|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
namespace Bolt\Helpers; |
|
4
|
|
|
|
|
5
|
|
|
use Bolt\Legacy\Content; |
|
6
|
|
|
|
|
7
|
|
|
class Excerpt |
|
8
|
|
|
{ |
|
9
|
|
|
/** @var Content|array|string */ |
|
10
|
|
|
protected $body; |
|
11
|
|
|
/** @var string */ |
|
12
|
|
|
protected $title; |
|
13
|
|
|
|
|
14
|
|
|
/** |
|
15
|
|
|
* Constructor. |
|
16
|
|
|
* |
|
17
|
|
|
* @param Content|array|string $body |
|
18
|
|
|
* @param string|null $title |
|
19
|
|
|
*/ |
|
20
|
|
|
public function __construct($body, $title = null) |
|
21
|
|
|
{ |
|
22
|
|
|
$this->body = $body; |
|
23
|
|
|
$this->title = trim($title); |
|
24
|
|
|
} |
|
25
|
|
|
|
|
26
|
|
|
/** |
|
27
|
|
|
* Get the excerpt of a given piece of text. |
|
28
|
|
|
* |
|
29
|
|
|
* @param int $length |
|
30
|
|
|
* @param bool $includeTitle |
|
31
|
|
|
* @param array|string|null $focus |
|
32
|
|
|
* |
|
33
|
|
|
* @return string|null |
|
34
|
|
|
*/ |
|
35
|
|
|
public function getExcerpt($length = 200, $includeTitle = false, $focus = null) |
|
36
|
|
|
{ |
|
37
|
|
|
$title = null; |
|
38
|
|
|
if ($includeTitle && $this->title !== null) { |
|
39
|
|
|
$title = Html::trimText(strip_tags($this->title), $length); |
|
40
|
|
|
$length = $length - strlen($title); |
|
41
|
|
|
} |
|
42
|
|
|
|
|
43
|
|
|
if ($this->body instanceof Content) { |
|
44
|
|
|
$this->body = $this->body->getValues(); |
|
45
|
|
|
} |
|
46
|
|
|
|
|
47
|
|
|
if (is_array($this->body)) { |
|
48
|
|
|
// Assume it's an array, strip some common fields that we don't need, implode the rest. |
|
49
|
|
|
$stripKeys = [ |
|
50
|
|
|
'id', |
|
51
|
|
|
'slug', |
|
52
|
|
|
'datecreated', |
|
53
|
|
|
'datechanged', |
|
54
|
|
|
'username', |
|
55
|
|
|
'ownerid', |
|
56
|
|
|
'title', |
|
57
|
|
|
'contenttype', |
|
58
|
|
|
'status', |
|
59
|
|
|
'taxonomy', |
|
60
|
|
|
'templatefields', |
|
61
|
|
|
]; |
|
62
|
|
|
|
|
63
|
|
|
foreach ($stripKeys as $key) { |
|
64
|
|
|
unset($this->body[$key]); |
|
65
|
|
|
} |
|
66
|
|
|
$excerpt = implode(' ', $this->body); |
|
67
|
|
|
} elseif (is_string($this->body) || (is_object($this->body) && method_exists($this->body, '__toString'))) { |
|
68
|
|
|
// otherwise we just use the string. |
|
69
|
|
|
$excerpt = (string) $this->body; |
|
70
|
|
|
} else { |
|
71
|
|
|
// Nope, got nothing. |
|
72
|
|
|
$excerpt = ''; |
|
73
|
|
|
} |
|
74
|
|
|
|
|
75
|
|
|
$excerpt = str_replace('>', '> ', $excerpt); |
|
76
|
|
|
|
|
77
|
|
|
if (empty($focus)) { |
|
78
|
|
|
$excerpt = Html::trimText(strip_tags($excerpt), $length); |
|
79
|
|
|
} else { |
|
80
|
|
|
$excerpt = $this->extractRelevant($focus, strip_tags($excerpt), $length); |
|
81
|
|
|
} |
|
82
|
|
|
|
|
83
|
|
|
if ($title !== null) { |
|
84
|
|
|
$excerpt = '<b>' . $title . '</b> ' . $excerpt; |
|
85
|
|
|
} |
|
86
|
|
|
|
|
87
|
|
|
return $excerpt; |
|
88
|
|
|
} |
|
89
|
|
|
|
|
90
|
|
|
/** |
|
91
|
|
|
* Find the locations of each of the words. |
|
92
|
|
|
* Nothing exciting here. The array_unique is required, unless you decide |
|
93
|
|
|
* to make the words unique before passing in. |
|
94
|
|
|
* |
|
95
|
|
|
* @param array $words |
|
96
|
|
|
* @param string $fulltext |
|
97
|
|
|
* |
|
98
|
|
|
* @return array |
|
99
|
|
|
*/ |
|
100
|
|
|
private function extractLocations(array $words, $fulltext) |
|
101
|
|
|
{ |
|
102
|
|
|
$locations = []; |
|
103
|
|
|
foreach ($words as $word) { |
|
104
|
|
|
$wordLen = strlen($word); |
|
105
|
|
|
$loc = stripos($fulltext, $word); |
|
106
|
|
|
while ($loc !== false) { |
|
107
|
|
|
$locations[] = $loc; |
|
108
|
|
|
$loc = stripos($fulltext, $word, $loc + $wordLen); |
|
109
|
|
|
} |
|
110
|
|
|
} |
|
111
|
|
|
$locations = array_unique($locations); |
|
112
|
|
|
sort($locations); |
|
113
|
|
|
|
|
114
|
|
|
return $locations; |
|
115
|
|
|
} |
|
116
|
|
|
|
|
117
|
|
|
/** |
|
118
|
|
|
* Work out which is the most relevant portion to display |
|
119
|
|
|
