|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
declare(strict_types=1); |
|
4
|
|
|
|
|
5
|
|
|
/** |
|
6
|
|
|
* This file is part of Scout Extended. |
|
7
|
|
|
* |
|
8
|
|
|
* (c) Algolia Team <[email protected]> |
|
9
|
|
|
* |
|
10
|
|
|
* For the full copyright and license information, please view the LICENSE |
|
11
|
|
|
* file that was distributed with this source code. |
|
12
|
|
|
*/ |
|
13
|
|
|
|
|
14
|
|
|
namespace Algolia\ScoutExtended\Splitters\HtmlSplitter; |
|
15
|
|
|
|
|
16
|
|
|
/** |
|
17
|
|
|
* @internal |
|
18
|
|
|
*/ |
|
19
|
|
|
final class NodeCollection |
|
20
|
|
|
{ |
|
21
|
|
|
/** |
|
22
|
|
|
* Collection of \Algolia\ScoutExtended\Splitters\HtmlSplitter\Node. |
|
23
|
|
|
* |
|
24
|
|
|
* @var array |
|
25
|
|
|
*/ |
|
26
|
|
|
private $nodes = []; |
|
27
|
|
|
|
|
28
|
|
|
/** |
|
29
|
|
|
* Clone of \Algolia\ScoutExtended\Splitters\HtmlSplitter\NodeCollection. |
|
30
|
|
|
* |
|
31
|
|
|
* @var array |
|
32
|
|
|
*/ |
|
33
|
|
|
private $cloneNodes = []; |
|
34
|
|
|
|
|
35
|
|
|
/** |
|
36
|
|
|
* The list of html tags. |
|
37
|
|
|
* |
|
38
|
|
|
* @var string[] |
|
39
|
|
|
*/ |
|
40
|
|
|
private $tags = []; |
|
41
|
|
|
|
|
42
|
|
|
/** |
|
43
|
|
|
* String. |
|
44
|
|
|
*/ |
|
45
|
|
|
private const IMPORTANCE = 'importance'; |
|
46
|
|
|
|
|
47
|
|
|
/** |
|
48
|
|
|
* String. |
|
49
|
|
|
*/ |
|
50
|
|
|
private const PARAGRAPH = 'p'; |
|
51
|
|
|
|
|
52
|
|
|
/** |
|
53
|
|
|
* NodeCollection constructor. |
|
54
|
|
|
* |
|
55
|
|
|
* @param array|null $tags |
|
56
|
|
|
*/ |
|
57
|
5 |
|
public function __construct(array $tags = null) |
|
58
|
|
|
{ |
|
59
|
5 |
|
if ($tags !== null) { |
|
60
|
5 |
|
$this->tags = $tags; |
|
61
|
|
|
} |
|
62
|
5 |
|
} |
|
63
|
|
|
|
|
64
|
|
|
/** |
|
65
|
|
|
* Add object to collection. |
|
66
|
|
|
* |
|
67
|
|
|
* @param Node $node |
|
68
|
|
|
*/ |
|
69
|
5 |
|
public function push(Node $node): void |
|
70
|
|
|
{ |
|
71
|
5 |
|
if ($this->lengthNodes() === 0) { |
|
72
|
5 |
|
$this->nodes[] = $node; |
|
73
|
5 |
|
$this->cloneNodes(); |
|
74
|
5 |
|
} elseif ($this->findWeight($node) > $this->findWeight(end($this->nodes))) { |
|
75
|
3 |
|
$this->nodes[] = $node; |
|
76
|
3 |
|
$this->cloneNodes(); |
|
77
|
|
|
} else { |
|
78
|
4 |
|
array_pop($this->nodes); |
|
79
|
4 |
|
$this->push($node); |
|
80
|
|
|
} |
|
81
|
5 |
|
} |
|
82
|
|
|
|
|
83
|
|
|
/** |
|
84
|
|
|
* Convert to array. |
|
85
|
|
|
* |
|
86
|
|
|
* @return array |
|
87
|
|
|
*/ |
|
88
|
5 |
|
public function toArray(): array |
|
89
|
|
|
{ |
|
90
|
5 |
|
$array = []; |
|
91
|
5 |
|
foreach ($this->cloneNodes as $nodes) { |
|
92
|
5 |
|
foreach ($nodes as $node) { |
|
93
|
5 |
|
if ($node instanceof Node) { |
|
94
|
5 |
|
$object[$node->getTag()] = $node->getContent(); |
|
95
|
|
|
} else { |
|
96
|
5 |
|
$object[self::IMPORTANCE] = $node; |
|
97
|
5 |
|
$array[] = $object; |
|
98
|
5 |
|
$object = []; |
|
99
|
|
|
} |
|
100
|
|
|
} |
|
101
|
|
|
} |
|
102
|
|
|
|
|
103
|
5 |
|
return $array; |
|
104
|
|
|
} |
|
105
|
|
|
|
|
106
|
|
|
/** |
|
107
|
|
|
* Importance need to be add after to avoid polluted queue. |
|
108
|
|
|
* |
|
109
|
|
|
* @return void |
|
110
|
|
|
*/ |
|
111
|
5 |
|
private function cloneNodes(): void |
|
112
|
|
|
{ |
|
113
|
5 |
|
$this->cloneNodes[] = $this->nodes; |
|
114
|
5 |
|
$this->cloneNodes[] = [self::IMPORTANCE => $this->importanceWeight(end($this->nodes))]; |
|
115
|
5 |
|
} |
|
116
|
|
|
|
|
117
|
|
|
/** |
|
118
|
|
|
* Importance formula. |
|
119
|
|
|
* Give integer from tags ranking. |
|
120
|
|
|
* |
|
121
|
|
|
* @param Node $node |
|
122
|
|
|
* |
|
123
|
|
|
* @return int |
|
124
|
|
|
*/ |
|
125
|
5 |
|
private function importanceWeight(Node $node): int |
|
126
|
|
|
{ |
|
127
|
5 |
|
if ($node->getTag() === self::PARAGRAPH) { |
|
128
|
3 |
|
$object = prev($this->nodes); |
|
129
|
3 |
|
if (empty(end($this->nodes)) || $this->lengthNodes() === 1) { |
|
130
|
|
|
return 0; |
|
131
|
|
|
} |
|
132
|
|
|
|
|
133
|
3 |
|
return (count($this->tags) - 1) + $this->findWeight($object); |
|
134
|
|
|
} |
|
135
|
|
|
|
|
136
|
5 |
|
return $this->findWeight($node); |
|
137
|
|
|
} |
|
138
|
|
|
|
|
139
|
|
|
/** |
|
140
|
|
|
* Find weight of current nodes. |
|
141
|
|
|
* |
|
142
|
|
|
* @param Node $node |
|
143
|
|
|
* |
|
144
|
|
|
* @return int |
|
145
|
|
|
*/ |
|
146
|
5 |
|
private function findWeight(Node $node): int |
|
147
|
|
|
{ |
|
148
|
5 |
|
return (int) array_search($node->getTag(), $this->tags, true); |
|
149
|
|
|
} |
|
150
|
|
|
|
|
151
|
|
|
/** |
|
152
|
|
|
* Give the length of the collection. |
|
153
|
|
|
* |
|
154
|
|
|
* @return int |
|
155
|
|
|
*/ |
|
156
|
5 |
|
private function lengthNodes(): int |
|
157
|
|
|
{ |
|
158
|
5 |
|
return count($this->nodes); |
|
159
|
|
|
} |
|
160
|
|
|
} |
|
161
|
|
|
|
algolia /
scout-extended
