Mercurial > hg > ywww
comparison goodreads/goodreadsReviews.php @ 6:077b0a0a3e6d
remaining originals according to dependency walk
author | Robert Boland <robert@markup.co.uk> |
---|---|
date | Thu, 16 Feb 2017 22:29:02 +0000 |
parents | |
children |
comparison
equal
deleted
inserted
replaced
5:55445b456ad0 | 6:077b0a0a3e6d |
---|---|
1 <?php | |
2 include_once('simple_html_dom.php'); | |
3 | |
4 $isbn=$_GET['isbn']; | |
5 | |
6 function scraping_goodreads($url) { | |
7 // create HTML DOM | |
8 $html = file_get_html($url); | |
9 | |
10 $loop = 0; | |
11 $ret = "<?xml version=\"1.0\"?>"; | |
12 $ret .= "<Reviews>"; | |
13 foreach($html->find('div[class="gr_review_text"]') as $review) { | |
14 $rating = $html->find('<span[class="gr_rating"]',$loop); | |
15 $link = $html->find('<a[class="gr_more_link"]',$loop); | |
16 $ret .= "<Review>"; | |
17 $ret .= "<Rating>" . substr( $rating->plaintext, 0, 1) . "</Rating>"; | |
18 $ret .= "<Body>" . trim(substr(trim($review->plaintext),0,-7)) . "</Body>"; | |
19 $ret .= "<URL><![CDATA[" . $link . "]]></URL>"; | |
20 $ret .= "</Review>"; | |
21 $loop++; | |
22 } | |
23 | |
24 $ret .= "</Reviews>"; | |
25 $html->clear(); | |
26 unset($html); | |
27 $ret = preg_replace("/[^[:alnum:][:punct:][:space:]]/","",$ret); | |
28 echo convert_utf8($ret); | |
29 } | |
30 | |
31 function convert_utf8($str){ | |
32 if(!seems_utf8($str)) | |
33 return mb_convert_encoding($str, 'UTF-8'); | |
34 return $str; | |
35 } | |
36 | |
37 function seems_utf8($Str) { | |
38 for ($i=0; $i<strlen($Str); $i++) { | |
39 if (ord($Str[$i]) < 0x80) continue; # 0bbbbbbb | |
40 elseif ((ord($Str[$i]) & 0xE0) == 0xC0) $n=1; # 110bbbbb | |
41 elseif ((ord($Str[$i]) & 0xF0) == 0xE0) $n=2; # 1110bbbb | |
42 elseif ((ord($Str[$i]) & 0xF8) == 0xF0) $n=3; # 11110bbb | |
43 elseif ((ord($Str[$i]) & 0xFC) == 0xF8) $n=4; # 111110bb | |
44 elseif ((ord($Str[$i]) & 0xFE) == 0xFC) $n=5; # 1111110b | |
45 else return false; # Does not match any model | |
46 for ($j=0; $j<$n; $j++) { # n bytes matching 10bbbbbb follow ? | |
47 if ((++$i == strlen($Str)) || ((ord($Str[$i]) & 0xC0) != 0x80)) | |
48 return false; | |
49 } | |
50 } | |
51 return true; | |
52 } | |
53 | |
54 $ret = scraping_goodreads('http://www.goodreads.com/api/reviews_widget_iframe?did=8vxmKmAn8AVolobtvBbMIg&isbn=' . $isbn); | |
55 ?> |