Java Tidy.setTidyMark方法代碼示例

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
public void tidy_init() {
	long bgn = System_.Ticks();
	wtr = new ByteArrayOutputStream();
	System.setProperty("line.separator", "\n");
	tidy = new Tidy(); // obtain a new Tidy instance
	tidy.setInputEncoding("utf-8");			// -utf8
	tidy.setOutputEncoding("utf-8");		// -utf8
	tidy.setDocType("\"\"");				// --doctype \"\"; set to empty else some wikis will show paragraph text with little vertical gap; PAGE:tr.b:
	tidy.setForceOutput(true);				// --force-output y 
	tidy.setQuiet(true);					// --quiet y
	tidy.setTidyMark(false);				// --tidy-mark n
	tidy.setWraplen(0);						// --wrap 0
	tidy.setIndentContent(true);			// --indent y; NOTE: true indents all content in edit box
	tidy.setQuoteNbsp(true);				// --quote-nbsp y
	tidy.setLiteralAttribs(true);			// --literal-attributes y
	tidy.setWrapAttVals(false);				// --wrap-attributes n
	tidy.setFixUri(false);					// --fix-url n
	tidy.setFixBackslash(false);			// --fix-backslash n
	tidy.setEncloseBlockText(true);			// --enclose-block-text y; NOTE: true creates extra <p>; very noticeable in sidebar
	tidy.setNumEntities(false);				// NOTE: true will convert all UTF-8 chars to &#val; which ruins readability
	tidy.setTrimEmptyElements(true);		// NOTE: tidy always trims (not even an option)
	tidy.setShowWarnings(false);			// NOTE: otherwise warnings printed to output window
	tidy.setShowErrors(0);					// NOTE: otherwise errors printed to output window; EX: Error: <time> is not recognized!
	app.Usr_dlg().Log_many("", "", "jtidy.init; elapsed=~{0}", System_.Ticks__elapsed_in_frac(bgn));
}
 

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
public static String formatXml(@NotNull String xml) throws TransformerException {
    StringReader stringReader = new StringReader(xml);
    Tidy tidy = new Tidy();
    tidy.setXmlOut(true);
    tidy.setInputEncoding("UTF-8");
    tidy.setOutputEncoding("UTF-8");
    tidy.setTidyMark(false);
    tidy.setForceOutput(true);
    tidy.setSmartIndent(true);
    tidy.setShowWarnings(false);
    tidy.setQuiet(true);
    StringWriter stringWriter = new StringWriter();
    tidy.parse(stringReader, stringWriter);
    return stringWriter.toString();
}
 

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
public static String formatHtml(String html) {
    StringReader stringReader = new StringReader(html);
    Tidy tidy = new Tidy();
    tidy.setXHTML(true);
    tidy.setInputEncoding("UTF-8");
    tidy.setOutputEncoding("UTF-8");
    tidy.setTidyMark(false);
    tidy.setSmartIndent(true);
    tidy.setForceOutput(true);
    tidy.setShowWarnings(false);
    tidy.setQuiet(true);
    StringWriter stringWriter = new StringWriter();
    tidy.parse(stringReader, stringWriter);
    return stringWriter.toString();
}
 

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
public void ConverterXhtml(String fileHtml, String fileXhtmlAux) throws Exception {

        Tidy tidy = new Tidy();

        FileInputStream in = new FileInputStream(fileHtml);
        FileOutputStream out = new FileOutputStream(fileXhtmlAux);

        tidy.setTidyMark(false);
        tidy.setDocType("omit");
        tidy.setAltText("");
        tidy.setFixBackslash(true);
        tidy.setFixComments(true);
        tidy.setXmlPi(true);
        tidy.setQuoteAmpersand(true);
        tidy.setQuoteNbsp(true);
        tidy.setNumEntities(true);
        tidy.setXmlOut(true);
        tidy.setWraplen(999);
        tidy.setWriteback(true);
        tidy.setQuoteMarks(true);
        tidy.setLogicalEmphasis(true);
        tidy.setEncloseText(true);
        tidy.setHideEndTags(true);
        tidy.setShowWarnings(false);
        tidy.setQuiet(true);
        tidy.setXHTML(true);
        tidy.parse(in, out);

        in.close();
        out.close();


    }
 

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
/**
 * Method used to transform the data from the html file given as parameter to 
 * xhtml format file which will be stored in the second file given.
 * @param fileHtml input html file path.
 * @param fileXhtmlAux output xhtml file path.
 * @throws Exception if the files can not be read or written.
 */
public void ConverterXhtml(String fileHtml, String fileXhtmlAux) throws Exception {

    Tidy tidy = new Tidy();

