Analysis_Response_Ops::add_occurrences() - Code Metrics - Inspection of "Fixed failing tests" - insideout10/wordlift-plugin - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — develop ( d9328c...0bc50d )

unknown

created 2020-12-08 10:03 UTC

Analysis_Response_Ops::add_occurrences() D

↳ Parent: Analysis_Response_Ops

Complexity

Conditions	15
Paths	26

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
cc	15
nc	26
nop	1
dl	0
loc	96
rs	4.8206
c	0
b	0
f	0

How to fix Long Method Complexity

<?php
/**
 * This file provides a class to manipulate the analysis response.
 *
 * @author David Riccitelli <[email protected]>
 * @since 3.25.0
 * @package Wordlift\Analysis\Response
 */

namespace Wordlift\Analysis\Response;

use stdClass;
use Wordlift\Entity\Entity_Helper;

class Analysis_Response_Ops {

	/**
	 * The analysis response json.
	 *
	 * @since 3.21.5
	 * @access private
	 * @var mixed $json Holds the analysis response json.
	 */
	private $json;

	/**
	 * Holds the {@link Wordlift_Entity_Uri_Service}.
	 *
	 * @since 3.21.5
	 * @access private
	 * @var \Wordlift_Entity_Uri_Service $entity_uri_service The {@link Wordlift_Entity_Uri_Service} instance.
	 */
	private $entity_uri_service;

	private $entity_service;

	/**
	 * @var \Wordlift_Entity_Type_Service
	 */
	private $entity_type_service;
	/**
	 * @var \Wordlift_Post_Image_Storage
	 */
	private $post_image_storage;

	/**
	 * @var Entity_Helper
	 */
	private $entity_helper;

	/**
	 * Analysis_Response_Ops constructor.
	 *
	 * @param \Wordlift_Entity_Uri_Service $entity_uri_service The {@link Wordlift_Entity_Uri_Service}.
	 * @param \Wordlift_Entity_Service $entity_service The {@link Wordlift_Entity_Service}.
	 * @param \Wordlift_Entity_Type_Service $entity_type_service The {@link Wordlift_Entity_Type_Service}.
	 * @param \Wordlift_Post_Image_Storage $post_image_storage A {@link Wordlift_Post_Image_Storage} instance.
	 * @param Entity_Helper $entity_helper The {@link Entity_Helper}.
	 * @param mixed $json The analysis response json.
	 *
	 * @since 3.21.5
	 */
	public function __construct( $entity_uri_service, $entity_service, $entity_type_service, $post_image_storage, $entity_helper, $json ) {


		$this->json                = $json;
		$this->entity_uri_service  = $entity_uri_service;
		$this->entity_service      = $entity_service;
		$this->entity_type_service = $entity_type_service;
		$this->post_image_storage  = $post_image_storage;
		$this->entity_helper       = $entity_helper;

	}

	/**
	 * Switches remote entities, i.e. entities with id outside the local dataset, to local entities.
	 *
	 * The function takes all the entities that have an id which is not local. For each remote entity, a list of URIs
	 * is built comprising the entity id and the sameAs. Then a query is issued in the local database to find potential
	 * matches from the local vocabulary.
	 *
	 * If found, the entity id is swapped with the local id and the remote id is added to the sameAs.
	 *
	 * @return Analysis_Response_Ops The current Analysis_Response_Ops instance.
	 */
	public function make_entities_local() {

		if ( ! isset( $this->json->entities ) ) {
			return $this;
		}

		// Get the URIs.
		$uris     = array_keys( get_object_vars( $this->json->entities ) );
		$mappings = $this->entity_helper->map_many_to_local( $uris );

		foreach ( $mappings as $external_uri => $internal_uri ) {

			// Move the data from the external URI to the internal URI.
			if ( ! isset( $this->json->entities->{$internal_uri} ) ) {
				$this->json->entities->{$internal_uri} = $this->json->entities->{$external_uri};
			}

			// Ensure sameAs is an array.
			if ( ! isset( $this->json->entities->{$internal_uri}->sameAs )
			     || ! is_array( $this->json->entities->{$internal_uri}->sameAs ) ) {
				$this->json->entities->{$internal_uri}->sameAs = array();
			}