* This is done by looping over each match and finding the smallest distance between two found |
|
120
|
|
|
* strings. The idea being that the closer the terms are the better match the snippet would be. |
|
121
|
|
|
* When checking for matches we only change the location if there is a better match. |
|
122
|
|
|
* The only exception is where we have only two matches in which case we just take the |
|
123
|
|
|
* first as will be equally distant. |
|
124
|
|
|
* |
|
125
|
|
|
* @param array $locations |
|
126
|
|
|
* @param integer $prevCount |
|
127
|
|
|
* |
|
128
|
|
|
* @return int |
|
129
|
|
|
*/ |
|
130
|
|
|
private function determineSnipLocation(array $locations, $prevCount) |
|
131
|
|
|
{ |
|
132
|
|
|
// If we only have 1 match we don't actually do the for loop so set to the first |
|
133
|
|
|
$startPos = $locations[0]; |
|
134
|
|
|
$loccount = count($locations); |
|
135
|
|
|
$smallestDiff = PHP_INT_MAX; |
|
136
|
|
|
|
|
137
|
|
|
// If we only have 2, skip as it's probably equally relevant |
|
138
|
|
|
if (count($locations) > 2) { |
|
139
|
|
|
// skip the first as we check 1 behind |
|
140
|
|
|
for ($i = 1; $i < $loccount; $i++) { |
|
141
|
|
|
if ($i === $loccount - 1) { // at the end |
|
142
|
|
|
$diff = $locations[$i] - $locations[$i - 1]; |
|
143
|
|
|
} else { |
|
144
|
|
|
$diff = $locations[$i + 1] - $locations[$i]; |
|
145
|
|
|
} |
|
146
|
|
|
|
|
147
|
|
|
if ($smallestDiff > $diff) { |
|
148
|
|
|
$smallestDiff = $diff; |
|
149
|
|
|
$startPos = $locations[$i]; |
|
150
|
|
|
} |
|
151
|
|
|
} |
|
152
|
|
|
} |
|
153
|
|
|
|
|
154
|
|
|
$startPos = $startPos > $prevCount ? $startPos - $prevCount : 0; |
|
155
|
|
|
|
|
156
|
|
|
return $startPos; |
|
157
|
|
|
} |
|
158
|
|
|
|
|
159
|
|
|
/** |
|
160
|
|
|
* Center on, and highlight search terms in excerpts. |
|
161
|
|
|
* |
|
162
|
|
|
* @see: http://www.boyter.org/2013/04/building-a-search-result-extract-generator-in-php/ |
|
163
|
|
|
* |
|
164
|
|
|
* @param string|array $words |
|
165
|
|
|
* @param string $fulltext |
|
166
|
|
|
* @param integer $relLength |
|
167
|
|
|
* |
|
168
|
|
|
* @return mixed|string |
|
169
|
|
|
*/ |
|
170
|
|
|
private function extractRelevant($words, $fulltext, $relLength = 300) |
|
171
|
|
|
{ |
|
172
|
|
|
if (!is_array($words)) { |
|
173
|
|
|
$words = explode(' ', $words); |
|
174
|
|
|
} |
|
175
|
|
|
|
|
176
|
|
|
// 1/6 ratio on prevcount tends to work pretty well and puts the terms |
|
177
|
|
|
// in the middle of the extract |
|
178
|
|
|
$prevCount = floor($relLength / 6); |
|
179
|
|
|
|
|
180
|
|
|
$indicator = '…'; |
|
181
|
|
|
|
|
182
|
|
|
$textlength = strlen($fulltext); |
|
183
|
|
|
if ($textlength <= $relLength) { |
|
184
|
|
|
return $fulltext; |
|
185
|
|
|
} |
|
186
|
|
|
|
|
187
|
|
|
$locations = $this->extractLocations($words, $fulltext); |
|
188
|
|
|
$startPos = $this->determineSnipLocation($locations, $prevCount); |
|
189
|
|
|
|
|
190
|
|
|
// if we are going to snip too much... |
|
191
|
|
|
if ($textlength - $startPos < $relLength) { |
|
192
|
|
|
$startPos = $startPos - ($textlength - $startPos) / 2; |
|
193
|
|
|
} |
|
194
|
|
|
|
|
195
|
|
|
$relText = substr($fulltext, $startPos, $relLength); |
|
196
|
|
|
|
|
197
|
|
|
// check to ensure we dont snip the last word if thats the match |
|
198
|
|
|
if ($startPos + $relLength < $textlength) { |
|
199
|
|
|
$relText = substr($relText, 0, strrpos($relText, ' ')) . $indicator; // remove last word |
|
200
|
|
|
} |
|
201
|
|
|
|
|
202
|
|
|
// If we trimmed from the front add '…' |
|
203
|
|
|
if ($startPos != 0) { |
|
204
|
|
|
$relText = $indicator . substr($relText, strpos($relText, ' ') + 1); // remove first word |
|
205
|
|
|
} |
|
206
|
|
|
|
|
207
|
|
|
// Highlight the words, using the `<mark>` tag. |
|
208
|
|
|
foreach ($words as $word) { |
|
209
|
|
|
$relText = preg_replace('/\b(' . $word . ')\b/i', '<mark>$1</mark>', $relText); |
|
210
|
|
|
} |
|
211
|
|
|
|
|
212
|
|
|
return $relText; |
|
213
|
|
|
} |
|
214
|
|
|
} |
|
215
|
|
|
|
bolt /
bolt
