如何模仿堆栈溢出自动链接行为

Question

使用PHP，我如何模仿Stack Overflow的自动链接行为（BTW非常酷）？

例如，以下网址：

  

http://www.stackoverflow.com/questions/1925455/how-to-mimic-stackoverflow-auto-link-behavior

转换成这个：

<a title="how to mimic stackoverflow auto link behavior" rel="nofollow" href="http://www.stackoverflow.com/questions/1925455/how-to-mimic-stackoverflow-auto-link-behavior">stackoverflow.com/questions/1925455/…</a>

在这种情况下，我并不真正关心title属性。

---

而且：

  

http://pt.php.net/manual/en/function.base-convert.php#52450

转换成这个：

<a rel="nofollow" href="http://pt.php.net/manual/en/function.base-convert.php#52450">pt.php.net/manual/en/…</a>

如何在PHP中创建类似的功能？

PS：查看我对此问题的评论，了解更多示例和行为。

Answer 1

试一试。 URL匹配的正则表达式模式来自Daring Fireball。

/**
 * Replace links in text with html links
 *
 * @param  string $text
 * @return string
 */
function auto_link_text($text)
{
   // a more readably-formatted version of the pattern is on http://daringfireball.net/2010/07/improved_regex_for_matching_urls
   $pattern  = '(?i)\b((?:[a-z][\w-]+:(?:/{1,3}|[a-z0-9%])|www\d{0,3}[.]|[a-z0-9.\-]+[.][a-z]{2,4}/)(?:[^\s()<>]+|\(([^\s()<>]+|(\([^\s()<>]+\)))*\))+(?:\(([^\s()<>]+|(\([^\s()<>]+\)))*\)|[^\s`!()\[\]{};:\'".,<>?«»“”‘’]))';

   $callback = create_function('$matches', '
       $url       = array_shift($matches);
       $url_parts = parse_url($url);

       $text = parse_url($url, PHP_URL_HOST) . parse_url($url, PHP_URL_PATH);
       $text = preg_replace("/^www./", "", $text);

       $last = -(strlen(strrchr($text, "/"))) + 1;
       if ($last < 0) {
           $text = substr($text, 0, $last) . "&hellip;";
       }

       return sprintf(\'<a rel="nofollow" href="%s">%s</a>\', $url, $text);
   ');

   return preg_replace_callback($pattern, $callback, $text);
}

输入文字：

This is my text.  I wonder if you know about asking questions on StackOverflow:
 Check This out http://www.stackoverflow.com/questions/1925455/how-to-mimic-stackoverflow-auto-link-behavior

 Also, base_convert php function?
http://pt.php.net/manual/en/function.base-convert.php#52450

http://pt.php.net/manual/en/function.base-convert.php?wtf=hehe#52450

输出文字：

This is my text.  I wonder if you know about asking questions on StackOverflow:
 Check This out <a rel="nofollow" href="http://www.stackoverflow.com/questions/1925455/how-to-mimic-stackoverflow-auto-link-behavior">stackoverflow.com/questions/1925455/&hellip;</a>

 Also, base_convert php function?
<a rel="nofollow" href="http://pt.php.net/manual/en/function.base-convert.php#52450">pt.php.net/manual/en/&hellip;</a>

<a rel="nofollow" href="http://pt.php.net/manual/en/function.base-convert.php?wtf=hehe#52450">pt.php.net/manual/en/&hellip;</a>

Answer 2

这是基于相同的daringfireball.net regular expression，但增加了比Eric Coleman的示例更多的逻辑，以及最大URL深度的配置（SO似乎是50），URL被截断时的最大路径深度（所以似乎是2）和省略号（…）。

据我所知，这复制了所有SO URL重写功能，至少就目前在评论和回复中所讨论的内容而言。

function auto_link_text($text) {
    $pattern  = '#\b(([\w-]+://?|www[.])[^\s()<>]+(?:\([\w\d]+\)|([^[:punct:]\s]|/)))#';
    return preg_replace_callback($pattern, 'auto_link_text_callback', $text);
}

function auto_link_text_callback($matches) {
    $max_url_length = 50;
    $max_depth_if_over_length = 2;
    $ellipsis = '&hellip;';

    $url_full = $matches[0];
    $url_short = '';

    if (strlen($url_full) > $max_url_length) {
        $parts = parse_url($url_full);
        $url_short = $parts['scheme'] . '://' . preg_replace('/^www\./', '', $parts['host']) . '/';

        $path_components = explode('/', trim($parts['path'], '/'));
        foreach ($path_components as $dir) {
            $url_string_components[] = $dir . '/';
        }

        if (!empty($parts['query'])) {
            $url_string_components[] = '?' . $parts['query'];
        }

        if (!empty($parts['fragment'])) {
            $url_string_components[] = '#' . $parts['fragment'];
        }

        for ($k = 0; $k < count($url_string_components); $k++) {
            $curr_component = $url_string_components[$k];
            if ($k >= $max_depth_if_over_length || strlen($url_short) + strlen($curr_component) > $max_url_length) {
                if ($k == 0 && strlen($url_short) < $max_url_length) {
                    // Always show a portion of first directory
                    $url_short .= substr($curr_component, 0, $max_url_length - strlen($url_short));
                }
                $url_short .= $ellipsis;
                break;
            }
            $url_short .= $curr_component;
        }

