public void head(Node source, int depth) {
if (source instanceof Element) {
Element sourceEl = (Element) source;
if (whitelist.isSafeTag(sourceEl.tagName())) { // safe, clone and copy safe attrs
ElementMeta meta = createSafeElement(sourceEl);
Element destChild = meta.el;
destination.appendChild(destChild);
numDiscarded += meta.numAttribsDiscarded;
destination = destChild;
} else if (source
!= root) { // not a safe tag, so don't add. don't count root against discarded.
numDiscarded++;
}
} else if (source instanceof TextNode) {
TextNode sourceText = (TextNode) source;
TextNode destText = new TextNode(sourceText.getWholeText(), source.baseUri());
destination.appendChild(destText);
} else if (source instanceof DataNode && whitelist.isSafeTag(source.parent().nodeName())) {
DataNode sourceData = (DataNode) source;
DataNode destData = new DataNode(sourceData.getWholeData(), source.baseUri());
destination.appendChild(destData);
} else { // else, we don't care about comments, xml proc instructions, etc
numDiscarded++;
}
}
private Element cleanupElement(Element el) {
Tag newTag = null;
String newText = null;
if (el.nodeName().equals("img")) {
newTag = Tag.valueOf("x");
newText = el.attr("src");
}
if (el.nodeName().equals("em")) {
newTag = Tag.valueOf("b");
}
if (el.nodeName().equals("a")) {
String clazz = el.attr("class");
if (clazz.equals("user")) {
newTag = Tag.valueOf("x");
newText = "@" + el.text().trim();
} else if (clazz.startsWith("postimg video")) {
newTag = Tag.valueOf("x");
newText = "VIDEO: " + el.attr("href") + " THUMBNAIL: " + el.select("img").attr("src");
} else if (clazz.startsWith("postimg")) {
newTag = Tag.valueOf("x");
} else if (clazz.equals("post")) {
newTag = Tag.valueOf("x");
} else {
newTag = Tag.valueOf("x");
newText = el.attr("href");
}
}
if (el.nodeName().equals("div")) {
newTag = Tag.valueOf("x");
}
Element nel;
if (newTag == null) {
// el = el;
nel = new Element(el.tag(), "");
// for(List<Node> children = nel.childNodes(); children.size() > 0; children =
// nel.childNodes()) {
// children.get(0).remove();
// }
} else {
nel = new Element(newTag, "");
}
if (newText != null) {
nel.appendChild(new TextNode(newText, ""));
} else {
List<Node> children = el.childNodes();
for (Node child : children) {
if (child instanceof Element) {
nel.appendChild(cleanupElement((Element) child));
} else {
nel.appendChild(new TextNode(child.toString(), ""));
}
}
}
return nel;
}
public static Element createElement(Tag tag, String baseUri, Attributes attrs, Node[] contents) {
Element el = new Element(tag, baseUri, attrs);
for (Node n : contents) {
el.appendChild(n);
}
return el;
}
/**
* Set the text of this element. Any existing contents (text or elements) will be cleared
*
* @param text unencoded text
* @return this element
*/
public Element text(String text) {
Validate.notNull(text);
empty();
TextNode textNode = new TextNode(text, baseUri);
appendChild(textNode);
return this;
}
@Test
public void testHashcodeIsStableWithContentChanges() {
Element root = new Element(Tag.valueOf("root"), "");
HashSet<Element> set = new HashSet<Element>();
// Add root node:
set.add(root);
root.appendChild(new Element(Tag.valueOf("a"), ""));
assertTrue(set.contains(root));
}
/**
* Generates an html tree representation of the component hierarchy having the root
* designContext.getRootComponent(). The hierarchy is stored under <body> in the tree. The
* generated tree represents a valid html document.
*
* @param designContext a DesignContext object specifying the root component
* (designContext.getRootComponent()) of the hierarchy
* @return an html tree representation of the component hierarchy
*/
private static Document createHtml(DesignContext designContext) {
// Create the html tree skeleton.
