Broken image paths

[jwarzech]

When extracting an article's html I am getting a lot of broken image links if the source article uses relative image paths (common on a lot of blogs).

Is there a way we could get relative paths to get the root domain prepended to them during the extraction?

If I get a chance I'll try to dive in and do a Pull-Request for this functionality but If someone has already done this or able to get to it quicker I'm sure everyone would find this very useful!

Thanks!

[codelucas]

I've also seen this bug, it's pretty annoying and a PR would be much appreciated!

[thoraxe]

Just hit this myself. I would have to dig in to see if I could even figure out how to do a PR.

But basically I'd think you would want to:

• detect that the image path is relative
• if relative, deconstruct the article's URL originally supplied to get the FQDN
• construct the full path by adding the FQDN and the relative path

[thoraxe]

I noticed that this also affects relative URLs.

[mercuree]

This is how I did it with links and images

def make_links_absoulte(article):

    if len(article.article_html) == 0:
        return ''
    top_node = article.top_node_article_html
    urlformat = urlparse(article.url)
    base = article.doc.base
    url_without_path = urlformat.scheme + "://" + urlformat.netloc
    if base:
        if urlparse(base).netloc:
            output_base = base
        else:
            output_base = url_without_path + base
    else:
        output_base = url_without_path + urlformat.path

    top_node.make_links_absolute(output_base)
    return tostring(top_node, encoding='unicode')