Code to automatically generate article abstracts [PHP version]

Code to automatically generate article abstracts [PHP version]_PHP tutorial

WBOY

Release： 2016-07-21 15:55:31

Original

940 people have browsed it

Implementation content: Truncate a piece of text containing HTML code, but there will be no problem that the containment tag is not closed.
Note: This is the PHP version, for use on the server side. If you need a client version, please read the next article
When we write programs like BLOG, we often need to display the previous part of the article. , but I am afraid that improper truncation will destroy the closing tag and cause damage to the entire document structure. Using my function can solve this problem without high requirements.
Everyone should consider whether this function is applied on the server side or on the client side. Because I consider that this function may be more machine-intensive to run, it can be placed on the client if security requirements are not high.
It is best to put this summary in a separate field in the data table, so that the corresponding database query is also optimized. It's still a good deal to sacrifice a little space in exchange for a lot of time.
Let’s talk about security issues again, mainly content security. If the client intends to change the normal summary information, generally the owner of the BLOG has this power, and it is his own business to destroy the consistency between the summary and the original text. Security other than content can be solved on the server side. Therefore, it is recommended to use this function on the client side.
Core code

Copy code The code is as follows:

// PHP 4.3 or above needed  
define("BRIEF_LENGTH", 800);     //Word amount of the Briefing of an Article  
function Generate_Brief($text){  
  global $Briefing_Length;  
  if(strlen($text) <= BRIEF_LENGTH ) return $text;     
  $Foremost = substr($text, 0, BRIEF_LENGTH);  
  $re = "/<(/?)(P|DIV|H1|H2|H3|H4|H5|H6|ADDRESS|PRE|TABLE|TR|TD|TH|INPUT|SELECT|TEXTAREA|OBJECT|A|UL|OL|LI|BASE|META|LINK|HR|BR|PARAM|IMG|AREA|INPUT|SPAN)[^>]*(>?)/i";  
  $Single = "/BASE|META|LINK|HR|BR|PARAM|IMG|AREA|INPUT/i";     
  $Stack = array(); $posStack = array();  
  preg_match_all($re,$Foremost,$matches, PREG_SET_ORDER | PREG_OFFSET_CAPTURE);  
  /*   [Child-matching Specification]:  
    $matches[$i][1] : A "/" charactor indicating whether current "<...>" Friction is Closing Part  
    $matches[$i][2] : Element Name.  
    $matches[$i][3] : Right > of a "<...>" Friction   */  
  for($i = 0 ; $i < count($matches); $i++){  
    if($matches[$i][1][0] == ""){  
        $Elem = $matches[$i][2][0];  
        if(preg_match($Single,$Elem) && $matches[$i][3][0] !=""){  
          continue;  
        }  
        array_push($Stack, strtoupper($matches[$i][2][0]));  
        array_push($posStack, $matches[$i][2][1]);           
        if($matches[$i][3][0] =="") break;  
    }else{  
        $StackTop = $Stack[count($Stack)-1];  
        $End = strtoupper($matches[$i][2][0]);  
        if(strcasecmp($StackTop,$End)==0){  
          array_pop($Stack);  
          array_pop($posStack);  
          if($matches[$i][3][0] ==""){  
            $Foremost = $Foremost.">";  
          }  
        }  
    }       
  }  
  $cutpos = array_shift($posStack) - 1;     
  $Foremost = substr($Foremost,0,$cutpos);  
  return $Foremost;  
}; 

若遇到问题（发现上面的函数对多字节字符集支持得不好）不烦试试下面的这个！

复制代码代码如下:

function Generate_Brief($text){  
  global $Briefing_Length;  
  mb_regex_encoding("UTF-8");  
  if(mb_strlen($text) <= BRIEF_LENGTH ) return $text;     
  $Foremost = mb_substr($text, 0, BRIEF_LENGTH);  
  $re = "<(/?)(P|DIV|H1|H2|H3|H4|H5|H6|ADDRESS|PRE|TABLE|TR|TD|TH|INPUT|SELECT|TEXTAREA|OBJECT|A|UL|OL|LI|BASE|META|LINK|HR|BR|PARAM|IMG|AREA|INPUT|SPAN)[^>]*(>?)";  
  $Single = "/BASE|META|LINK|HR|BR|PARAM|IMG|AREA|INPUT|BR/i";     
  $Stack = array(); $posStack = array();  
  mb_ereg_search_init($Foremost, $re, 'i');  
  while($pos = mb_ereg_search_pos()){  
    $match = mb_ereg_search_getregs();  
    /*   [Child-matching Formulation]:  
        $matche[1] : A "/" charactor indicating whether current "<...>" Friction is Closing Part  
        $matche[2] : Element Name.  
        $matche[3] : Right > of a "<...>" Friction     
    */  
    if($match[1]==""){  
        $Elem = $match[2];  
        if(mb_eregi($Single, $Elem) && $match[3] !=""){  
          continue;  
        }  
        array_push($Stack, mb_strtoupper($Elem));  
        array_push($posStack, $pos[0]);           
    }else{  
        $StackTop = $Stack[count($Stack)-1];  
        $End = mb_strtoupper($match[2]);  
        if(strcasecmp($StackTop,$End)==0){  
          array_pop($Stack);  
          array_pop($posStack);  
          if($match[3] ==""){  
            $Foremost = $Foremost.">";  
          }  
        }  
    }  
  }  
  $cutpos = array_shift($posStack) - 1;     
  $Foremost = mb_substr($Foremost,0,$cutpos,"UTF-8");  
  return $Foremost;  
}; 