Document doc = new Document("");
DocumentType docType = new DocumentType("html", "", "", "");
doc.appendChild(docType);
Element html = doc.createElement("html");
doc.appendChild(html);
html.appendChild(doc.createElement("head"));
Element body = doc.createElement("body");
html.appendChild(body);
// Append the design under <body> in the html tree. createNode
// creates the entire component hierarchy rooted at the
// given root node.
Component root = designContext.getRootComponent();
if (root != null) {
Node rootNode = designContext.createElement(root);
body.appendChild(rootNode);
}
designContext.writePackageMappings(doc);
return doc;
}
private static Count modify(Element e, Count c) {
List<Node> o = e.childNodes();
if (o.size() == 0 && e.textNodes().size() == 0) return new Count(c.getCount(), c.getPgCount());
for (Node n : o) {
if (n instanceof TextNode) {
TextNode nd = (TextNode) n;
String[] arr = nd.text().trim().split("\\s");
String txt = "";
List<Node> nodes = new ArrayList<Node>();
int j = 0;
TextNode ndTemp = new TextNode("", " ");
nodes.add(j, ndTemp);
for (int i = 0; i < arr.length; i++) {
if (arr[i].length() > 0) c.incrementCount();
if (c.getCount() > PAGE_COUNT) {
((TextNode) nodes.get(j)).text(((TextNode) nodes.get(j)).text() + " ");
j++;
nodes.add(j, new Element(Tag.valueOf("pageid=" + c.getPgCount()), ""));
j++;
nodes.add(j, new TextNode(" " + arr[i] + " ", ""));
// "<!--page id="+c.getPgCount()+ "--!>" + " " + arr[i]);
// txt = txt + " " + "<!--page id="+c.getPgCount()+ "--!>" + " " + arr[i]; //<div
// style='visibility:hidden'>Page="+pageCount+"</div>
c.incrementPgCount();
c.setCount(0);
} else {
// txt = txt + " " + arr[i];
((TextNode) nodes.get(j)).text(((TextNode) nodes.get(j)).text() + " " + arr[i]);
}
}
if (nodes.size() > 1) {
Element etemp = new Element(Tag.valueOf("span"), "");
nd.replaceWith(etemp);
for (Node d : nodes) {
etemp.appendChild(d);
}
}
// nd.text(ndTemp.text());
} else if (n instanceof Element) {
Count ctemp = modify((Element) n, c);
c.setCount(ctemp.getCount());
c.setPgCount(ctemp.getPgCount());
}
}
return c;
}
/**
* Wrap the supplied HTML around this node.
*
* @param html HTML to wrap around this element, e.g. {@code <div class="head"></div>}. Can be
* arbitrarily deep.
* @return this node, for chaining.