			// Add the external URI as sameAs.
			$this->json->entities->{$internal_uri}->sameAs[] = $external_uri;

			// Finally remove the external URI.
			unset( $this->json->entities->{$external_uri} );
		}

		if ( isset( $this->json->annotations ) ) {
			foreach ( $this->json->annotations as $key => $annotation ) {
				if ( isset( $annotation->entityMatches ) ) {
					foreach ( $annotation->entityMatches as $match ) {
						if ( isset( $match->entityId ) && isset( $mappings[ $match->entityId ] ) ) {
							$match->entityId = $mappings[ $match->entityId ];
						}
					}
				}
			}
		}

		return $this;
	}

	/**
	 * Add occurrences by parsing the provided html content.
	 *
	 * @param string $content The html content with annotations.
	 *
	 * @return Analysis_Response_Ops The {@link Analysis_Response_Ops} instance.
	 *
	 * @since 3.23.7 refactor the regex pattern to take into account that there might be css classes between textannotation
	 *  and disambiguated.
	 *
	 * @link https://github.com/insideout10/wordlift-plugin/issues/1001
	 */
	public function add_occurrences( $content ) {

		// Try to get all the disambiguated annotations and bail out if an error occurs.
		if ( false === preg_match_all(
				'|<span\s+id="([^"]+)"\s+class="textannotation\s+(?:\S+\s+)?disambiguated(?=[\s"])[^"]*"\s+itemid="([^"]*)">(.*?)</span>|',
				$content,
				$matches,
				PREG_OFFSET_CAPTURE
			) ) {
			return $this;
		}

		if ( empty( $matches ) ) {
			return $this;
		}

		$parse_data = array_reduce( range( 0, count( $matches[1] ) - 1 ), function ( $carry, $i ) use ( $matches ) {
			if ( empty( $matches[0] ) ) {
				return $carry;
			}

			$start         = $matches[0][ $i ][1];
			$end           = $start + strlen( $matches[0][ $i ][0] );
			$annotation_id = $matches[1][ $i ][0];
			$item_id       = $matches[2][ $i ][0];
			$text          = $matches[3][ $i ][0];

			$annotation               = new StdClass;
			$annotation->annotationId = $annotation_id;
			$annotation->start        = $start;
			$annotation->end          = $end;
			$annotation->text         = $text;

			$entity_match                = new StdClass;
			$entity_match->confidence    = 100;
			$entity_match->entityId      = $item_id;
			$annotation->entityMatches[] = $entity_match;

			$carry['annotations'][ $annotation_id ] = $annotation;
			$carry['occurrences'][ $item_id ][]     = $annotation_id;

			return $carry;
		}, array( 'annotations' => array(), 'occurrences' => array(), ) );

		$annotations = $parse_data['annotations'];
		$occurrences = $parse_data['occurrences'];

		foreach ( array_keys( $occurrences ) as $item_id ) {

			// If the entity isn't there, add it.
			if ( ! is_bool( $this->json ) && ! isset( $this->json->entities->{$item_id} ) ) {
				$entity = $this->get_local_entity( $item_id );

				// Entity not found in the local vocabulary, continue to the next one.
				if ( false === $entity ) {
					continue;
				}

				$this->json->entities->{$item_id} = $entity;
			}
		}

		// Here we're adding back some data structures required by the client-side code.
		//
		// We're adding:
		//  1. the .entities[entity_id].occurrences array with the annotations' ids.
		//  2. the .entities[entity_id].annotations[annotation_id] = { id: annotation_id } map.
		//
		// Before 3.23.0 this was done by the client-side code located in src/coffee/editpost-widget/app.services.AnalysisService.coffee
		// function `preselect`, which was called by src/coffee/editpost-widget/app.services.EditorService.coffee in
		// `embedAnalysis`.

		if ( ! is_bool( $this->json ) ) {
			foreach ( $this->json->entities as $id => $entity ) {
				$this->json->entities->{$id}->occurrences = isset( $occurrences[ $id ] ) ? $occurrences[ $id ] : array();;

				foreach ( $this->json->entities->{$id}->occurrences as $annotation_id ) {
					$this->json->entities->{$id}->annotations[ $annotation_id ] = array(
						'id' => $annotation_id,
					);
				}
			}
		}

