|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
namespace Integrations\Connectors\Wikipedia; |
|
4
|
|
|
|
|
5
|
|
|
use Log; |
|
6
|
|
|
use App\Models\User; |
|
7
|
|
|
|
|
8
|
|
|
/* PHP-Wiki-API: This is a simple class to get short Wikipedia info boxes from a given Keyword. |
|
9
|
|
|
* |
|
10
|
|
|
* @package PHP-Wiki-API |
|
11
|
|
|
* @copyright Copyright (c) 2019 Igor Gaffling <[email protected]> |
|
12
|
|
|
* @license https://www.gnu.org/licenses/old-licenses/lgpl-2.1.txt LGPL License |
|
13
|
|
|
* @version Release: @1.1@ |
|
14
|
|
|
* @link https://github.com/gaffling/PHP-Wiki-API |
|
15
|
|
|
* @since Class available since Release 1.0 |
|
16
|
|
|
* |
|
17
|
|
|
* @example <php> |
|
18
|
|
|
* require_once __DIR__.'/wiki2api.php'; // Include the Wikipedia API Class |
|
19
|
|
|
* $wiki = new wiki(); // Start the Wikipedia API Class |
|
20
|
|
|
* echo $wiki->api($_GET['q']); // Output the API Response |
|
21
|
|
|
* </php> |
|
22
|
|
|
*/ |
|
23
|
|
|
|
|
24
|
|
|
|
|
25
|
|
|
class WikiToApi |
|
26
|
|
|
{ |
|
27
|
|
|
|
|
28
|
|
|
|
|
29
|
|
|
// Read and set Parameters |
|
30
|
|
|
public function __construct($params=array()) |
|
31
|
|
|
{ |
|
32
|
|
|
|
|
33
|
|
|
// Default Values |
|
34
|
|
|
$defaults = array( |
|
35
|
|
|
'language' => 'de', |
|
36
|
|
|
'userAgent' => 'WikiBot/1.0 (+http://'.$_SERVER['SERVER_NAME'].'/)', |
|
37
|
|
|
'betterResults' => true, |
|
38
|
|
|
'proxy' => '', |
|
39
|
|
|
'imageProxy' => true, |
|
40
|
|
|
'DEBUG' => '', |
|
41
|
|
|
); |
|
42
|
|
|
|
|
43
|
|
|
// Merge Parameters and Defaults |
|
44
|
|
|
$this->params = array_merge($defaults, $params); |
|
|
|
|
|
|
45
|
|
|
} |
|
46
|
|
|
|
|
47
|
|
|
// Helper Function to get the Content from the API URL |
|
48
|
|
|
private function getContent($url, $user_agent, $proxy='') |
|
49
|
|
|
{ |
|
50
|
|
|
|
|
51
|
|
|
// Hopfully we run PHP 4 >= 4.3.0 |
|
52
|
|
|
if (function_exists('file_get_contents')) { |
|
53
|
|
|
|
|
54
|
|
|
// Set User-Agent and Proxy |
|
55
|
|
|
$context = array ( |
|
56
|
|
|
'http' => array ( |
|
57
|
|
|
'user_agent' => $user_agent, |
|
58
|
|
|
'proxy' => $proxy, |
|
59
|
|
|
'request_fulluri' => true, |
|
60
|
|
|
), |
|
61
|
|
|
); |
|
62
|
|
|
|
|
63
|
|
|
// Build Stream Context |
|
64
|
|
|
$context = stream_context_create($context); |
|
65
|
|
|
|
|
66
|
|
|
// Use file_get_contents() Function and hide Error with @ |
|
67
|
|
|
$content = @file_get_contents($url, null, $context); |
|
68
|
|
|
} |
|
69
|
|
|
else // We run PHP < 4.3.0 - OMG :-o |
|
70
|
|
|
{ |
|
71
|
|
|
|
|
72
|
|
|
// Ini Var |
|
73
|
|
|
$content = ''; |
|
74
|
|
|
|
|
75
|
|
|
// Open URL and hide Error with @ |
|
76
|
|
|
if($handle = @fopen($url, 'r')) { |
|
77
|
|
|
|
|
78
|
|
|
// While there is Data |
|
79
|
|
|
while (!feof($handle)) |
|
80
|
|
|
{ |
|
81
|
|
|
|
|
82
|
|
|
// Read the Data-Line |
|
83
|
|
|
$line = fgets($handle, 4096); |
|
84
|
|
|
|
|
85
|
|
|
// Add the Data-Line to the Content Var |
|
86
|
|
|
$content .= $line; |
|
87
|
|
|
} |
|
88
|
|
|