    } else {
        $url_short = $url_full;
    }

    return "<a rel=\"nofollow\" href=\"$url_full\">$url_short</a>";
}

示例输入：

This is my text.  I wonder if you know about asking questions on StackOverflow:
Check This out http://www.stackoverflow.com/questions/1925455/how-to-mimic-stackoverflow-auto-link-behavior

Also, base_convert php function?
http://pt.php.net/manual/en/function.base-convert.php#52450

http://pt.php.net/manual/en/function.base-convert.php?wtf=hehe#52450

http://a.b/c/d/e/f/test

and http://a.b/c/d/e/f/g/h/i/j/k/l/m/n/o/p/q/r/s/t/u/v/z/y/w/z/test

示例输出：

This is my text.  I wonder if you know about asking questions on StackOverflow:
Check This out <a rel="nofollow" href="http://www.stackoverflow.com/questions/1925455/how-to-mimic-stackoverflow-auto-link-behavior">http://stackoverflow.com/questions/1925455/&hellip;</a> 

Also, base_convert php function?
<a rel="nofollow" href="http://pt.php.net/manual/en/function.base-convert.php#52450">http://pt.php.net/manual/en/&hellip;</a> 

<a rel="nofollow" href="http://pt.php.net/manual/en/function.base-convert.php?wtf=hehe#52450">http://pt.php.net/manual/en/&hellip;</a> 

<a rel="nofollow" href="http://a.b/c/d/e/f/test">http://a.b/c/d/e/f/test</a> 

and <a rel="nofollow" href="http://a.b/c/d/e/f/g/h/i/j/k/l/m/n/o/p/q/r/s/t/u/v/z/y/w/z/test">http://a.b/c/d/&hellip;</a>

Answer 3

这会将示例字符串转换为您所使用的字符串。我遗漏了title，因为它来自不同的来源，而不仅仅是一个独立的网址，你说这并不重要。

<?php
$urlInput="http://www.stackoverflow.com/questions/1925455/how-to-mimic-stackoverflow-auto-link-behavior";
preg_match('@http://(?:www\.)?(\S+/)\S*(?:\s|$)@i', $urlInput, $matches);
print('<a rel="nofollow" href="' . trim($matches[0]) . '">' . $matches[1] . '...</a>');
?>

根据需要进行扩展以扫描您的文字。

如果您只想匹配一定数量的URL路径元素，请使用此RE：

'@http://(?:www\.)?((?:\S+?/){1,3})\S*(?:\s|$)@i'

这将最多提取3个路径元素（主机和最多两个目录）。您可以改变{1,3}中的上限来定义所需的最大路径元素数。

更改结尾\S以允许零匹配。

Answer 4

如果您有一个像SO这样的可预测网址，那么应该很容易获取带有正则表达式的链接并过滤掉与该模式匹配的链接。因此，如果您的网址为http://example.com/stuff/1234，则使用正则表达式查找http://example.com/stuff/1234/how-to-mimic将非常简单。

<?php
preg_match('/http:\/\/example.com\/(\w*)\/(\d)[\/*]/', $text, $matches);

if (is_array($matches))
{
  foreach ($matches as $match)
  {
    // do something...
  }
}
?>

Answer 5

基于Kevin Brock的回答，但允许可配置的参数（文件夹深度和URL长度），并接受没有尾部斜杠的URL：

$url = 'http://www.stackoverflow.com/questions/1925455/how-to-mimic-stackoverflow-auto-link-behavior';
$output = '';
$params = array (
    'length' => 10,
    'depth' => 2,
);
preg_match ('@http://(?:www\.)?([^/?# ]+)(/\S+)?(?=\s|$)@i', $url, $matches);
if (isset ($matches[2]))
{
    $parts = explode('/', substr($matches[2], 1));
    if (count($parts) > $params['depth'] && strlen($matches[1].$matches[2]) > $params['length'])
        $output = $matches[1].'/'.implode('/', array_slice($parts, 0, 2)).'/...';
    else
        $output = $matches[1].$matches[2];
}
else
    $output = $matches[1];

echo '<a href="'.$matches[0].'">'.$output.'</a>';

希望这有帮助

Answer 6

请参阅Regex (regular expression) to match a URL：

https?://([-\w\.]+)+(:\d+)?(/([\w/_\.]*(\?\S+)?)?)?

     

PHP示例：自动链接URL的内部文本。

     

$text = preg_replace('@(https?://([-\w\.]+)+(:\d+)?(/([\w/_\.]*(\?\S+)?)?)?)@', '<a href="$1">$1</a>', $text);