		// Add the missing annotations. This allows the analysis response to work also if we didn't receive results
		// from the analysis API.
		foreach ( $annotations as $annotation_id => $annotation ) {

			if ( ! is_bool( $this->json ) && ! isset( $this->json->annotations->{$annotation_id} ) ) {
				$this->json->annotations->{$annotation_id} = $annotation;
			}

		}

		return $this;
	}

	/**
	 * Add local entities
	 *
	 * @return Analysis_Response_Ops The {@link Analysis_Response_Ops} instance.
	 *
	 * @since 3.27.6
	 *
	 * @link https://github.com/insideout10/wordlift-plugin/issues/1178
	 */
	public function add_local_entities(){

		// Populating the local entities object
		$entities = array();
		foreach ( $this->json->annotations as $annotation ) {
			foreach ($annotation->entityMatches as $entity_matches){

				$entity_id = url_to_postid( htmlspecialchars_decode($entity_matches->entityId) );
				$serialized_entity = wl_serialize_entity( $entity_id );

				if($serialized_entity){
					$serialized_entity['entityId'] = $serialized_entity['id'];
					unset($serialized_entity['id']);

					$entities[$entity_matches->entityId] = $serialized_entity;
				}

			}
		}

		// Adding occurrences and annotations data structures required by the client-side code.
		foreach ($entities as $entity_id => $entity ){
			foreach ( $this->json->annotations as $annotation ) {
				if($annotation->entityMatches[0]->entityId === $entity_id){
					$entities[$entity_id]['occurrences'][] = $annotation->annotationId;
					$entities[$entity_id]['annotations'][$annotation->annotationId]['id'] = $annotation->annotationId;
				}
			}
		}

		$this->json->entities = $entities;

		return $this;

	}

	private function get_local_entity( $uri ) {

		$entity = $this->entity_uri_service->get_entity( $uri );

		if ( null === $entity ) {
			return false;
		}

		$type   = $this->entity_type_service->get( $entity->ID );
		$images = $this->post_image_storage->get( $entity->ID );

		return (object) array(
			'id'          => $uri,
			'label'       => $entity->post_title,
			/*
			 * As of 2020.06.29 we're comment out the `post_content` because Gutenberg posts will return here
			 * the whole Gutenberg source including potentially our own wordlift/classification block, which means
			 * that data may grow quickly to more than a 100 KBytes and could break web servers.
			 *
			 * We don't really need the description for local entities (because the description is indeed taken from
			 * the local WordPress database) and we're not using it anywhere in the UI.
			 *
			 * So take extra care in enabling this line: eventually consider using the post_excerpt.
			 *
			 * PS: We didn't test using the WordLift Post Excerpt Helper.
			 */
			// 'description' => $entity->post_content,
			'description' => '',
			'sameAs'      => wl_schema_get_value( $entity->ID, 'sameAs' ),
			'mainType'    => str_replace( 'wl-', '', $type['css_class'] ),
			'types'       => wl_get_entity_rdf_types( $entity->ID ),
			'images'      => $images,
		);
	}

	/**
	 * Return the JSON response.
	 *
	 * @return mixed The JSON response.
	 * @since 3.24.2
	 */
	public function get_json() {

		return $this->json;
	}

	/**
	 * Get the string representation of the JSON.
	 *
	 * @return false|string The string representation or false in case of error.
	 */
	public function to_string() {

		// Add the `JSON_UNESCAPED_UNICODE` only for PHP 5.4+.
		$options = ( version_compare( PHP_VERSION, '5.4', '>=' )
			? 256 : 0 );