|
|
89
|
|
|
// Better Close the FileHandle after the fgets() |
|
90
|
|
|
fclose($handle); |
|
91
|
|
|
} |
|
92
|
|
|
} |
|
93
|
|
|
|
|
94
|
|
|
// The Function returns the Content |
|
95
|
|
|
return $content; |
|
96
|
|
|
} |
|
97
|
|
|
|
|
98
|
|
|
// Call the API Main Function |
|
99
|
|
|
public function api($query) |
|
100
|
|
|
{ |
|
101
|
|
|
|
|
102
|
|
|
// Ini Vars |
|
103
|
|
|
$text = $image = $description = ''; |
|
104
|
|
|
|
|
105
|
|
|
// Convert Query to Lowercase for Headline |
|
106
|
|
|
$strtolower = mb_strtolower($query); |
|
107
|
|
|
|
|
108
|
|
|
// Convert Headlie to UTF-8 Uppercase Words |
|
109
|
|
|
$headline = mb_convert_case($strtolower, MB_CASE_TITLE, 'UTF-8'); |
|
110
|
|
|
|
|
111
|
|
|
// If Query is complete Uppercase make also complete Uppercase Headline |
|
112
|
|
|
if ($query === strtoupper($query)) { |
|
113
|
|
|
$headline = mb_strtoupper($query); |
|
114
|
|
|
} |
|
115
|
|
|
|
|
116
|
|
|
// Replace spaces in Query to Underscore and use Uppercase Words from Headline |
|
117
|
|
|
$query = str_replace(' ', '_', $headline); |
|
118
|
|
|
|
|
119
|
|
|
// In DEBUG Mode print Query |
|
120
|
|
View Code Duplication |
if ($this->params['DEBUG']=='KEY' || $this->params['DEBUG']=='ALL') { |
|
|
|
|
|
|
121
|
|
|
echo '<tt><b>Search-Keyword </b><xmp>#'.$query.'#</xmp></tt>'; |
|
122
|
|
|
} |
|
123
|
|
|
|
|
124
|
|
|
// First search the API if betterResults==true |
|
125
|
|
|
if ($this->params['betterResults'] == true) { |
|
126
|
|
|
|
|
127
|
|
|
// Wikipedia API URL 1 - https://en.wikipedia.org/w/api.php |
|
128
|
|
|
$url = 'https://'.$this->params['language'].'.wikipedia.org/w/api.php'. |
|
129
|
|
|
'?action=query&format=json&list=search&srsearch=intitle:'.$query. |
|
130
|
|
|
'&maxlag=1'; /* stop if wiki server is busy */ |
|
131
|
|
|
|
|
132
|
|
|
// If API Call 1 could be reached |
|
133
|
|
|
if ($api = $this->getContent($url, $this->params['userAgent'], $this->params['proxy'])) { |
|
134
|
|
|
|
|
135
|
|
|
// Decode the 1 Response |
|
136
|
|
|
$data = json_decode($api, true); |
|
137
|
|
|
|
|
138
|
|
|
// In DEBUG Mode print 1 Response |
|
139
|
|
View Code Duplication |
if ($this->params['DEBUG']=='API1' || $this->params['DEBUG']=='ALL') { |
|
|
|
|
|
|
140
|
|
|
echo '<pre><b>Search API-Call (1) Response</b> '; |
|
141
|
|
|
echo var_dump($data); |
|
|
|
|
|
|
142
|
|
|
echo '</pre>'; |
|
143
|
|
|
} |
|
144
|
|
|
|
|
145
|
|
|
// If there is a search Result |
|
146
|
|
|
if (isset($data['query']['search'][0]['title'])) { |
|
147
|
|
|
|
|
148
|
|
|
// Set Headline |
|
149
|
|
|
$headline = $data['query']['search'][0]['title']; |
|
150
|
|
|
|
|
151
|
|
|
// Set the Query to the first Search Result (and replace Spaces with Underscores) |
|
152
|
|
|
$query = str_replace(' ', '_', $data['query']['search'][0]['title']); |
|
153
|
|
|
|
|
154
|
|
|
// In DEBUG Mode print Found Keyword |
|
155
|
|
View Code Duplication |
if ($this->params['DEBUG']=='KEY' || $this->params['DEBUG']=='ALL') { |
|
|
|
|
|
|
156
|
|
|
