Saving image via API and Filename Error

[qubism]

Hey Girls & Boys,

I'm currently importing data with some basic scraping from another page and adding new pages with the ProcessWire API.
The data is mostly text which works fine but some images can't get downloaded from the other page.

My guess would be due to the filename.

Here's the Error:

ProcessWire\WireException

File could not be downloaded (https://************.com/content/v1/5e8119f5232766b98/1613689-FJ73MYGC/Außenansicht+2+Galerie.jpg) 400 Bad Request:  (tried: curl) search

My Code snippet (throws the error also without sanitizer):

if($html->find('img.thumb-image', 0)->{'data-src'}) {
    $image = $html->find('img.thumb-image', 0)->{'data-src'};
    $p->article_thumb = $sanitizer->url($image);
}

Is there a way to do this?

 

Thanks for your time ?

 

Edit: Found a solution. urlencode changed the whole URL and made the API upload empty images. So I kept the URL until the last slash and just changed the filename like so:

        $image = $html->find('img.thumb-image', 0)->{'data-src'};

        $imageURL = $image;
        $pos = strrpos($imageURL, '/') + 1;
        $result = substr($imageURL, 0, $pos) . urlencode(substr($imageURL, $pos));

 

[Robin S]

Try using urlencode() on the image URL to deal with spaces and other potentially problematic characters.

[JayGee]

Also check the source server doesn't hasn't blocked image hotlinking or downloading of images without a referrer. I had an issue with an import script once along these lines.

[qubism]

7 hours ago, Robin S said:

Try using urlencode() on the image URL to deal with spaces and other potentially problematic characters.

I tried that and also rawurlencode() and the API just uploads an blank image with the whole absolute url as filename.

7 hours ago, Guy Incognito said:

Also check the source server doesn't hasn't blocked image hotlinking or downloading of images without a referrer. I had an issue with an import script once along these lines.

Most images work though, so I guess that's not the problem. 

 

Another option would be to check for special characters in the filename and skip that images, so ProcessWire at least doesn't throw an error.
Not optimal, but still something

[horst]

When using urlencode, check if the + signs are changed too! Otherwise use a str_replace("+", "%20", $URL) additionally!!