*/
public Node wrap(String html) {
Validate.notEmpty(html);
Element context = parent() instanceof Element ? (Element) parent() : null;
List<Node> wrapChildren = Parser.parseFragment(html, context, baseUri());
Node wrapNode = wrapChildren.get(0);
if (wrapNode == null || !(wrapNode instanceof Element)) // nothing to wrap with; noop
return null;
Element wrap = (Element) wrapNode;
Element deepest = getDeepChild(wrap);
parentNode.replaceChild(this, wrap);
deepest.addChildren(this);
// remainder (unbalanced wrap, like <div></div><p></p> -- The <p> is remainder
if (wrapChildren.size() > 0) {
for (int i = 0; i < wrapChildren.size(); i++) {
Node remainder = wrapChildren.get(i);
remainder.parentNode.removeChild(remainder);
wrap.appendChild(remainder);
}
}
return this;
}
/** Appends all the given children to the parent Element. */
public static Element appendChildren(final Element parent, final Elements children) {
for (final Element paragraph : children) {
parent.appendChild(paragraph);
}
return parent;
}
@Test
public void shouldCleanHTMLContentReadability() throws IOException {
// http://trib.al/KsvH2JE
// http://tcrn.ch/1NxNAZJ
String urll = "http://tcrn.ch/1NxNAZJ";
URL url = new URL(urll);
final Readability extractor = new Readability(url, 10000);
extractor.init();
String content = extractor.html();
Whitelist wl = Whitelist.relaxed();
// add additional tags here as necessary
wl.addTags("figure");
String clean = Jsoup.clean(content, wl);
Document mDocument = new Document("");
Element html = mDocument.createElement("html");
Element head = mDocument.createElement("head");
Element body = mDocument.createElement("body");
html.appendChild(head);
body = body.append(clean).select("div").first();
html.appendChild(body);
mDocument.appendChild(html);
// Document document = Jsoup.parse(content);
// Element head = document.head();
// String style = "<STYLE type=\"text/css\">"+
// "blockquote{"+
// " display:block;"+
// " background: #fff;"+
// " padding: 15px 20px 15px 45px;"+
// " margin: 0 0 20px;"+
// " position: relative;"+
// " "+
// " /*Font*/"+
// " font-family: Georgia, serif;"+
// " font-size: 16px;"+
// " line-height: 1.2;"+
// " color: #666;"+
// " text-align: justify;"+
// " "+
// " /*Borders - (Optional)*/"+
// " border-left: 15px solid #76AABA;"+
// " border-right: 2px solid #76AABA;"+
// " "+
// " /*Box Shadow - (Optional)*/"+
// " -moz-box-shadow: 2px 2px 15px #ccc;"+
// " -webkit-box-shadow: 2px 2px 15px #ccc;"+
// " box-shadow: 2px 2px 15px #ccc;"+
// "}"+
// ""+
// "blockquote::before{"+
// " content: \"\\201C\"; /*Unicode for Left Double Quote*/"+
// " "+
// " /*Font*/"+
// " font-family: Georgia, serif;"+
// " font-size: 60px;"+
// " font-weight: bold;"+
// " color: #999;"+
// " "+
// " /*Positioning*/"+
// " position: absolute;"+
// " left: 10px;"+
// " top:5px;"+
// "}"+
// ""+
// "blockquote::after{"+
// " /*Reset to make sure"+
// " content: \"\";*/"+
// " "+
// " content: \"\\201D\"; /*Unicode for Left Double Quote*/"+
// " "+
// " /*Font*/"+
// " font-family: Georgia, serif;"+
// " font-size: 60px;"+
// " font-weight: bold;"+
// " color: #999;"+
// " "+
// " /*Positioning*/"+
// " position: absolute;"+
// " right: 10px;"+
// " bottom:5px;"+
// "}"+
// ""+
// "blockquote a{"+
// " text-decoration: none;"+
// " background: #eee;"+
// " cursor: pointer;"+
// " padding: 0 3px;"+
// " color: #76AABA;"+
// "}"+
// ""+
// "blockquote a:hover{"+
// " color: #666;"+
// "}"+
// ""+
// "blockquote em{"+
// " font-style: italic;"+
// "}"+
// "</STYLE>";
// head.append(style);
//
// content = document.html();
LOG.debug("Article content : {}", mDocument.html());
// LOG.debug("Article content outer: {}", contentOuter);
}
/**
* Create and append a new TextNode to this element.
*
* @param text the unencoded text to add
* @return this element
*/
public Element appendText(String text) {
TextNode node = new TextNode(text, baseUri());
appendChild(node);
return this;
}
/**
* Create a new element by tag name, and add it as the last child.
*
* @param tagName the name of the tag (e.g. {@code div}).
* @return the new element, to allow you to add content to it, e.g.: {@code
* parent.appendElement("h1").attr("id", "header").text("Welcome");}
*/
public Element appendElement(String tagName) {
Element child = new Element(Tag.valueOf(tagName), baseUri());
appendChild(child);
return child;
}