 | } |
||
| 252 | } |
||
| 253 | } |
||
| 254 | |||
| 255 | |||
| 256 | private function getFormatForDate($class) { |
||
| 257 | $format = 'y-M-d'; // default |
||
| 258 | switch ($class) { |
||
| 259 | case 'Date': |
||
| 260 | $format = 'y-M-d'; |
||
| 261 | break; |
||
| 262 | case 'SS_Datetime': |
||
| 263 | $format = 'y-M-d H:m:s'; |
||
| 264 | break; |
||
| 265 | case 'Datetime': |
||
| 266 | $format = 'y-M-d H:m:s'; |
||
| 267 | break; |
||
| 268 | case 'Time': |
||
| 269 | $format = 'H:m:s'; |
||
| 270 | break; |
||
| 271 | } |
||
| 272 | |||
| 273 | return $format; |
||
| 274 | } |
||
| 275 | |||
| 276 | |||
| 277 | /** |
||
| 278 | * Get the elasticsearch mapping for the current document/type |
||
| 279 | * |
||
| 280 | * @return \Elastica\Type\Mapping |
||
| 281 | */ |
||
| 282 | public function getElasticaMapping() { |
||
| 313 | |||
| 314 | |||
| 315 | /** |
||
| 316 | * Get an elasticsearch document |
||
| 317 | * |
||
| 318 | * @return \Elastica\Document |
||
| 319 | */ |
||
| 320 | public function getElasticaDocument() { |
||
| 357 | |||
| 358 | |||
| 359 | public function getFieldValuesAsArray($recurse = true) { |
||
| 421 | |||
| 422 | |||
| 423 | /** |
||
| 424 | * Returns whether to include the document into the search index. |
||
| 425 | * All documents are added unless they have a field "ShowInSearch" which is set to false |
||
| 426 | * |
||
| 427 | * @return boolean |
||
| 428 | */ |
||
| 429 | public function showRecordInSearch() { |
||
| 432 | |||
| 433 | |||
| 434 | /** |
||
| 435 | * Delete the record from the search index if ShowInSearch is deactivated (non-SiteTree). |
||
| 436 | */ |
||
| 437 | public function onBeforeWrite() { |
||
| 445 | |||
| 446 | |||
| 447 | /** |
||
| 448 | * Delete the record from the search index if ShowInSearch is deactivated (SiteTree). |
||
| 449 | */ |
||
| 450 | public function onBeforePublish() { |
||
| 461 | |||
| 462 | |||
| 463 | /** |
||
| 464 | * Updates the record in the search index (non-SiteTree). |
||
| 465 | */ |
||
| 466 | public function onAfterWrite() { |
||
| 469 | |||
| 470 | |||
| 471 | /** |
||
| 472 | * Updates the record in the search index (SiteTree). |
||
| 473 | */ |
||
| 474 | public function onAfterPublish() { |
||
| 477 | |||
| 478 | |||
| 479 | /** |
||
| 480 | * Updates the record in the search index. |
||
| 481 | */ |
||
| 482 | protected function doIndexDocument() { |
||
| 489 | |||
| 490 | |||
| 491 | /** |
||
| 492 | * Removes the record from the search index (non-SiteTree). |
||
| 493 | */ |
||
| 494 | public function onAfterDelete() { |
||
| 497 | |||
| 498 | |||
| 499 | /** |
||
| 500 | * Removes the record from the search index (non-SiteTree). |
||
| 501 | */ |
||
| 502 | public function onAfterUnpublish() { |
||
| 505 | |||
| 506 | |||
| 507 | /** |
||
| 508 | * Removes the record from the search index if the "ShowInSearch" attribute is set to true. |
||
| 509 | */ |
||
| 510 | protected function doDeleteDocumentIfInSearch() { |
||
| 515 | |||
| 516 | |||
| 517 | /** |
||
| 518 | * Removes the record from the search index. |
||
| 519 | */ |
||
| 520 | protected function doDeleteDocument() { |
||
| 532 | |||
| 533 | |||
| 534 | /** |
||
| 535 | * Return all of the searchable fields defined in $this->owner::$searchable_fields and all the parent classes. |
||
| 536 | * |
||
| 537 | * @param $recuse Whether or not to traverse relationships. First time round yes, subsequently no |
||
| 538 | * @return array searchable fields |
||
| 539 | */ |
||
| 540 | public function getAllSearchableFields($recurse = true) { |
||
| 594 | |||
| 595 | |||
| 596 | /* |
||
| 597 | Evaluate each field, e.g. 'Title', 'Member.Name' |
||
| 598 | */ |
||
| 599 | private function fieldsToElasticaConfig($fields) { |
||
| 616 | |||
| 617 | |||
| 618 | public function requireDefaultRecords() { |
||
| 669 | |||
| 670 | |||
| 671 | private function getListRelationshipMethods() { |
||
| 683 | |||
| 684 | |||
| 685 | private function isInSiteTree($classname) { |
||
| 699 | |||
| 700 | |||
| 701 | /* |
||
| 702 | Allow the option of overriding the default template with one of <ClassName>ElasticSearchResult |
||
| 703 | */ |
||
| 704 | public function RenderResult($linkToContainer = '') { |
||
| 709 | |||
| 710 | |||
| 711 | |||
| 712 | public function getTermVectors() { |
||
| 715 | |||
| 716 | |||
| 717 | public function updateCMSFields(\FieldList $fields) { |
||
| 782 | |||
| 783 | |||
| 784 | } |
||
| 785 |
