|
1
|
|
|
<?php |
|
2
|
|
|
namespace ApacheSolrForTypo3\Solr\Domain\Search\ApacheSolrDocument; |
|
3
|
|
|
|
|
4
|
|
|
/*************************************************************** |
|
5
|
|
|
* Copyright notice |
|
6
|
|
|
* |
|
7
|
|
|
* (c) 2017 Timo Hund <[email protected]> |
|
8
|
|
|
* All rights reserved |
|
9
|
|
|
* |
|
10
|
|
|
* This script is part of the TYPO3 project. The TYPO3 project is |
|
11
|
|
|
* free software; you can redistribute it and/or modify |
|
12
|
|
|
* it under the terms of the GNU General Public License as published by |
|
13
|
|
|
* the Free Software Foundation; either version 2 of the License, or |
|
14
|
|
|
* (at your option) any later version. |
|
15
|
|
|
* |
|
16
|
|
|
* The GNU General Public License can be found at |
|
17
|
|
|
* http://www.gnu.org/copyleft/gpl.html. |
|
18
|
|
|
* |
|
19
|
|
|
* This script is distributed in the hope that it will be useful, |
|
20
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
21
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
22
|
|
|
* GNU General Public License for more details. |
|
23
|
|
|
* |
|
24
|
|
|
* This copyright notice MUST APPEAR in all copies of the script! |
|
25
|
|
|
***************************************************************/ |
|
26
|
|
|
|
|
27
|
|
|
use Apache_Solr_Document; |
|
28
|
|
|
use ApacheSolrForTypo3\Solr\Access\Rootline; |
|
29
|
|
|
use ApacheSolrForTypo3\Solr\Domain\Site\SiteRepository; |
|
30
|
|
|
use ApacheSolrForTypo3\Solr\Domain\Variants\IdBuilder; |
|
31
|
|
|
use ApacheSolrForTypo3\Solr\Site; |
|
32
|
|
|
use ApacheSolrForTypo3\Solr\Typo3PageContentExtractor; |
|
33
|
|
|
use ApacheSolrForTypo3\Solr\Util; |
|
34
|
|
|
use TYPO3\CMS\Core\Utility\GeneralUtility; |
|
35
|
|
|
use TYPO3\CMS\Frontend\Controller\TypoScriptFrontendController; |
|
36
|
|
|
|
|
37
|
|
|
/** |
|
38
|
|
|
* Builder class to build an ApacheSolrDocument |
|
39
|
|
|
* |
|
40
|
|
|
* Responsible to build Apache_Solr_Documents |
|
41
|
|
|
* |
|
42
|
|
|
* @author Timo Hund <[email protected]> |
|
43
|
|
|
*/ |
|
44
|
|
|
class Builder |
|
45
|
|
|
{ |
|
46
|
|
|
/** |
|
47
|
|
|
* @var IdBuilder |
|
48
|
|
|
*/ |
|
49
|
|
|
protected $variantIdBuilder; |
|
50
|
|
|
|
|
51
|
|
|
/** |
|
52
|
|
|
* Builder constructor. |
|
53
|
|
|
* @param IdBuilder|null $variantIdBuilder |
|
54
|
|
|
*/ |
|
55
|
59 |
|
public function __construct(IdBuilder $variantIdBuilder = null) |
|
56
|
|
|
{ |
|
57
|
59 |
|
$this->variantIdBuilder = is_null($variantIdBuilder) ? GeneralUtility::makeInstance(IdBuilder::class) : $variantIdBuilder; |
|
58
|
59 |
|
} |
|
59
|
|
|
|
|
60
|
|
|
/** |
|
61
|
|
|
* This method can be used to build an Apache_Solr_Document from a TYPO3 page. |
|
62
|
|
|
* |
|
63
|
|
|
* @param TypoScriptFrontendController $page |
|
64
|
|
|
* @param string $url |
|
65
|
|
|
* @param Rootline $pageAccessRootline |
|
66
|
|
|
* @param string $mountPointParameter |
|
67
|
|
|
* @return Apache_Solr_Document|object |
|
68
|
|
|
*/ |
|
69
|
59 |
|
public function fromPage(TypoScriptFrontendController $page, $url, Rootline $pageAccessRootline, $mountPointParameter): \Apache_Solr_Document |
|
70
|
|
|
{ |
|
71
|
|
|
/* @var $document \Apache_Solr_Document */ |
|
72
|
59 |
|
$document = GeneralUtility::makeInstance(Apache_Solr_Document::class); |
|
73
|
59 |
|
$site = $this->getSiteByPageId($page->id); |
|
74
|
59 |
|
$pageRecord = $page->page; |
|
75
