1
0

Sexactu.php 3.5 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687
  1. <?php
  2. /**
  3. * Unfortunatly, Sexactu do not provide a RSS stream, which prevents me from upgrading this to the mugnificent RssExpander framework
  4. * @name Sexactu
  5. * @description Sexactu via rss-bridge
  6. * @update 04/02/2014
  7. */
  8. define("GQ", "http://www.gqmagazine.fr");
  9. class Sexactu extends BridgeAbstract{
  10. public function collectData(array $param){
  11. $find = array('janvier', 'février', 'mars', 'avril', 'mai', 'juin', 'juillet', 'août', 'septembre', 'novembre', 'décembre');
  12. $replace = array('January', 'February', 'March', 'April', 'May', 'June', 'July', 'August', 'September', 'October', 'November', 'December');
  13. $html = file_get_html($this->getURI()) or $this->returnError('Could not request '.$this->getURI(), 404);
  14. foreach($html->find('.content-holder') as $contentHolder) {
  15. // only use first list as second one only contains pages numbers
  16. $articles = $contentHolder->find('ul', 0);
  17. foreach($articles->find('li') as $element) {
  18. // if you ask about that method_exists, there seems to be a bug in simple html dom
  19. // see stackoverflow for more details : http://stackoverflow.com/a/10828479/15619
  20. if(is_object($element)) {
  21. $item = new Item();
  22. // various metadata
  23. $titleBlock = $element->find('.title-holder', 0);
  24. if(is_object($titleBlock)) {
  25. $titleDetails = $titleBlock->find('.article-title',0);
  26. $titleData = $titleDetails->find('h2', 0)->find('a',0);
  27. $titleTimestamp =$titleDetails->find('h4',0);
  28. $item->title = $this->correctCase(trim($titleData->innertext));
  29. $item->uri = GQ.$titleData->href;
  30. // Fugly date parsing due to the fact my DNS-323 doesn't support php intl extension
  31. $dateText = $titleTimestamp->innertext;
  32. $dateText = substr($dateText, strpos($dateText,',')+1);
  33. $dateText = str_replace($find, $replace, strtolower($dateText));
  34. $date = strtotime($dateText);
  35. $item->timestamp = $date;
  36. $item->name = "Maïa Mazaurette";
  37. $elementText = $element->find('.text-container', 0);
  38. // don't forget to replace images server url with gq one
  39. foreach($elementText->find('img') as $image) {
  40. $image->src = GQ.$image->src;
  41. }
  42. $item->content = $elementText->innertext;
  43. $this->items[] = $item;
  44. }
  45. }
  46. }
  47. }
  48. }
  49. public function getName(){
  50. return 'Sexactu';
  51. }
  52. public function getURI(){
  53. return GQ.'/sexactu';
  54. }
  55. public function getCacheDuration(){
  56. return 7200; // 2h hours
  57. }
  58. public function getDescription(){
  59. return "Sexactu via rss-bridge";
  60. }
  61. public function correctCase($str) {
  62. $sentences=explode('.', mb_strtolower($str, "UTF-8"));
  63. $str="";
  64. $sep="";
  65. foreach ($sentences as $sentence)
  66. {
  67. //upper case first char
  68. $sentence=ucfirst(trim($sentence));
  69. //append sentence to output
  70. $str=$str.$sep.$sentence;
  71. $sep=". ";
  72. }
  73. return $str;
  74. }
  75. }