Java Tidy.setDropEmptyParas方法代碼示例

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
public JTidyBookProcessor()
    {
        tidy = new Tidy();
//        tidy.setConfigurationFromFile(JTidyBookProcessor.class.getResource("/jtidy.properties").getFile());
        tidy.setSpaces(2);
        tidy.setIndentContent(true);
        tidy.setSmartIndent(true);
        tidy.setXHTML(true);
        tidy.setQuoteMarks(false);
        tidy.setQuoteAmpersand(true);
        tidy.setDropEmptyParas(false);
        tidy.setTidyMark(false);
        tidy.setJoinClasses(true);
        tidy.setJoinStyles(true);
        tidy.setWraplen(0);
        tidy.setDropProprietaryAttributes(true);
        tidy.setEscapeCdata(true);
        Properties props = new Properties();
        props.put("new-blocklevel-tags", "svg image  altGlyph altGlyphDef altGlyphItem animate animateColor animateMotion animateTransform circle clipPath color-profile cursor defs desc ellipse feBlend feColorMatrix feComponentTransfer feComposite feConvolveMatrix feDiffuseLighting feDisplacementMap feDistantLight feFlood feFuncA feFuncB feFuncG feFuncR feGaussianBlur feImage feMerge feMergeNode feMorphology feOffset fePointLight feSpecularLighting feSpotLight feTile feTurbulence filter font font-face font-face-format font-face-name font-face-src font-face-uri foreignObject g glyph glyphRef hkern image line linearGradient marker mask metadata missing-glyph mpath path pattern polygon polyline radialGradient rect script set stop style svg switch symbol text textPath title tref tspan use view vkern");
        tidy.getConfiguration().addProps(props);
    }
 

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
/**
 * Initializes the tidy document builder.
 */
private void initializeTidyBuilder() {
    tidyBuilder = new Tidy();
    tidyBuilder.setInputEncoding("UTF-8");
    tidyBuilder.setOutputEncoding("UTF-8");
    tidyBuilder.setXmlOut(true);
    tidyBuilder.setShowWarnings(false);
    tidyBuilder.setQuiet(true);
    tidyBuilder.setDropEmptyParas(false);
    tidyBuilder.setTidyMark(false);
    tidyBuilder.setFixComments(false);
    tidyBuilder.setTrimEmptyElements(false);
    tidyBuilder.setJoinStyles(false);
    tidyBuilder.setXmlTags(true); // important, otherwise jtidy manipulates the markup
}
 

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
/**
 * Configures a new JTidy instance.
 */
private static Tidy newTidy() {
    Tidy tidy = new Tidy();
    tidy.setMessageListener(new TidyMessageListener() {

        @Override
        public void messageReceived(TidyMessage msg) {
        	logger.warn(String.format("HTML warning at %s:%s: %s", msg.getLine(), msg.getColumn(), msg.getMessage()));
        }
    });
    tidy.setDropEmptyParas(false);
    tidy.setDropFontTags(false);
    tidy.setDropProprietaryAttributes(false);
    tidy.setTrimEmptyElements(false);
    tidy.setXHTML(true);
    tidy.setIndentAttributes(false);
    tidy.setIndentCdata(false);
    tidy.setIndentContent(false);
    tidy.setQuiet(true);
    tidy.setShowWarnings(!Options.isQuietEnabled());
    tidy.setShowErrors(0);
    tidy.setEncloseBlockText(false);
    tidy.setEscapeCdata(false);
    tidy.setDocType("omit");
    tidy.setInputEncoding("UTF-8");
    tidy.setRawOut(true);
    tidy.setOutputEncoding("UTF-8");
    tidy.setFixUri(false);
    Properties prop = new Properties();
    prop.put("new-blocklevel-tags", "canvas");
    tidy.getConfiguration().addProps(prop);
    return tidy;
}
 

