|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
/** |
|
4
|
|
|
* @file HTMLConverter.php |
|
5
|
|
|
* @brief This file contains the HTMLConverter class. |
|
6
|
|
|
* @details |
|
7
|
|
|
* @author Filippo F. Fadda |
|
8
|
|
|
*/ |
|
9
|
|
|
|
|
10
|
|
|
|
|
11
|
|
|
namespace Converter; |
|
12
|
|
|
|
|
13
|
|
|
/** |
|
14
|
|
|
* @brief A rudimental converter that takes as input HTML and replaces tags with related BBCodes. |
|
15
|
|
|
* @details This converter doesn't touch the HTML inside pre or code tags. |
|
16
|
|
|
*/ |
|
17
|
|
|
class HTMLConverter extends Converter { |
|
18
|
|
|
protected $snippets = []; |
|
19
|
|
|
|
|
20
|
|
|
|
|
21
|
|
|
/** |
|
22
|
|
|
* @brief Finds all code snippets inside the body, replacing them with appropriate markers. |
|
23
|
|
|
* @details The code can be inside `<pre></pre>`, `<code></code>`, or `[code][/code]` in case you are using BBCode |
|
24
|
|
|
* markup language. |
|
25
|
|
|
*/ |
|
26
|
|
|
protected function removeSnippets() { |
|
27
|
|
|
$pattern = '%(?P<openpre><pre>)(?P<contentpre>[\W\D\w\s]*?)(?P<closepre></pre>)|(?P<opencode><code>)(?P<contentcode>[\W\D\w\s]*?)(?P<closecode></code>)|(?P<openbbcode>\[code=?\w*\])(?P<contentbbcode>[\W\D\w\s]*?)(?P<closebbcode>\[/code\])%iu'; |
|
28
|
|
|
|
|
29
|
|
|
if (preg_match_all($pattern, $this->text, $this->snippets)) { |
|
30
|
|
|
|
|
31
|
|
|
$pattern = '%<pre>[\W\D\w\s]*?</pre>|<code>[\W\D\w\s]*?</code>|\[code=?\w*\][\W\D\w\s]*?\[/code\]%iu'; |
|
32
|
|
|
|
|
33
|
|
|
// Replaces the code snippet with a special marker to be able to inject the code in place. |
|
34
|
|
|
$this->text = preg_replace($pattern, '___SNIPPET___', $this->text); |
|
35
|
|
|
} |
|
36
|
|
|
} |
|
37
|
|
|
|
|
38
|
|
|
|
|
39
|
|
|
/** |
|
40
|
|
|
* @brief Restores the snippets, converting the HTML tags to BBCode tags. |
|
41
|
|
|
*/ |
|
42
|
|
|
protected function restoreSnippets() { |
|
43
|
|
|
$snippetsCount = count($this->snippets[0]); |
|
44
|
|
|
|
|
45
|
|
|
for ($i = 0; $i < $snippetsCount; $i++) { |
|
46
|
|
|
// We try to determine which tags the code is inside: <pre></pre>, <code></code>, [code][/code] |
|
47
|
|
|
if (!empty($this->snippets['openpre'][$i])) |
|
48
|
|
|
$snippet = "[code]".PHP_EOL.trim($this->snippets['contentpre'][$i]).PHP_EOL."[/code]"; |
|
49
|
|
|
elseif (!empty($this->snippets['opencode'][$i])) |
|
50
|
|
|
$snippet = "[code]".PHP_EOL.trim($this->snippets['contentcode'][$i]).PHP_EOL."[/code]"; |
|
51
|
|
|
else |
|
52
|
|
|
$snippet = $this->snippets['openbbcode'][$i].PHP_EOL.trim($this->snippets['contentbbcode'][$i]).PHP_EOL.$this->snippets['closebbcode'][$i]; |
|
53
|
|
|
|
|
54
|
|
|
$this->text = preg_replace('/___SNIPPET___/', PHP_EOL.trim($snippet).PHP_EOL, $this->text, 1); |
|
55
|
|
|
} |
|
56
|
|
|
} |
|
57
|
|
|
|
|
58
|
|
|
|
|
59
|
|
|
/** |
|
60
|
|
|
* @brief Replace links. |
|
61
|
|
|
*/ |
|
62
|
|
|
protected function replaceLinks() { |
|
63
|
|
|
|
|
64
|
|
|
$this->text = preg_replace_callback('%<a[^>]+>(.+?)</a>%iu', |
|
65
|
|
|
|
|
66
|
|
|
function ($matches) { |
|
67
|
|
|
|
|
68
|
|
|
// Extracts the url. |
|
69
|
|
|
if (preg_match('/\s*href\s*=\s*("([^"]*")|\'[^\']*\'|([^\'">\s]+))/iu', $matches[0], $others) === 1) { |
|
70
|
|
|
$href = trim($others[1], '"'); |
|
71
|
|
|
|
|
72
|
|
|
// Extracts the target. |
|
73
|
|
|
if (preg_match('/\s*target\s*=\s*("([^"]*")|\'[^\']*\'|([^\'">\s]+))/iu', $matches[0], $others) === 1) |
|
74
|
|
|
$target = strtolower(trim($others[1], '"')); |
|
75
|
|
|
else |
|
76
|
|
|
$target = "_self"; |
|
77
|
|
|
} |
|
78
|
|
|
else |
|
79
|
|
|
throw new \RuntimeException(sprintf("Text identified by '%d' has malformed links", $this->id)); |