echo '<tt><b>Found Search-Keyword </b><xmp>#'.$query.'#</xmp></tt>'; |
|
157
|
|
|
} |
|
158
|
|
|
} |
|
159
|
|
|
|
|
160
|
|
|
// If Search Result is a 'Did you mean:' Hint |
|
161
|
|
|
if (isset($data['query']['searchinfo']['suggestion'])) { |
|
162
|
|
|
|
|
163
|
|
|
// Set Text Hints depending on selected Language |
|
164
|
|
|
if ($this->params['language'] == 'de') { |
|
165
|
|
|
$suggestionText = 'Meinten Sie: '; |
|
166
|
|
|
} |
|
167
|
|
|
else |
|
168
|
|
|
{ |
|
169
|
|
|
$suggestionText = 'Did you mean: '; |
|
170
|
|
|
} |
|
171
|
|
|
|
|
172
|
|
|
// Remove 'q=' Variable=Value Pair from Querystring |
|
173
|
|
|
$QUERY_STRING = preg_replace('/'.('q'?'(\&|)q(\=(.*?)((?=&(?!amp\;))|$)|(.*?)\b)':'(\?.*)').'/i', '', $_SERVER['QUERY_STRING']); |
|
174
|
|
|
|
|
175
|
|
|
// Delete 'intitle:' from Suggestion Keyword |
|
176
|
|
|
$suggestion = str_replace('intitle:', '', $data['query']['searchinfo']['suggestion']); |
|
177
|
|
|
|
|
178
|
|
|
// Make Suggestion UTF-8 Uppercase Words |
|
179
|
|
|
$suggestion = mb_convert_case($suggestion, MB_CASE_TITLE, 'UTF-8'); |
|
180
|
|
|
|
|
181
|
|
|
// Make HTML Link for Suggestion |
|
182
|
|
|
$description = $suggestionText.'<a href="?q='. |
|
183
|
|
|
str_replace(' ', '_', $suggestion).$QUERY_STRING.'">'.$suggestion.'</a>'; |
|
184
|
|
|
} |
|
185
|
|
|
} |
|
186
|
|
|
} |
|
187
|
|
|
|
|
188
|
|
|
// Wikipedia API URL 2 - https://en.wikipedia.org/w/api.php |
|
189
|
|
|
$url = 'https://'.$this->params['language']. |
|
190
|
|
|
'.wikipedia.org/api/rest_v1/page/summary/'.$query. |
|
191
|
|
|
'?maxlag=1'; /* stop if wiki server is busy */ |
|
192
|
|
|
|
|
193
|
|
|
// If API Call 2 could be reached |
|
194
|
|
|
if ($api = $this->getContent($url, $this->params['userAgent'], $this->params['proxy'])) { |
|
195
|
|
|
// Decode the 2 Response |
|
196
|
|
|
$data = json_decode($api, true); |
|
197
|
|
|
|
|
198
|
|
|
// In DEBUG Mode print 2 Response |
|
199
|
|
View Code Duplication |
if ($this->params['DEBUG']=='API2' || $this->params['DEBUG']=='ALL') { |
|
|
|
|
|
|
200
|
|
|
echo '<pre><b>Main API-Call (2) Response</b> '; |
|
201
|
|
|
echo var_dump($data); |
|
202
|
|
|
echo '</pre>'; |
|
203
|
|
|
} |
|
204
|
|
|
|
|
205
|
|
|
// If there is an Image in the Search Result |
|
206
|
|
|
if (isset($data['originalimage']['source'])) { |
|
207
|
|
|
|
|
208
|
|
|
// If the DSGVO imageProxy should be use define it |
|
209
|
|
|
$proxy = ''; |
|
210
|
|
|
if ($this->params['imageProxy']==true) { |
|
211
|
|
|
$proxy = 'wiki-image-proxy.php?url='; |
|
212
|
|
|
} |
|
213
|
|
|
|
|
214
|
|
|
// Build HTML for Image |
|
215
|
|
|
$image = '<img src="'.$proxy.$data['thumbnail']['source'].'" />'; |
|
216
|
|
|
} |
|
217
|
|
|
|
|
218
|
|
|
// Correct the Text |
|
219
|
|
|
$text = str_replace('#', ': ', $data['extract_html']); |
|
220
|
|
|
|
|
221
|
|
|
// If there is a Description |
|
222
|
|
|
if (isset($data['description'])) { |
|
223
|
|
|
|
|
224
|
|
|
// Correct the Description depending on selected Language |
|
225
|
|
|
$description = str_replace( |