		return wp_json_encode( $this->json, $options );
	}

}


1		<?php
2		/**
3		* This file provides a class to manipulate the analysis response.
4		*
5		* @author David Riccitelli <[email protected]>
6		* @since 3.25.0
7		* @package Wordlift\Analysis\Response
8		*/
9
10		namespace Wordlift\Analysis\Response;
11
12		use stdClass;
13		use Wordlift\Entity\Entity_Helper;
14
15		class Analysis_Response_Ops {
16
17		/**
18		* The analysis response json.
19		*
20		* @since 3.21.5
21		* @access private
22		* @var mixed $json Holds the analysis response json.
23		*/
24		private $json;
25
26		/**
27		* Holds the {@link Wordlift_Entity_Uri_Service}.
28		*
29		* @since 3.21.5
30		* @access private
31		* @var \Wordlift_Entity_Uri_Service $entity_uri_service The {@link Wordlift_Entity_Uri_Service} instance.
32		*/
33		private $entity_uri_service;
34
35		private $entity_service;
36
37		/**
38		* @var \Wordlift_Entity_Type_Service
39		*/
40		private $entity_type_service;
41		/**
42		* @var \Wordlift_Post_Image_Storage
43		*/
44		private $post_image_storage;
45
46		/**
47		* @var Entity_Helper
48		*/
49		private $entity_helper;
50
51		/**
52		* Analysis_Response_Ops constructor.
53		*
54		* @param \Wordlift_Entity_Uri_Service $entity_uri_service The {@link Wordlift_Entity_Uri_Service}.
55		* @param \Wordlift_Entity_Service $entity_service The {@link Wordlift_Entity_Service}.
56		* @param \Wordlift_Entity_Type_Service $entity_type_service The {@link Wordlift_Entity_Type_Service}.
57		* @param \Wordlift_Post_Image_Storage $post_image_storage A {@link Wordlift_Post_Image_Storage} instance.
58		* @param Entity_Helper $entity_helper The {@link Entity_Helper}.
59		* @param mixed $json The analysis response json.
60		*
61		* @since 3.21.5
62		*/
63	View Code Duplication	public function __construct( $entity_uri_service, $entity_service, $entity_type_service, $post_image_storage, $entity_helper, $json ) {
		0 ignored issues – show Duplication introduced 2020-02-15 12:51 UTC by Report Bug Copy Issue Report This method seems to be duplicated in your project. Duplicated code is one of the most pungent code smells. If you need to duplicate the same code in three or more different places, we strongly encourage you to look into extracting the code into a single class or operation. You can also find more detailed suggestions in the “Code” section of your repository. Loading history...
64
65		$this->json = $json;
66		$this->entity_uri_service = $entity_uri_service;
67		$this->entity_service = $entity_service;
68		$this->entity_type_service = $entity_type_service;
69		$this->post_image_storage = $post_image_storage;
70		$this->entity_helper = $entity_helper;
71
72		}
73
74		/**
75		* Switches remote entities, i.e. entities with id outside the local dataset, to local entities.
76		*
77		* The function takes all the entities that have an id which is not local. For each remote entity, a list of URIs
78		* is built comprising the entity id and the sameAs. Then a query is issued in the local database to find potential
79		* matches from the local vocabulary.
80		*
81		* If found, the entity id is swapped with the local id and the remote id is added to the sameAs.
82		*
83		* @return Analysis_Response_Ops The current Analysis_Response_Ops instance.
84		*/
85		public function make_entities_local() {
86
87		if ( ! isset( $this->json->entities ) ) {
88		return $this;
89		}
90
91		// Get the URIs.
92		$uris = array_keys( get_object_vars( $this->json->entities ) );
93		$mappings = $this->entity_helper->map_many_to_local( $uris );
94
95		foreach ( $mappings as $external_uri => $internal_uri ) {
96
97		// Move the data from the external URI to the internal URI.
98		if ( ! isset( $this->json->entities->{$internal_uri} ) ) {
99		$this->json->entities->{$internal_uri} = $this->json->entities->{$external_uri};
100		}
101
102		// Ensure sameAs is an array.
103		if ( ! isset( $this->json->entities->{$internal_uri}->sameAs )
104		\|\| ! is_array( $this->json->entities->{$internal_uri}->sameAs ) ) {
105		$this->json->entities->{$internal_uri}->sameAs = array();
106		}