|
|
|
|
|
76
|
59 |
|
$accessGroups = $this->getDocumentIdGroups($pageAccessRootline); |
|
77
|
59 |
|
$documentId = $this->getPageDocumentId($page, $accessGroups, $mountPointParameter); |
|
78
|
|
|
|
|
79
|
59 |
|
$document->setField('id', $documentId); |
|
80
|
59 |
|
$document->setField('site', $site->getDomain()); |
|
81
|
59 |
|
$document->setField('siteHash', $site->getSiteHash()); |
|
82
|
59 |
|
$document->setField('appKey', 'EXT:solr'); |
|
83
|
59 |
|
$document->setField('type', 'pages'); |
|
84
|
|
|
|
|
85
|
|
|
// system fields |
|
86
|
59 |
|
$document->setField('uid', $page->id); |
|
87
|
59 |
|
$document->setField('pid', $pageRecord['pid']); |
|
88
|
|
|
|
|
89
|
|
|
// variantId |
|
90
|
59 |
|
$variantId = $this->variantIdBuilder->buildFromTypeAndUid('pages', $page->id); |
|
91
|
59 |
|
$document->setField('variantId', $variantId); |
|
92
|
|
|
|
|
93
|
59 |
|
$document->setField('typeNum', $page->type); |
|
94
|
59 |
|
$document->setField('created', $pageRecord['crdate']); |
|
95
|
59 |
|
$document->setField('changed', $pageRecord['SYS_LASTCHANGED']); |
|
96
|
|
|
|
|
97
|
59 |
|
$rootline = $this->getRootLineFieldValue($page->id, $mountPointParameter); |
|
98
|
59 |
|
$document->setField('rootline', $rootline); |
|
99
|
|
|
|
|
100
|
|
|
// access |
|
101
|
59 |
|
$this->addAccessField($document, $pageAccessRootline); |
|
102
|
59 |
|
$this->addEndtimeField($document, $pageRecord); |
|
103
|
|
|
|
|
104
|
|
|
// content |
|
105
|
59 |
|
$contentExtractor = $this->getExtractorForPageContent($page->content); |
|
106
|
59 |
|
$document->setField('title', $contentExtractor->getPageTitle()); |
|
107
|
59 |
|
$document->setField('subTitle', $pageRecord['subtitle']); |
|
108
|
59 |
|
$document->setField('navTitle', $pageRecord['nav_title']); |
|
109
|
59 |
|
$document->setField('author', $pageRecord['author']); |
|
110
|
59 |
|
$document->setField('description', $pageRecord['description']); |
|
111
|
59 |
|
$document->setField('abstract', $pageRecord['abstract']); |
|
112
|
59 |
|
$document->setField('content', $contentExtractor->getIndexableContent()); |
|
113
|
59 |
|
$document->setField('url', $url); |
|
114
|
|
|
|
|
115
|
59 |
|
$this->addKeywordsField($document, $pageRecord); |
|
116
|
59 |
|
$this->addTagContentFields($document, $contentExtractor->getTagContent()); |
|
117
|
|
|
|
|
118
|
59 |
|
return $document; |
|
119
|
|
|
} |
|
120
|
|
|
|
|
121
|
|
|
|
|
122
|
|
|
/** |
|
123
|
|
|
* Creates a Solr document with the basic / core fields set already. |
|
124
|
|
|
* |
|
125
|
|
|
* @param array $itemRecord |
|
126
|
|
|
* @param string $type |
|
127
|
55 |
|
* @param int $rootPageUid |
|
128
|
|
|
* @param string $accessRootLine |
|
129
|
55 |
|
* @return Apache_Solr_Document |
|
130
|
|
|
*/ |
|
131
|
|
|
public function fromRecord(array $itemRecord, string $type, int $rootPageUid, string $accessRootLine): \Apache_Solr_Document |
|
132
|
|
|
{ |
|
133
|
|
|
/* @var $document Apache_Solr_Document */ |
|
134
|
|
|
$document = GeneralUtility::makeInstance(Apache_Solr_Document::class); |
|
135
|
|
|
|
|
136
|
55 |
|
$site = $this->getSiteByPageId($rootPageUid); |
|
137
|
|
|
|
|
138
|
55 |
|
$documentId = $this->getDocumentId($type, $site->getRootPageId(), $itemRecord['uid']); |
|
139
|
55 |
|
|
|
140
|
|
|
// required fields |
|
141
|
|
|
$document->setField('id', $documentId); |
|
142
|
|
|
$document->setField('type', $type); |
