Filter HTML tags using Java Regular Expressions

Import  Java. util. RegEx. matcher;  Import  Java. util. RegEx. pattern;  /**  * <P> * Title: HTML-related regular expression tool * </P> * <p> * Description: includes HTML Tag filtering and HTML Tag conversion, replace a specific HTML Tag * </P> * <p> * copyright: Copyright (c) 2006 * </P> **  @ Author  Hejian *  @ Version  1.0 * @ createtime 2006-10-16  */      Public  Class  Htmlregexputil {  Private   Final   Static String regxpforhtml = "<([^>] *)> "; //  Filter all labels starting with <and ending with>        Private   Final   Static String regxpforimgtag = "<\ s * IMG \ s + ([^>] *) \ s *> "; //  Find the IMG tag        Private   Final  Static String regxpforimatagsrcattrib = "src = \" ([^ \ "] + )\""; //  Locate the src attribute of the IMG tag        /**  *  */        Public  Htmlregexputil (){  //  Todo auto-generated constructor stub  }  /**  ** Basic function: Replace the tag for normal display ** <p> **  @ Param Input *  @ Return  String  */        Public  String replacetag (string input ){  If (! Hasspecialchars (input )){  Return  Input;} stringbuffer filtered = New  Stringbuffer (input. Length ());  Char  C; For ( Int I = 0; I <= input. Length ()-1; I ++ ) {C = Input. charat (I );  Switch  (C ){  Case '<' : Filtered. append ( "& Lt ;" );  Break  ;  Case '>': Filtered. append ( "& Gt ;" );  Break  ;  Case '"' : Filtered. append ( "& Quot ;" );  Break  ;  Case '&' : Filtered. append ( "& Amp ;" ); Break  ;  Default  : Filtered. append (c );}}  Return  (Filtered. tostring ());}  /**  ** Basic function: determines whether a tag exists. ** <p> **  @ Param  Input *  @ Return  Boolean  */        Public  Boolean  Hasspecialchars (string input ){  Boolean Flag = False  ;  If (Input! = Null ) & (Input. Length ()> 0 )){  Char  C;  For ( Int I = 0; I <= input. Length ()-1; I ++ ) {C =Input. charat (I );  Switch  (C ){  Case '>' : Flag = True  ;  Break  ;  Case '<' : Flag = True  ; Break  ;  Case '"' : Flag = True  ;  Break  ;  Case '&' : Flag = True  ;  Break ;}}}  Return  Flag ;}  /**  ** Basic function: filter all tags starting with "<" and ending with ">" * <p> **  @ Param  Str *  @ Return  String  */        Public   Static  String filterhtml (string Str) {pattern Pattern =Pattern. Compile (regxpforhtml); matcher = Pattern. matcher (STR); stringbuffer sb = New  Stringbuffer ();  Boolean Result1 = Matcher. Find ();  While  (Result1) {matcher. appendreplacement (SB, "" ); Result1 = Matcher. Find ();} matcher. appendtail (SB );  Return SB. tostring ();}  /**  ** Basic function: Filter specified tags ** <p> **  @ Param  Str *  @ Param  Tag * specify tag *  @ Return  String  */        Public   Static  String fiterhtmltag (string STR, string tag) {string regxp = "<\ S *" + tag + "\ s + ([^>] *) \ s *>"; Pattern Pattern = Pattern. Compile (regxp); matcher = Pattern. matcher (STR); stringbuffer sb = New  Stringbuffer ();  Boolean Result1 = Matcher. Find ();  While  (Result1) {matcher. appendreplacement (SB, "" ); Result1 =Matcher. Find ();} matcher. appendtail (SB );  Return  SB. tostring ();}  /**  ** Basic function: Replace the specified tag ** <p> **  @ Param  Str *  @ Param  Beforetag * tag to be replaced *  @ Param  Tagattrib * tag attribute value to be replaced *  @ Param Starttag * new tag start tag *  @ Param  Endtag * new tag end tag *  @ Return  String * @ for example, if the src attribute value of the IMG label is replaced by the [img] attribute value [/img]  */        Public   Static  String replacehtmltag (string STR, string beforetag, string tagattrib, string starttag, string endtag) {string regxpfortag = "<\ S *" + beforetag + "\ s + ([^>] *) \ s *>" ; String regxpfortagattrib = Tagattrib + "= \" ([^ \ "] + )\""; Pattern patternfortag = Pattern. Compile (regxpfortag); pattern patternforattrib = Pattern. Compile (regxpfortagattrib); matcher matcherfortag = Patternfortag. matcher (STR); stringbuffer sb = New  Stringbuffer ();  Boolean Result = Matcherfortag. Find ();  While  (Result) {stringbuffer sbreplace =New  Stringbuffer (); matcher matcherforattrib = Patternforattrib. matcher (matcherfortag. Group ( 1 ));  If  (Matcherforattrib. Find () {matcherforattrib. appendreplacement (sbreplace, starttag + Matcherforattrib. Group (1) + Endtag);} matcherfortag. appendreplacement (SB, sbreplace. tostring (); Result = Matcherfortag. Find ();} matcherfortag. appendtail (SB ); Return  SB. tostring ();}} 

 

Reprinted: http://aguang520.iteye.com/blog/1056686