107
108		// Add the external URI as sameAs.
109		$this->json->entities->{$internal_uri}->sameAs[] = $external_uri;
110
111		// Finally remove the external URI.
112		unset( $this->json->entities->{$external_uri} );
113		}
114
115		if ( isset( $this->json->annotations ) ) {
116		foreach ( $this->json->annotations as $key => $annotation ) {
117		if ( isset( $annotation->entityMatches ) ) {
118		foreach ( $annotation->entityMatches as $match ) {
119		if ( isset( $match->entityId ) && isset( $mappings[ $match->entityId ] ) ) {
120		$match->entityId = $mappings[ $match->entityId ];
121		}
122		}
123		}
124		}
125		}
126
127		return $this;
128		}
129
130		/**
131		* Add occurrences by parsing the provided html content.
132		*
133		* @param string $content The html content with annotations.
134		*
135		* @return Analysis_Response_Ops The {@link Analysis_Response_Ops} instance.
136		*
137		* @since 3.23.7 refactor the regex pattern to take into account that there might be css classes between textannotation
138		* and disambiguated.
139		*
140		* @link https://github.com/insideout10/wordlift-plugin/issues/1001
141		*/
142		public function add_occurrences( $content ) {
143
144		// Try to get all the disambiguated annotations and bail out if an error occurs.
145		if ( false === preg_match_all(
146		'\|<span\s+id="([^"]+)"\s+class="textannotation\s+(?:\S+\s+)?disambiguated(?=[\s"])[^"]"\s+itemid="([^"])">(.*?)</span>\|',
147		$content,
148		$matches,
149		PREG_OFFSET_CAPTURE
150		) ) {
151		return $this;
152		}
153
154		if ( empty( $matches ) ) {
155		return $this;
156		}
157
158		$parse_data = array_reduce( range( 0, count( $matches[1] ) - 1 ), function ( $carry, $i ) use ( $matches ) {
159		if ( empty( $matches[0] ) ) {
160		return $carry;
161		}
162
163		$start = $matches[0][ $i ][1];
164		$end = $start + strlen( $matches[0][ $i ][0] );
165		$annotation_id = $matches[1][ $i ][0];
166		$item_id = $matches[2][ $i ][0];
167		$text = $matches[3][ $i ][0];
168
169		$annotation = new StdClass;
170		$annotation->annotationId = $annotation_id;
171		$annotation->start = $start;
172		$annotation->end = $end;
173		$annotation->text = $text;
174
175		$entity_match = new StdClass;
176		$entity_match->confidence = 100;
177		$entity_match->entityId = $item_id;
178		$annotation->entityMatches[] = $entity_match;
179
180		$carry['annotations'][ $annotation_id ] = $annotation;
181		$carry['occurrences'][ $item_id ][] = $annotation_id;
182
183		return $carry;
184		}, array( 'annotations' => array(), 'occurrences' => array(), ) );
185
186		$annotations = $parse_data['annotations'];
187		$occurrences = $parse_data['occurrences'];
188
189		foreach ( array_keys( $occurrences ) as $item_id ) {
190
191		// If the entity isn't there, add it.
192		if ( ! is_bool( $this->json ) && ! isset( $this->json->entities->{$item_id} ) ) {
193		$entity = $this->get_local_entity( $item_id );
194
195		// Entity not found in the local vocabulary, continue to the next one.
196		if ( false === $entity ) {
197		continue;
198		}
199
200		$this->json->entities->{$item_id} = $entity;
201		}
202		}
203
204		// Here we're adding back some data structures required by the client-side code.
205		//
206		// We're adding:
207		// 1. the .entities[entity_id].occurrences array with the annotations' ids.
208		// 2. the .entities[entity_id].annotations[annotation_id] = { id: annotation_id } map.
209		//
210		// Before 3.23.0 this was done by the client-side code located in src/coffee/editpost-widget/app.services.AnalysisService.coffee
211		// function `preselect`, which was called by src/coffee/editpost-widget/app.services.EditorService.coffee in
212		// `embedAnalysis`.
213
214		if ( ! is_bool( $this->json ) ) {
215		foreach ( $this->json->entities as $id => $entity ) {
216		$this->json->entities->{$id}->occurrences = isset( $occurrences[ $id ] ) ? $occurrences[ $id ] : array();;
217
218		foreach ( $this->json->entities->{$id}->occurrences as $annotation_id ) {
219		$this->json->entities->{$id}->annotations[ $annotation_id ] = array(
220		'id' => $annotation_id,
221		);
222		}