    FileInputStream in = new FileInputStream(fileHtml);
    FileOutputStream out = new FileOutputStream(fileXhtmlAux);

    tidy.setTidyMark(false);
    tidy.setDocType("omit");
    tidy.setAltText("");
    tidy.setFixBackslash(true);
    tidy.setFixComments(true);
    tidy.setXmlPi(true);
    tidy.setQuoteAmpersand(true);
    tidy.setQuoteNbsp(true);
    tidy.setNumEntities(true);
    tidy.setXmlOut(true);
    tidy.setWraplen(999);
    tidy.setWriteback(true);
    tidy.setQuoteMarks(true);
    tidy.setLogicalEmphasis(true);
    tidy.setEncloseText(true);
    tidy.setHideEndTags(true);
    tidy.setShowWarnings(false);
    tidy.setQuiet(true);
    tidy.setXHTML(true);
    tidy.parse(in, out);

    in.close();
    out.close();


}
 

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
public
HtmlToXml_Tidy()
{
    tidy_ = new Tidy();
    tidy_.setTidyMark(false);
    configuration_ = tidy_.getConfiguration();
    Properties properties = new Properties();        

    // SAX parser requires XML
    properties.put("output-xml", "true");

    // No need to add document generator to output
    properties.put("tidy-mark", "false");

    // Request that entities be output in their numeric form
    // rather than the character mnemonics: some browsers don't 
    // understand all the mnemonics.  Unfortunately this forces
    // all entities except for  , <, >, and & 
    // to be output as numerics, even if they were input as 
    // characters.
    properties.put("numeric-entities", "true");

    // Tell Tidy not to perform any line wrapping.
    // Otherwise the browser whitespace bug may reappear
    // if the line gets too long.
    properties.put("wrap", "0");

    // Tell Tidy not to output messages that aren't errors.
    // This makes builds look cleaner.
    properties.put("quiet", "true");
    
    configuration_.addProps(properties);
}
 

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
/**
 * Return a document based on file,
 * Document tpe 
 * 
 * @param file
 * @return
 */
public Document parseFileToDocument(File file, Boolean setXML, Boolean setXHTML, Boolean setParseMark, String  docType){
	
	tidy = new Tidy();
	
	//default format is XML 
	tidy.setXmlOut(setXML);
	
	//set if extensible html
	tidy.setXHTML(setXHTML);
	
	//this removes the tidy meta tag in the header
	tidy.setTidyMark(setParseMark);
	
	tidy.setDocType(docType);
	
	FileInputStream fis;
	Document document = null;
	try {
		fis = new FileInputStream(file);
	
		//parse input stream and return a DOM Document
		document = tidy.parseDOM(fis, null);
		
		fis.close();

	} catch (Exception e) {
		throw new RuntimeException("Unable to parse the file :" + file.getAbsolutePath() + " throws", e);
	}
	
	return document;
}
 

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
public static void main(String[] args) throws IOException, ParserConfigurationException, DocumentException
   {
       if (args.length != 2)
       {
           System.out.println("Error: You should type as: convertToPDF htmlFile pdfOut");
           System.exit(1);
       }

       String htmlFile = args[0];
       String pdfOutFile = args[1];//"/home/sardor/Downloads/CV_B/test.pdf";

       FileInputStream input = reOpenStream(htmlFile);


       //Append required HTML elements via JSoup
       HtmlValidator validator = new HtmlValidator(input, htmlFile);
       if(validator.hasError()==1)
       {
           System.out.println("Error: Something wrong with HTML file");
           System.exit(1);
       }

       //Redefine file stream
       input = reOpenStream(htmlFile);

       //Validate out html document
       Tidy tidy=new Tidy();
       tidy.setInputEncoding("UTF-8");
       tidy.setOutputEncoding("UTF-8");
       tidy.setIndentCdata(true);
       tidy.setQuiet(true);
       tidy.setForceOutput(true);
       tidy.setTidyMark(false);
       tidy.setXHTML(true);
       tidy.setJoinClasses(true);
       tidy.setSmartIndent(true);

       //Convert to XML view
       Document xhtmlDoc = tidy.parseDOM(input, null);

       //Create PDF file
       OutputStream os = new FileOutputStream(pdfOutFile);

       //Render pdf doc
       ITextRenderer renderer = new ITextRenderer();
       renderer.setDocument(xhtmlDoc,null);
       renderer.layout();
       renderer.createPDF(os);

       os.close();
}
 

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
/**
 * Return a document from the input stream
 * 
 * @param is
 * @return
 */
public Document parseInputStreamToDocument(InputStream is, Boolean setXML, Boolean setXHTML, Boolean setParseMark, String  docType){
	
	tidy = new Tidy();
	
	//default format is XML 
	tidy.setXmlOut(setXML);
	
	//set if extensible html
	tidy.setXHTML(setXHTML);
	
	//this removes the tidy meta tag in the header
	tidy.setTidyMark(setParseMark);
	
	tidy.setDocType(docType);
	
	Document document = null;
	try {
		//parse input stream and return a DOM Document
		document = tidy.parseDOM(is, null);

	} catch (Exception e) {
		throw new RuntimeException("Unable to parse the given input stream ");
	}
	

	
	return document;
}