PHP 得到HTML中的文本信息
转载：
?
<?php
function SpHtml2Text($str)
{
?$str = preg_replace("/<sty(.*)\\/style>|<scr(.*)\\/script>|<!--(.*)-->/isU","",$str);
?$alltext = "";
?$start = 1;
?for($i=0;$i<strlen($str);$i++)
?{
? if($start==0 && $str[$i]==">")
? {
?? $start = 1;
? }
? else if($start==1)
? {
?? if($str[$i]=="<")
?? {
??? $start = 0;
??? $alltext .= " ";
?? }
?? else if(ord($str[$i])>31)
?? {
??? $alltext .= $str[$i];
?? }
? }
?}
?$alltext = str_replace(" "," ",$alltext);
?$alltext = preg_replace("/&([^;&]*)(;|&)/","",$alltext);
?$alltext = preg_replace("/[ ]+/s"," ",$alltext);
?return $alltext;
}
function Html2Text($str,$r=0)
{
?if($r==0)
?{
? return SpHtml2Text($str);
?}
?else
?{
? $str = SpHtml2Text(stripslashes($str));
? return addslashes($str);
?}
}
$html=file_get_contents('http://www.baidu.com');
echo Html2Text($html);
?>
