Page MenuHomeDevCentral
Files F11708811

No OneTemporary
Actions

View Options

diff --git a/pages/lalibrebelgique.php b/pages/lalibrebelgique.php
new file mode 100644
index 0000000..25d15a7
--- /dev/null
+++ b/pages/lalibrebelgique.php
@@ -0,0 +1,62 @@
+<?php
+
+//Page analysis for www.lesoir.be
+class LaLibreBelgiquePage extends Page {
+ function analyse () {
+ //La Libre uses ISO-8859-1 and not UTF-8
+ $this->data = iconv('iso-8859-1', 'utf-8', $this->data);
+
+ //Cals parent analyzer
+ parent::analyse();
+
+ //Hardcoded known info
+ $this->site = "La Libre Belgique";
+ $this->skipYMD = true;
+
+ //Gets date
+ //meta tag 'archi_id' has t-YYYYMMDD-HHMMhh as format (where hh = AM/PM)
+ // e.g. t-20120722-0211PM
+ $date = trim(self::between('Mis en ligne le ', '</p>'));
+ $yyyy = substr($date, 6, 4);
+ $mm = substr($date, 3, 2);
+ $dd = substr($date, 0, 2);
+ $this->date = strftime(LONG_DATE_FORMAT, mktime(12, 0, 0, $mm, $dd, $yyyy));
+
+ //Gets authors
+ $authors = trim(self::between('<p id="writer">', '</p>'));
+ if (strpos($authors, 'daction ') > 0) {
+ //"rédaction en ligne", "Rédaction web","Rédaction en ligne (avec afp)", etc.
+ //(they're not coherent about case).
+ $this->skipAuthor = true;
+ } else {
+ $authors = preg_split('/( et |, )/', $authors);
+ $start = true;
+ foreach ($authors as $author) {
+ //Fixes some authors
+ switch ($author) {
+ case 'G. Dt': $author = 'Guy Duplat'; break;
+ case 'afp': $author = 'AFP'; break;
+ }
+ if ($author == 'G. Dt') $author = 'Guy Duplat';
+ if ($start) {
+ $this->author = $author;
+ $start = false;
+ } else {
+ $this->coauthors[] = $author;
+ }
+ }
+ }
+
+ //Gets title
+ if (!$this->title = $this->meta_tags['og:title']) {
+ $this->title = self::between("<title>Lalibre.be - ", "</title>");
+ }
+ }
+
+ function get_meta_tags () {
+ //This site doesn't always use <meta name="" value=""> but sometimes property= or itemprop=
+ return $this->get_all_meta_tags();
+ }
+}
+
+?>

File Metadata

Mime Type
text/x-diff
Expires
Mon, Sep 15, 07:22 (1 d, 9 h)
Storage Engine
blob
Storage Format
Raw Data
Storage Handle
2983856
Default Alt Text
(2 KB)

Event Timeline

Nasqueron DevCentral · If it had been much bigger the moon would have had a core of ice. · Powered by Phabricator