téléchargement des images d'un article

3 files changed, 134 insertions, 16 deletions

diff --git a/inc/config.php b/inc/config.php
index d468228d..51a25d36 100644
--- a/inc/config.php
+++ b/inc/config.php
@@ -7,7 +7,10 @@
  * @copyright  2013
  * @license    http://www.wtfpl.net/ see COPYING file
  */
+
+if(!is_dir('db/')){mkdir('db/',0705);}
 define ('DB_PATH', 'sqlite:./db/poche.sqlite');
+define ('ABS_PATH', 'archiveImg/');
 
 include 'db.php';
 include 'functions.php';
@@ -44,4 +47,4 @@ $token              = (isset ($_REQUEST['token'])) ? $_REQUEST['token'] : '';
 
 if ($action != '') {
     action_to_do($action, $id, $url, $token);
-}
\ No newline at end of file
+}
diff --git a/inc/db.php b/inc/db.php
index 705fc0c4..a1674ea7 100644
--- a/inc/db.php
+++ b/inc/db.php
@@ -10,12 +10,15 @@
 
 class db {
     var $handle;
-    function __construct($path) {
+    function __construct($path) 
+    {
         $this->handle = new PDO($path);
+        $this->handle->exec('CREATE TABLE IF NOT EXISTS "entries" ("id" INTEGER PRIMARY KEY  AUTOINCREMENT  NOT NULL  UNIQUE , "title" VARCHAR, "url" VARCHAR UNIQUE , "is_read" INTEGER DEFAULT 0, "is_fav" INTEGER DEFAULT 0, "content" BLOB)');
         $this->handle->setAttribute(PDO::ATTR_ERRMODE, PDO::ERRMODE_EXCEPTION);
     }
 
-    public function getHandle() {
+    public function getHandle() 
+    {
         return $this->handle;
     }
 }
\ No newline at end of file
diff --git a/inc/functions.php b/inc/functions.php
index 02d2c9dc..37abafd4 100755
--- a/inc/functions.php
+++ b/inc/functions.php
@@ -16,8 +16,9 @@ function get_poche_url()
 }
 
 // function define to retrieve url content
-function get_external_file($url, $timeout)
+function get_external_file($url)
 {
+    $timeout = 15;
     // spoofing FireFox 18.0
     $useragent="Mozilla/5.0 (Windows NT 5.1; rv:18.0) Gecko/20100101 Firefox/18.0";
 
@@ -45,10 +46,7 @@ function get_external_file($url, $timeout)
     } else {
 
         // create http context and add timeout and user-agent
-        $context = stream_context_create(array('http'=>array('timeout' => $timeout, // Timeout : time until we stop waiting for the response.
-        'header'=> "User-Agent: ".$useragent, // spoot Mozilla Firefox
-        'follow_location' => true
-        )));
+        $context = stream_context_create(array('http'=>array('timeout' => $timeout,'header'=> "User-Agent: ".$useragent,/*spoot Mozilla Firefox*/'follow_location' => true)));
 
         // only download page lesser than 4MB
         $data = @file_get_contents($url, false, $context, -1, 4000000); // We download at most 4 MB from source.
@@ -90,7 +88,7 @@ function get_external_file($url, $timeout)
 /**
  * Préparation de l'URL avec récupération du contenu avant insertion en base
  */
-function prepare_url($url)
+function prepare_url($url, $id)
 {
     $parametres = array();
     $url    = html_entity_decode(trim($url));
@@ -112,19 +110,127 @@ function prepare_url($url)
         $r->convertLinksToFootnotes = TRUE;
         if($r->init())
         {
-            $title = $r->articleTitle->innerHTML;
+            $content = $r->articleContent->innerHTML;
+            $parametres['title'] = $r->articleTitle->innerHTML;
+            $parametres['content'] = filtre_picture($content, $url, $id);
+            return $parametres;
         }
     }
 