|
143
|
|
|
$document->setField('appKey', 'EXT:solr'); |
|
144
|
|
|
|
|
145
|
|
|
// site, siteHash |
|
146
|
55 |
|
$document->setField('site', $site->getDomain()); |
|
147
|
|
|
$document->setField('siteHash', $site->getSiteHash()); |
|
148
|
55 |
|
|
|
149
|
|
|
// uid, pid |
|
150
|
|
|
$document->setField('uid', $itemRecord['uid']); |
|
151
|
|
|
$document->setField('pid', $itemRecord['pid']); |
|
152
|
|
|
|
|
153
|
|
|
// variantId |
|
154
|
|
|
$variantId = $this->variantIdBuilder->buildFromTypeAndUid($type, $itemRecord['uid']); |
|
155
|
|
|
$document->setField('variantId', $variantId); |
|
156
|
|
|
|
|
157
|
|
|
// created, changed |
|
158
|
59 |
|
if (!empty($GLOBALS['TCA'][$type]['ctrl']['crdate'])) { |
|
159
|
|
|
$document->setField('created', $itemRecord[$GLOBALS['TCA'][$type]['ctrl']['crdate']]); |
|
160
|
59 |
|
} |
|
161
|
59 |
|
if (!empty($GLOBALS['TCA'][$type]['ctrl']['tstamp'])) { |
|
162
|
47 |
|
$document->setField('changed', $itemRecord[$GLOBALS['TCA'][$type]['ctrl']['tstamp']]); |
|
163
|
|
|
} |
|
164
|
59 |
|
|
|
165
|
|
|
// access, endtime |
|
166
|
|
|
$document->setField('access', $accessRootLine); |
|
167
|
|
|
if (!empty($GLOBALS['TCA'][$type]['ctrl']['enablecolumns']['endtime']) |
|
168
|
|
|
&& $itemRecord[$GLOBALS['TCA'][$type]['ctrl']['enablecolumns']['endtime']] != 0 |
|
169
|
|
|
) { |
|
170
|
|
|
$document->setField('endtime', $itemRecord[$GLOBALS['TCA'][$type]['ctrl']['enablecolumns']['endtime']]); |
|
171
|
|
|
} |
|
172
|
|
|
|
|
173
|
|
|
return $document; |
|
174
|
59 |
|
} |
|
175
|
|
|
|
|
176
|
59 |
|
/** |
|
177
|
59 |
|
* @param TypoScriptFrontendController $page |
|
178
|
|
|
* @param string $accessGroups |
|
179
|
59 |
|
* @param string $mountPointParameter |
|
180
|
13 |
|
* @return string |
|
181
|
|
|
*/ |
|
182
|
|
|
protected function getPageDocumentId(TypoScriptFrontendController $page, string $accessGroups, string $mountPointParameter): string |
|
183
|
59 |
|
{ |
|
184
|
|
|
return Util::getPageDocumentId($page->id, $page->type, $page->sys_language_uid, $accessGroups, $mountPointParameter); |
|
185
|
59 |
|
} |
|
186
|
|
|
|
|
187
|
|
|
/** |
|
188
|
|
|
* @param string $type |
|
189
|
|
|
* @param int $rootPageId |
|
190
|
|
|
* @param int $recordUid |
|
191
|
|
|
* @return string |
|
192
|
|
|
*/ |
|
193
|
|
|
protected function getDocumentId(string $type, int $rootPageId, int $recordUid): string |
|
194
|
59 |
|
{ |
|
195
|
|
|
return Util::getDocumentId($type, $rootPageId, $recordUid); |
|
196
|
59 |
|
} |
|
197
|
59 |
|
|
|
198
|
44 |
|
/** |
|
199
|
|
|
* @param integer $pageId |
|
200
|
59 |
|
* @return Site |
|
201
|
|
|
*/ |
|
202
|
|
|
protected function getSiteByPageId($pageId) |
|
203
|
|
|
{ |
|
204
|
|
|
$siteRepository = GeneralUtility::makeInstance(SiteRepository::class); |
|
205
|
|
|
return $siteRepository->getSiteByPageId($pageId); |
|
206
|
|
|
} |
|
207
|
|
|
|
|
208
|
59 |
|
/** |
|
209
|
|
|
* @param string $pageContent |
|
210
|
59 |
|
* @return Typo3PageContentExtractor |
|
211
|
1 |
|
*/ |
|
212
|
|
|
protected function getExtractorForPageContent($pageContent) |
|
213
|
59 |
|
{ |
|
214
|
|
|
return GeneralUtility::makeInstance(Typo3PageContentExtractor::class, $pageContent); |
|
215
|
|
|
} |
|
216
|
|
|
|
|
217
|
|
|
/** |
|
218
|
|
|
* Builds the content for the rootline field. |
|
219
|
|
|
* |
|
220
|
|
|
* @param int $pageId |
|
221
|
59 |
