|
1
|
|
|
<?php declare(strict_types = 1); |
|
2
|
|
|
|
|
3
|
|
|
/** |
|
4
|
|
|
* Created by PhpStorm. |
|
5
|
|
|
* User: gordon |
|
6
|
|
|
* Date: 25/3/2561 |
|
7
|
|
|
* Time: 17:01 น. |
|
8
|
|
|
*/ |
|
9
|
|
|
|
|
10
|
|
|
namespace Suilven\FreeTextSearch\Helper; |
|
11
|
|
|
|
|
12
|
|
|
use League\CLImate\CLImate; |
|
|
|
|
|
|
13
|
|
|
use SilverStripe\CMS\Model\SiteTree; |
|
14
|
|
|
use SilverStripe\Core\Config\Config; |
|
15
|
|
|
use SilverStripe\ORM\DataObject; |
|
16
|
|
|
use SilverStripe\ORM\DB; |
|
17
|
|
|
use SilverStripe\ORM\Queries\SQLUpdate; |
|
18
|
|
|
use SilverStripe\SiteConfig\SiteConfig; |
|
19
|
|
|
use Suilven\FreeTextSearch\Factory\BulkIndexerFactory; |
|
20
|
|
|
use Suilven\FreeTextSearch\Indexes; |
|
21
|
|
|
|
|
22
|
|
|
class BulkIndexingHelper |
|
23
|
|
|
{ |
|
24
|
|
|
/** |
|
25
|
|
|
* @param string $indexName |
|
26
|
|
|
* @param boolean $dirty Set this to true to only index 'dirty' DataObjects, false to reindex all |
|
27
|
|
|
* @param CLImate|null $climate |
|
28
|
|
|
*/ |
|
29
|
|
|
public function bulkIndex($indexName, $dirty = false, $climate = null) |
|
30
|
|
|
{ |
|
31
|
|
|
$indexes = new Indexes(); |
|
32
|
|
|
$index = $indexes->getIndex($indexName); |
|
33
|
|
|
|
|
34
|
|
|
/** @var string $clazz */ |
|
35
|
|
|
$clazz = $index->getClass(); |
|
36
|
|
|
|
|
37
|
|
|
|
|
38
|
|
|
$startTime = \microtime(true); |
|
39
|
|
|
|
|
40
|
|
|
if (!(is_null($climate))) { |
|
41
|
|
|
$climate->border('*'); |
|
42
|
|
|
$climate->green()->bold('Indexing sitetree'); |
|
43
|
|
|
$climate->border(); |
|
44
|
|
|
} |
|
45
|
|
|
|
|
46
|
|
|
$filters = ['ShowInSearch' => true]; |
|
47
|
|
|
if ($dirty) { |
|
48
|
|
|
$filters['IsDirtyFreeTextSearch'] = true; |
|
49
|
|
|
} |
|
50
|
|
|
|
|
51
|
|
|
$nDocuments = SiteTree::get()->filter($filters)->count(); |
|
52
|
|
|
|
|
53
|
|
|
error_log('N DOCUMENTS: ' . $nDocuments); |
|
54
|
|
|
|
|
55
|
|
|
if ($nDocuments > 0) { |
|
56
|
|
|
$config = SiteConfig::current_site_config(); |
|
57
|
|
|
|
|
58
|
|
|
// * @phpstan-ignore-next-line |
|
59
|
|
|
$bulkSize = $config->BulkSize; |
|
60
|
|
|
$pages = 1+\round($nDocuments / $bulkSize); |
|
61
|
|
|
$progress = !is_null($climate) ? $climate->progress()->total($nDocuments) : null; |
|
62
|
|
|
|
|
63
|
|
|
if (!is_null($climate)) { |
|
64
|
|
|
$climate->green('Pages: ' . $pages); |
|
65
|
|
|
$climate->green()->info('Indexing ' . $nDocuments .' objects'); |
|
66
|
|
|
} |
|
67
|
|
|
|
|
68
|
|
|
$factory = new BulkIndexerFactory(); |
|
69
|
|
|
$bulkIndexer = $factory->getBulkIndexer(); |
|
70
|
|
|
$bulkIndexer->setIndex($indexName); |
|
71
|
|
|
|
|
72
|
|
|
for ($i = 0; $i < $pages; $i++) { |
|
73
|
|
|
$dataObjects = $clazz::get()->limit($bulkSize, $i*$bulkSize)->filter($filters); |
|
74
|
|
|
foreach ($dataObjects as $do) { |
|
75
|
|
|
// Note this adds data to the payload, does not actually indexing against the third party search engine |
|
76
|
|
|
$bulkIndexer->addDataObject($do); |
|
77
|
|
|
} |
|
78
|
|
|
|
|
79
|
|
|
// index objects up to configured bulk size |
|
80
|
|
|
$bulkIndexer->indexDataObjects(); |
|
81
|
|
|
$current = $bulkSize * ($i+1); |
|
82
|
|
|
if ($current > $nDocuments) { |
|
83
|
|
|
$current = $nDocuments; |
|
84
|
|
|
} |
|
85
|
|
|
if (!is_null($progress)) { |
|
86
|
|
|
$progress->current($current); |
|
87
|
|
|
} |
|
88
|
|
|
} |
|
89
|
|
|
} |
|
90
|
|
|
|
|
91
|
|
|
|
|
92
|
|
|
$endTime = \microtime(true); |
|
93
|
|
|
$delta = $endTime-$startTime; |
|
94
|
|
|
|
|
95
|
|
|
$rate = \round($nDocuments / $delta, 2); |
|
96
|
|
|
|
|
97
|
|
|
$elapsedStr = \round($delta, 2); |
|
98
|
|
|
|
|
99
|
|
|
if (!is_null($climate)) { |
|
100
|
|
|
$climate->bold()->blue()->inline("{$nDocuments}"); |
|
101
|
|
|
$climate->blue()->inline(' objects indexed in '); |
|
102
|
|
|
$climate->bold()->blue()->inline("{$elapsedStr}"); |
|
103
|
|
|
$climate->blue()->inline('s, '); |
|
104
|
|
|
$climate->bold()->blue()->inline("{$rate}"); |
|
105
|
|
|
$climate->blue(' per second '); |
|
106
|
|
|
} |
|
107
|
|
|
|
|
108
|
|
|
$clazz = $index->getClass(); |
|
109
|
|
|
$table = Config::inst()->get($clazz, 'table_name'); |
|
110
|
|
|
|
|
111
|
|
|
|
|
112
|
|
|
DB::query("UPDATE \"{$table}\" SET \"IsDirtyFreeTextSearch\" = 0"); |
|
113
|
|
|
|
|
114
|
|
|
// @todo How to get the table name from versions? |
|
115
|
|
|
DB::query("UPDATE \"{$table}_Live\" SET \"IsDirtyFreeTextSearch\" = 0"); |
|
116
|
|
|
|
|
117
|
|
|
|
|
118
|
|
|
} |
|
119
|
|
|
} |
|
120
|
|
|
|
The issue could also be caused by a filter entry in the build configuration. If the path has been excluded in your configuration, e.g.
excluded_paths: ["lib/*"], you can move it to the dependency path list as follows:For further information see https://scrutinizer-ci.com/docs/tools/php/php-scrutinizer/#list-dependency-paths