223		}
224		}
225
226		// Add the missing annotations. This allows the analysis response to work also if we didn't receive results
227		// from the analysis API.
228		foreach ( $annotations as $annotation_id => $annotation ) {
229
230		if ( ! is_bool( $this->json ) && ! isset( $this->json->annotations->{$annotation_id} ) ) {
231		$this->json->annotations->{$annotation_id} = $annotation;
232		}
233
234		}
235
236		return $this;
237		}
238
239		/**
240		* Add local entities
241		*
242		* @return Analysis_Response_Ops The {@link Analysis_Response_Ops} instance.
243		*
244		* @since 3.27.6
245		*
246		* @link https://github.com/insideout10/wordlift-plugin/issues/1178
247		*/
248		public function add_local_entities(){
249
250		// Populating the local entities object
251		$entities = array();
252		foreach ( $this->json->annotations as $annotation ) {
253		foreach ($annotation->entityMatches as $entity_matches){
254
255		$entity_id = url_to_postid( htmlspecialchars_decode($entity_matches->entityId) );
256		$serialized_entity = wl_serialize_entity( $entity_id );
257
258		if($serialized_entity){
259		$serialized_entity['entityId'] = $serialized_entity['id'];
260		unset($serialized_entity['id']);
261
262		$entities[$entity_matches->entityId] = $serialized_entity;
263		}
264
265		}
266		}
267
268		// Adding occurrences and annotations data structures required by the client-side code.
269		foreach ($entities as $entity_id => $entity ){
270		foreach ( $this->json->annotations as $annotation ) {
271		if($annotation->entityMatches[0]->entityId === $entity_id){
272		$entities[$entity_id]['occurrences'][] = $annotation->annotationId;
273		$entities[$entity_id]['annotations'][$annotation->annotationId]['id'] = $annotation->annotationId;
274		}
275		}
276		}
277
278		$this->json->entities = $entities;
279
280		return $this;
281
282		}
283
284		private function get_local_entity( $uri ) {
285
286		$entity = $this->entity_uri_service->get_entity( $uri );
287
288		if ( null === $entity ) {
289		return false;
290		}
291
292		$type = $this->entity_type_service->get( $entity->ID );
293		$images = $this->post_image_storage->get( $entity->ID );
294
295		return (object) array(
296		'id' => $uri,
297		'label' => $entity->post_title,
298		/*
299		* As of 2020.06.29 we're comment out the `post_content` because Gutenberg posts will return here
300		* the whole Gutenberg source including potentially our own wordlift/classification block, which means
301		* that data may grow quickly to more than a 100 KBytes and could break web servers.
302		*
303		* We don't really need the description for local entities (because the description is indeed taken from
304		* the local WordPress database) and we're not using it anywhere in the UI.
305		*
306		* So take extra care in enabling this line: eventually consider using the post_excerpt.
307		*
308		* PS: We didn't test using the WordLift Post Excerpt Helper.
309		*/
310		// 'description' => $entity->post_content,
311		'description' => '',
312		'sameAs' => wl_schema_get_value( $entity->ID, 'sameAs' ),
313		'mainType' => str_replace( 'wl-', '', $type['css_class'] ),
314		'types' => wl_get_entity_rdf_types( $entity->ID ),
315		'images' => $images,
316		);
317		}
318
319		/**
320		* Return the JSON response.
321		*
322		* @return mixed The JSON response.
323		* @since 3.24.2
324		*/
325		public function get_json() {
326
327		return $this->json;
328		}
329
330		/**
331		* Get the string representation of the JSON.
332		*
333		* @return false\|string The string representation or false in case of error.
334		*/
335		public function to_string() {
336
337		// Add the `JSON_UNESCAPED_UNICODE` only for PHP 5.4+.
338		$options = ( version_compare( PHP_VERSION, '5.4', '>=' )
339		? 256 : 0 );
340
341		return wp_json_encode( $this->json, $options );
342		}
343
344		}
345

insideout10 / wordlift-plugin

Push — develop ( d9328c...0bc50d )

Analysis_Response_Ops::add_occurrences() D

Complexity

Size

Duplication

Importance

How to fix Long Method Complexity

Long Method

Duplication Side-by-Side

Filter issues like