|
* @param string $mountPointParameter |
|
222
|
|
|
* @return string |
|
223
|
59 |
|
*/ |
|
224
|
58 |
|
protected function getRootLineFieldValue($pageId, $mountPointParameter) |
|
225
|
|
|
{ |
|
226
|
|
|
$rootline = $pageId; |
|
227
|
1 |
|
if ($mountPointParameter !== '') { |
|
228
|
1 |
|
$rootline .= ',' . $mountPointParameter; |
|
229
|
1 |
|
} |
|
230
|
|
|
return $rootline; |
|
231
|
1 |
|
} |
|
232
|
|
|
|
|
233
|
|
|
/** |
|
234
|
|
|
* Gets a comma separated list of frontend user groups to use for the |
|
235
|
|
|
* document ID. |
|
236
|
|
|
* |
|
237
|
|
|
* @param Rootline $pageAccessRootline |
|
238
|
|
|
* @return string A comma separated list of frontend user groups. |
|
239
|
59 |
|
*/ |
|
240
|
|
|
protected function getDocumentIdGroups(Rootline $pageAccessRootline) |
|
241
|
59 |
|
{ |
|
242
|
1 |
|
$groups = $pageAccessRootline->getGroups(); |
|
243
|
|
|
$groups = Rootline::cleanGroupArray($groups); |
|
244
|
59 |
|
|
|
245
|
|
|
if (empty($groups)) { |
|
246
|
|
|
$groups[] = 0; |
|
247
|
|
|
} |
|
248
|
|
|
|
|
249
|
|
|
$groups = implode(',', $groups); |
|
250
|
|
|
|
|
251
|
|
|
return $groups; |
|
252
|
|
|
} |
|
253
|
|
|
|
|
254
|
|
|
/** |
|
255
|
|
|
* Adds the access field to the document if needed. |
|
256
|
|
|
* |
|
257
|
|
|
* @param \Apache_Solr_Document $document |
|
258
|
|
|
* @param Rootline $pageAccessRootline |
|
259
|
|
|
*/ |
|
260
|
|
|
protected function addAccessField(\Apache_Solr_Document $document, Rootline $pageAccessRootline) |
|
261
|
|
|
{ |
|
262
|
|
|
$access = (string)$pageAccessRootline; |
|
263
|
|
|
if (trim($access) !== '') { |
|
264
|
|
|
$document->setField('access', $access); |
|
265
|
|
|
} |
|
266
|
|
|
} |
|
267
|
|
|
|
|
268
|
|
|
/** |
|
269
|
|
|
* Adds the endtime field value to the Apache_Solr_Document. |
|
270
|
|
|
* |
|
271
|
|
|
* @param \Apache_Solr_Document $document |
|
272
|
|
|
* @param array $pageRecord |
|
273
|
|
|
*/ |
|
274
|
|
|
protected function addEndtimeField(\Apache_Solr_Document $document, $pageRecord) |
|
275
|
|
|
{ |
|
276
|
|
|
if ($pageRecord['endtime']) { |
|
277
|
|
|
$document->setField('endtime', $pageRecord['endtime']); |
|
278
|
|
|
} |
|
279
|
|
|
} |
|
280
|
|
|
|
|
281
|
|
|
/** |
|
282
|
|
|
* Adds keywords, multi valued. |
|
283
|
|
|
* |
|
284
|
|
|
* @param \Apache_Solr_Document $document |
|
285
|
|
|
* @param array $pageRecord |
|
286
|
|
|
*/ |
|
287
|
|
|
protected function addKeywordsField(\Apache_Solr_Document $document, $pageRecord) |
|
288
|
|
|
{ |
|
289
|
|
|
if (!isset($pageRecord['keywords'])) { |
|
290
|
|
|
return; |
|
291
|
|
|
} |
|
292
|
|
|
|
|
293
|
|
|
$keywords = array_unique(GeneralUtility::trimExplode(',', $pageRecord['keywords'], true)); |
|
294
|
|
|
foreach ($keywords as $keyword) { |
|
295
|
|
|
$document->addField('keywords', $keyword); |
|
296
|
|
|
} |
|
297
|
|
|
} |
|
298
|
|
|
|
|
299
|
|
|
/** |
|
300
|
|
|
* Add content from several tags like headers, anchors, ... |
|
301
|
|
|
* |
|
302
|
|
|
* @param \Apache_Solr_Document $document |
|
303
|
|
|
* @param array $tagContent |
|
304
|
|
|
*/ |
|
305
|
|
|
protected function addTagContentFields(\Apache_Solr_Document $document, $tagContent = []) |
|
306
|
|
|
{ |
|
307
|
|
|
foreach ($tagContent as $fieldName => $fieldValue) { |
|
308
|
|
|
$document->setField($fieldName, $fieldValue); |
|
309
|
|
|
} |
|
310
|
|
|
} |
|
311
|
|
|
} |
|
312
|
|
|
|