How to import image using Feeds which doesn't have a correct file extension?

Question

I am using Drupal 7 and Feeds to import an image field into a content type article and it does work when the image URL is a direct link to the image.

However images that are linked to image handlers, like:

http://business.iafrica.com/apps/imageUtil/view/article/1008937/1/630x385/

are not importing.

I have tried:

• to use the Feeds Tamper module and Feeds Tamper PHP and alter the image name, but this resulted in an invalid URL loading,
• to use Field Image Grabber, but I have an XML source, not HTML.

Here is my source feed URL:

http://resource.thumbtribe.mobi/ds1.portal/portal/1/4/resource/view/maximized/9500187?format=atom_extern

and my Image: URI XPath mapping is:

string(link[@rel="enclosure"]/@href)

Answer 1

I managed to solve this by using the feeds tamper module with the following php snippet to save the image and then reference the path on my server.

$checkext = substr($field, -5);

if (strpos($checkext,'.') == false)
{
  $filename = str_replace(".","_",microtime(true)).".jpg";
  $output = "public://field/image/".$filename;
  file_put_contents($output, file_get_contents($field));
  return  $output;
}
else
{
  return $field;
}

Answer 2

The problem is that current version of Feeds gets the filename from the URL (last part separated by / without anything after ?) where the logic is basically (in FeedsParser.inc):

// Strip any query string or fragment from file name.
list($filename) = explode('?', $this->getValue());
list($filename) = explode('#', $filename);
$filename = rawurldecode(drupal_basename($filename));

so it doesn't detect the extension automatically.

So the solution is to patch Feeds by fetching the file name either from headers (Content Disposition, check this patch) or based on its mime type:

$filename = $this->getLocalValue();
$filemime = file_get_mimetype($filename);

if ($filemime != $this->mime_type) {
  include_once DRUPAL_ROOT . "/includes/file.mimetypes.inc";
  $mapping = file_mimetype_mapping();
  if ($ext_id = array_search($this->mime_type, $mapping["mimetypes"])) {
    $extensions = array_keys($mapping["extensions"], $ext_id);
    foreach ($extensions as $extension) {
      if (in_array($extension, explode(' ', $this->allowedExtensions), TRUE)) {
        $filename .= ".$extension";
        break;
      }
    }
  }
}

Please check the following Drupal issues which have appropriate patches fixing it:

• Allow passing custom filename to FeedsParser.inc::getFile()
• Feeds doesn't import remote images which have no proper extension

which should be included soon into Feeds.

---

Alternative solution could be to rename the file before it gets saved.