-    $parametres['title']    = $title;
-    $parametres['content']  = $r->articleContent->innerHTML;
+    logm('error during url preparation');
+    return FALSE;
+}
+
+/**
+ * On modifie les URLS des images dans le corps de l'article
+ */
+function filtre_picture($content, $url, $id)
+{
+    $matches = array();
+    preg_match_all('#<\s*(img)[^>]+src="([^"]*)"[^>]*>#Si', $content, $matches, PREG_SET_ORDER);
+    foreach($matches as $i => $link)
+    {
+        $link[1] = trim($link[1]);
+        if (!preg_match('#^(([a-z]+://)|(\#))#', $link[1]) )
+        {
+            $absolute_path = get_absolute_link($link[2],$url);
+            $filename = basename(parse_url($absolute_path, PHP_URL_PATH));
+            $directory = create_assets_directory($id);
+            $fullpath = $directory . '/' . $filename;
+            download_pictures($absolute_path, $fullpath);
+            $content = str_replace($matches[$i][2], $fullpath, $content);
+        }
+
+    }
+
+    return $content;
+}
+
+/**
+ * Retourne le lien absolu
+ */
+function get_absolute_link($relative_link, $url)
+{
+    /* return if already absolute URL */
+    if (parse_url($relative_link, PHP_URL_SCHEME) != '') return $relative_link;
+
+    /* queries and anchors */
+    if ($relative_link[0]=='#' || $relative_link[0]=='?') return $url . $relative_link;
+
+    /* parse base URL and convert to local variables:
+       $scheme, $host, $path */
+    extract(parse_url($url));
+
+    /* remove non-directory element from path */
+    $path = preg_replace('#/[^/]*$#', '', $path);
+
+    /* destroy path if relative url points to root */
+    if ($relative_link[0] == '/') $path = '';
+
+    /* dirty absolute URL */
+    $abs = $host . $path . '/' . $relative_link;
+
+    /* replace '//' or '/./' or '/foo/../' with '/' */
+    $re = array('#(/\.?/)#', '#/(?!\.\.)[^/]+/\.\./#');
+    for($n=1; $n>0; $abs=preg_replace($re, '/', $abs, -1, $n)) {}
+
+    /* absolute URL is ready! */
+    return $scheme.'://'.$abs;
+}
+
+/**
+ * Téléchargement des images
+ */
+
+function download_pictures($absolute_path, $fullpath)
+{
+    $rawdata = get_external_file($absolute_path);
 
-    return $parametres;
+    if(file_exists($fullpath)) {
+        unlink($fullpath);
+    }
+    $fp = fopen($fullpath, 'x');
+    fwrite($fp, $rawdata);
+    fclose($fp);
+}
+
+/**
+ * Crée un répertoire de médias pour l'article
+ */
+function create_assets_directory($id)
+{
+    $assets_path = ABS_PATH;
+    if(!is_dir($assets_path)) {
+        mkdir($assets_path, 0705);
+    }
+
+    $article_directory = $assets_path . $id;
+    if(!is_dir($article_directory)) {
+        mkdir($article_directory, 0705);
+    }
+
+    return $article_directory;
+}
+
+/**
+ * Suppression du répertoire d'images
+ */
+function remove_directory($directory)
+{
+    if(is_dir($directory)) {
+        $files = array_diff(scandir($directory), array('.','..'));
+        foreach ($files as $file) {
+            // FIXME c'est quoi delTree ??
+            (is_dir("$directory/$file")) ? delTree("$directory/$file") : unlink("$directory/$file");
+        }
+        return rmdir($directory);
+    }
 }
 
 /**
  * Appel d'une action (mark as fav, archive, delete)
  */
+
 function action_to_do($action, $id, $url, $token)
 {
     global $db;
@@ -135,13 +241,19 @@ function action_to_do($action, $id, $url, $token)
             if ($url == '')
                 continue;
 
-            $parametres_url = prepare_url($url);
-            $sql_action     = 'INSERT INTO entries ( url, title, content ) VALUES (?, ?, ?)';
-            $params_action  = array($url, $parametres_url['title'], $parametres_url['content']);
+            $req = $db->getHandle()->query("SELECT id FROM entries ORDER BY id DESC");
+            $id = $req->fetchColumn()+1;
+
+            if($parametres_url = prepare_url($url, $id)) {
+                $sql_action     = 'INSERT INTO entries ( id, url, title, content ) VALUES (?,?, ?, ?)';
+                $params_action  = array($id,$url, $parametres_url['title'], $parametres_url['content']);
+            }
+
             logm('add link ' . $url);
             break;
         case 'delete':
             if (verif_token($token)) {
+                remove_directory(ABS_PATH . $id);
                 $sql_action     = "DELETE FROM entries WHERE id=?";
                 $params_action  = array($id);
                 logm('delete link #' . $id);