Java Tidy.setIndentContent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
/**
 * 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
/** Convenience method if prettified HTML is needed. E.g. in order to simplify debugging generated HTML
 * often available as a single one liner. DONT prettify production HTML as it might slow down the browser */
public static String prettifyHTML(String htmlString) {
	StringWriter stringWriter = new StringWriter();
	
	Tidy tidy = new Tidy();
	tidy.setIndentContent(true);
	tidy.parse(new ByteArrayInputStream(htmlString.getBytes()), stringWriter);
	
	return stringWriter.toString();
}
 

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
public Beautifier() {
    tidy = new Tidy();
    
    tidy.setWraplen(0);
    //tidy.setDropEmptyParas(false);
    tidy.setDropFontTags(false);
    tidy.setDropProprietaryAttributes(false);
    tidy.setIndentContent(true);

}
 

import org.w3c.tidy.Tidy; //導入方法依賴的package包/類
public String valueOf(final String string) {
    if (StringUtils.isBlank(string)) {
        return removeBlankspaces ? null : string;
    }

    final Tidy tidy = new Tidy(); // obtain a new Tidy instance
    tidy.setXHTML(false); // set desired config options using tidy setters
    tidy.setQuiet(true);
    tidy.setShowErrors(0);
    tidy.setShowWarnings(false);
    tidy.setIndentContent(false);
    tidy.setXmlOut(true);

    final Document document = tidy.parseDOM(new StringReader(string), null);
    removeBadNodes(document);

    final NodeList bodies = document.getElementsByTagName("body");
    if (bodies.getLength() == 0) {
        // No body element? return null
        return null;
    } else {
        // Result will contain the xml header plus the body element itself. We need to body content only
        String result = XmlHelper.toString(bodies.item(0));
        result = result.substring(result.indexOf("<body>") + "<body>".length(), result.indexOf("</body>"));
        // Remove the nbsps
        if (removeBlankspaces) {
            int begin = 0;
            while (result.charAt(begin) == NBSP) {
                begin++;
                if (begin == result.length()) {
                    // All the string was NBSPs
                    return null;
                }
            }
            int end = result.length();
            while (result.charAt(end - 1) == NBSP) {
                end--;
            }
            return StringUtils.trimToNull(result.substring(begin, end));
        } else {
            return StringUtils.trimToNull(result);
        }
    }
}
 

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;
}