Combining CURL and simple html dom

Question

I have been working with CURL to scrape websites for a while and also Simple HTML DOM. I experienced that CURL is much better for scraping websites. However I really like the simplicity of Simple HTML DOM. So I figured why not combine the two, I tried:

require_once('simple_html_dom.php');

    $url = 'http://news.yahoo.com/';

    $ch = curl_init($url);
    curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
    $curl_scraped_page = curl_exec($ch);

    $html = new simple_html_dom();
    $html->load($curl_scraped_page);


    foreach($html->find('head') as $d) {
        $d->innertext = "<base href='$url'>" . $d->innertext;
    }

    echo $html->save();

I did my best but it doesn't work. What else can I try?

Answer 1

Try changing this:

$html->load($curl_scraped_page);

To this:

$html->load($curl_scraped_page, true, false);

The problem is that simple_html_dom removes all \\r \\n by default and in this case it breaks javascript code since yahoo don't end it with a semicolon.

You can see this error at the browser console and you can also see that simple_html_dom removes linebreaks viewing the source.

Answer 2

I think I would add a function to the class

function loadWithoutRemovingStuff($str, $lowercase=true, $stripRN=true, $defaultBRText=DEFAULT_BR_TEXT, $defaultSpanText=DEFAULT_SPAN_TEXT)
{
    $this->prepare($str, $lowercase, $stripRN, $defaultBRText, $defaultSpanText);

    while ($this->parse());

    $this->root->_[HDOM_INFO_END] = $this->cursor;
    $this->parse_charset();

    return $this;
}

and then call that function instead of the default load function.

Or, since everything is public in this class,

 $html = new simple_html_dom();
 $html->prepare($str, $lowercase, $stripRN, $defaultBRText, $defaultSpanText);

 while ($html->parse());
 $html->root->_[HDOM_INFO_END] = $html->cursor;
 $html->parse_charset();

but the first way is better (cleaner)