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
public cfData execute(cfSession _session, cfArgStructData argStruct ) throws cfmRunTimeException {
	try{
		Tidy tidy	= new Tidy();

		final StringBuilder	errors	= new StringBuilder(32);
		
		tidy.setMessageListener( new TidyMessageListener(){
		public void messageReceived(TidyMessage mess) {
			errors.append( "Line: " + mess.getLine() + "." + mess.getColumn() + "; " + mess.getMessage() + "\r\n" );
		}
		});

  	tidy.setSmartIndent( false );
  	tidy.setSpaces( 2 );
  	tidy.setTabsize( 2 );
  	tidy.setWraplen( 0 );

  	tidy.setLogicalEmphasis( true );
  	tidy.setMakeClean( true );
  	tidy.setQuiet( true );
  	tidy.setDropEmptyParas( true );
  	tidy.setXHTML( true );
  	tidy.setXmlSpace( true );
  	tidy.setTrimEmptyElements( true );
  	tidy.setBreakBeforeBR( false );
  	tidy.setUpperCaseTags( false );
  	tidy.setUpperCaseAttrs( false );
  	tidy.setWord2000( true );

  	tidy.setFixUri(false);
  	tidy.setFixBackslash( false );
  	tidy.setIndentAttributes( false );
  	tidy.setShowWarnings( false );
  	tidy.setShowErrors( 1 );
  	tidy.setOnlyErrors( false );

  	tidy.setPrintBodyOnly( false );
  	tidy.setJoinClasses( true );
  	tidy.setJoinStyles( true );
  	
  	String inHtml = getNamedStringParam(argStruct,"string","");
  	
  	StringReader reader = new StringReader( inHtml );
  	StringWriter writer = new StringWriter();
  	tidy.parse( reader, writer );
  	
  	if ( errors.length() != 0 ){
  		throwException( _session, errors.toString() );
  		return null;
  	}else{
  	
   	String outHtml	= writer.toString();
   	int c1	= outHtml.indexOf("<body>");
   	if ( c1 >= 0 ){
   		outHtml	= outHtml.substring( c1 + 6 );
   		c1 = outHtml.lastIndexOf("</body>");
   		if ( c1 >= 0 ){
   			outHtml	= outHtml.substring( 0, c1 );
   		}
   	}

   	return new cfStringData( outHtml );
  	}
  	
	}catch( Exception e ){
		throwException( _session, e.getMessage() );
		return null;
	}
}
 

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
private static String tidyHTML(String content)
								throws IOException{
	Tidy tidy = new Tidy();
	
	// set configuration values
	tidy.setDropEmptyParas(false); // drop empty P elements
	tidy.setDocType("omit"); // omit the doctype
	tidy.setEncloseBlockText(true); // wrap blocks of text in P elements
	tidy.setEncloseText(true); // wrap text right under BODY element in P elements
	tidy.setHideEndTags(false); // force optional end tags
	tidy.setIndentContent(false); // indent content for easy reading
	tidy.setLiteralAttribs(false); // no new lines in attributes
	tidy.setLogicalEmphasis(false); // replace i and b by em and strong, respectively
	tidy.setMakeClean(false); // strip presentational cruft
	tidy.setNumEntities(true); // convert entities to their numeric form
	tidy.setWord2000(true); // strip Word 2000 cruft
	tidy.setXHTML(true); // output XHTML
	tidy.setXmlPi(true); // add <?xml?> processing instruction
	
	// parse
	StringReader in = new StringReader(content);
	StringWriter out = new StringWriter();
	tidy.parse(in, out);
	in.close();
	out.close();
	String results = out.toString();
	
	// remove the XML namespace declaration,
	// since it makes trouble for us in the XPath
	// evaluator
	// FIXME: this is ghetto and needs to be fixed
	// with a namespace evaluator in the XPath section,
	// but namespace evaluators are a pain in the butt
	// to get working
	
	// String.replace() does not work on 1.4 JVMs when compiled
	// with 1.5 JVMs, even with target="1.4" (this is a known Java
	// bug). Using workaround instead. -- Brad Neuberg
	//results = results.replace("xmlns=\"http://www.w3.org/1999/xhtml\"", "");
	
	StringBuffer buffer = new StringBuffer(results);
	int startCut = buffer.indexOf("xmlns=\"http://www.w3.org/1999/xhtml\"");
	if(startCut != -1){
		buffer.replace(startCut, 
						startCut + "xmlns=\"http://www.w3.org/1999/xhtml\"".length(), 
						"");
		results = buffer.toString();
	}
	
	//System.out.println("tidied results="+results);
			
	return results;
}