|
80
|
|
|
|
|
81
|
|
|
return "[url=".$href." t=".$target."]".$matches[1]."[/url]"; |
|
82
|
|
|
|
|
83
|
|
|
}, |
|
84
|
|
|
|
|
85
|
|
|
$this->text |
|
86
|
|
|
); |
|
87
|
|
|
|
|
88
|
|
|
} |
|
89
|
|
|
|
|
90
|
|
|
|
|
91
|
|
|
/** |
|
92
|
|
|
* @brief Replace images. |
|
93
|
|
|
*/ |
|
94
|
|
|
protected function replaceImages() { |
|
95
|
|
|
$this->text = preg_replace_callback('/<img[^>]+>/iu', |
|
96
|
|
|
|
|
97
|
|
|
function ($matches) { |
|
98
|
|
|
|
|
99
|
|
|
// Extracts the src. |
|
100
|
|
|
if (preg_match('/\s*src\s*=\s*("([^"]*")|\'[^\']*\'|([^\'">\s]+))/iu', $matches[0], $others) === 1) |
|
101
|
|
|
$src = trim($others[1], '"'); |
|
102
|
|
|
else |
|
103
|
|
|
throw new \RuntimeException(sprintf("Text identified by '%d' has malformed images", $this->id)); |
|
104
|
|
|
|
|
105
|
|
|
return "[img]".$src."[/img]"; |
|
106
|
|
|
|
|
107
|
|
|
}, |
|
108
|
|
|
|
|
109
|
|
|
$this->text |
|
110
|
|
|
); |
|
111
|
|
|
|
|
112
|
|
|
} |
|
113
|
|
|
|
|
114
|
|
|
|
|
115
|
|
|
/** |
|
116
|
|
|
* @brief Replace all other simple tags, even the lists. |
|
117
|
|
|
*/ |
|
118
|
|
|
protected function replaceOtherTags() { |
|
119
|
|
|
$this->text = preg_replace_callback('%</?[a-z][a-z0-9]*[^<>]*>%iu', |
|
120
|
|
|
|
|
121
|
|
|
function ($matches) { |
|
122
|
|
|
$tag = strtolower($matches[0]); |
|
123
|
|
|
|
|
124
|
|
|
switch ($tag) { |
|
125
|
|
|
case ($tag == '<strong>' || $tag == '<b>'): |
|
126
|
|
|
return '[b]'; |
|
127
|
|
|
|
|
128
|
|
|
case ($tag == '</strong>' || $tag == '</b>'): |
|
129
|
|
|
return '[/b]'; |
|
130
|
|
|
|
|
131
|
|
|
case ($tag == '<em>' || $tag == '<i>'): |
|
132
|
|
|
return '[i]'; |
|
133
|
|
|
|
|
134
|
|
|
case ($tag == '</em>' || $tag == '</i>'): |
|
135
|
|
|
return '[/i]'; |
|
136
|
|
|
|
|
137
|
|
|
case '<u>': |
|
138
|
|
|
return '[u]'; |
|
139
|
|
|
|
|
140
|
|
|
case '</u>': |
|
141
|
|
|
return '[/u]'; |
|
142
|
|
|
|
|
143
|
|
|
case ($tag == '<strike>' || $tag == '<del>'): |
|
144
|
|
|
return '[s]'; |
|
145
|
|
|
|
|
146
|
|
|
case ($tag == '</strike>' || $tag == '</del>'): |
|
147
|
|
|
return '[/s]'; |
|
148
|
|
|
|
|
149
|
|
|
case '<ul>': |
|
150
|
|
|
return '[list]'; |
|
151
|
|
|
|
|
152
|
|
|
case '</ul>': |
|
153
|
|
|
return '[/list]'; |
|
154
|
|
|
|
|
155
|
|
|
case '<ol>': |
|
156
|
|
|
return '[list=1]'; |
|
157
|
|
|
|
|
158
|
|
|
case '</ol>': |
|
159
|
|
|
return '[/list]'; |
|
160
|
|
|
|
|
161
|
|
|
case '<li>': |
|
162
|
|
|
return '[*]'; |
|
163
|
|
|
|
|
164
|
|
|
case '</li>': |
|
165
|
|
|
return ''; |
|
166
|
|
|
|
|
167
|
|
|
case '<center>': |
|
168
|
|
|
return '[center]'; |
|
169
|
|
|
|
|
170
|
|
|
case '</center>': |
|
171
|
|
|
return '[/center]'; |
|
172
|
|
|
|
|
173
|
|
|
case ($tag == '<br>' || $tag == '<br/>' || $tag == '<br />'): |
|
174
|
|
|
return PHP_EOL; |
|
175
|
|
|
|
|
176
|
|
|
default: |
|
177
|
|
|
return $tag; |
|
178
|
|
|
} |
|
179
|
|
|
}, |
|
180
|
|
|
|
|
181
|
|
|
$this->text |
|
182
|
|
|
); |
|
183
|
|
|
|
|
184
|
|
|
} |
|
185
|
|
|
|
|
186
|
|
|
|
|
187
|
|
|
/** |
|
188
|
|
|
* @brief Converts the provided HTML text into BBCode. |
|
189
|
|
|
*/ |
|
190
|
|
|
public function toBBCode() { |
|
191
|
|
|
// We don't want any HTML entities. |
|
192
|
|
|
$this->text = htmlspecialchars_decode($this->text); |
|
193
|
|
|
|
|
194
|
|
|
$this->removeSnippets(); |
|
195
|
|
|
$this->replaceLinks(); |
|
196
|
|
|
$this->replaceImages(); |
|
197
|
|
|
$this->replaceOtherTags(); |
|
198
|
|
|
$this->text = strip_tags($this->text); |
|
199
|
|
|
$this->restoreSnippets(); |
|
200
|
|
|
|
|
201
|
|
|
return $this->text; |
|
202
|
|
|
} |
|
203
|
|
|
|
|
204
|
|
|
} |
|
205
|
|
|
|
mtawil /
converter