|
226
|
|
|
array( |
|
227
|
|
|
'Wikimedia-Begriffsklärungsseite', |
|
228
|
|
|
'Disambiguation page providing links to topics that could be referred to by the same search term' |
|
229
|
|
|
), |
|
230
|
|
|
array( |
|
231
|
|
|
'kann sich auf Folgendes beziehen', |
|
232
|
|
|
'may refer to the following' |
|
233
|
|
|
), |
|
234
|
|
|
$data['description'] |
|
235
|
|
|
); |
|
236
|
|
|
|
|
237
|
|
|
// Set Keyword to UTF-8 Uppercase Words of Query |
|
238
|
|
|
$keyword = mb_convert_case($strtolower, MB_CASE_TITLE, 'UTF-8'); |
|
239
|
|
|
|
|
240
|
|
|
// Highlight the Query in the Text and Delete some Text |
|
241
|
|
|
$text = str_replace( |
|
242
|
|
|
array($keyword, ' may refer to', ' steht für:'), |
|
243
|
|
|
array('<b class="hint">'.$keyword.'</b>', '', ''), |
|
244
|
|
|
$text |
|
245
|
|
|
); |
|
246
|
|
|
} |
|
247
|
|
|
|
|
248
|
|
|
// If there is no Article Text set a Default depending on selected Language |
|
249
|
|
|
// e.g. q=Leonardo%20di%20caprio&language=de OR q=100&language=de |
|
250
|
|
|
if ($text == '') { |
|
251
|
|
|
$description = $image = ''; |
|
|
|
|
|
|
252
|
|
View Code Duplication |
if ($this->params['language'] == 'de') { |
|
|
|
|
|
|
253
|
|
|
$text = 'Zu diesem Stichwort ist kein Artikel vorhanden.'; |
|
|
|
|
|
|
254
|
|
|
} |
|
255
|
|
|
else if($text == '') { |
|
256
|
|
|
$text = 'There is no article available for this keyword.'; |
|
|
|
|
|
|
257
|
|
|
} |
|
258
|
|
|
return; // ONLY IF YOU WHANT NO OUTPUT !! |
|
259
|
|
|
} |
|
260
|
|
|
} |
|
261
|
|
|
|
|
262
|
|
|
// Build the HTML Output |
|
263
|
|
View Code Duplication |
if ($this->params['language']=='de') { |
|
|
|
|
|
|
264
|
|
|
$moreAbout = 'Mehr über'; |
|
265
|
|
|
$from = 'bei'; |
|
266
|
|
|
} |
|
267
|
|
|
else |
|
268
|
|
|
{ |
|
269
|
|
|
$moreAbout = 'More about'; |
|
270
|
|
|
$from = 'from'; |
|
271
|
|
|
} |
|
272
|
|
|
|
|
273
|
|
|
// Without any Search Result return nothing |
|
274
|
|
|
if ($text == '' && $description == '') { |
|
275
|
|
|
return ''; |
|
276
|
|
|
} |
|
277
|
|
|
|
|
278
|
|
|
// With a Search Resuld build a Footer Link |
|
279
|
|
|
if ($text != '') { |
|
280
|
|
|
$footer = $moreAbout.' »'.$headline.'« '.$from; |
|
281
|
|
|
$url = 'https://'.$this->params['language'].'.wikipedia.org/wiki/'.$query; |
|
282
|
|
|
} |
|
283
|
|
|
else if ($description != '') { |
|
284
|
|
|
// Footer Link for Suggestion-Link |
|
285
|
|
|
$footer = ''; |
|
286
|
|
|
$url = 'https://'.$this->params['language'].'.wikipedia.org/'; |
|
287
|
|
|
} |
|
288
|
|
|
|
|
289
|
|
|
// Use the Template |
|
290
|
|
|
ob_start(); |
|
291
|
|
|
include 'wiki2tpl.phtm'; |
|
292
|
|
|
|
|
293
|
|
|
// Return the HTML |
|
294
|
|
|
return ob_get_clean(); |
|
295
|
|
|
|
|
296
|
|
|
} |
|
297
|
|
|
|
|
298
|
|
|
} |
|
299
|
|
|
|
In PHP it is possible to write to properties without declaring them. For example, the following is perfectly valid PHP code:
Generally, it is a good practice to explictly declare properties to avoid accidental typos and provide IDE auto-completion: