日期:2014-05-17  浏览次数:20619 次

HtmlUnit解析html会丢掉不可见的Element

最近使用htmlunit来作为后端抽取数据,htmlunit的DOM解析,使用xpath定位结点的过程发现有这个问题。

不知道是故意这么做,还是个bug。

于是对重写了前端获取xpath的代码,让它也忽略不可见Element,保持一致,可以解决我们的问题:

function getXPath(element){
	if (element && element.id)
		return '//*[@id="' + element.id + '"]';

	var paths = [];

	for (; element && element.nodeType == 1; element = element.parentNode){
		var index = 0;
		for (var sibling = element.previousSibling; sibling; sibling = sibling.previousSibling){
			if (sibling.localName == element.localName && isVisible(sibling))
				++index;
		}
		var tagName = element.localName.toLowerCase();
		var pathIndex = (index ? "[" + (index+1) + "]" : "");
		paths.splice(0, 0, tagName + pathIndex);
	}
	return paths.length ? "/" + paths.join("/") : null;    
};
?

?判断是否可见:

function isVisible(element){
	var doc = element.ownerDocument;
	var docView = XPCOMUtils.QI(doc,_CI.nsIDOMDocumentView);
	var viewCss = XPCOMUtils.QI(docView.defaultView,_CI.nsIDOMViewCSS);
	var computedCss = viewCss.getComputedStyle(element,"");
	var visiable = computedCss.getPropertyCSSValue("visibility").getStringValue();
	var display = computedCss.getPropertyCSSValue("display").getStringValue();
	return visiable != "hidden" && display != "none";
}

其中XPCOMUtils是XPCOM的QueryInterface, createInstance,getService的一个代码简化的封装:

var _CI = Components.interfaces;
var _CC = Components.classes;

function XPCOMUtils() {}

(function() {

this.CCSV = function(cName, ifaceName)
{
	if (_CC[cName])
		return _CC[cName].getService(_CI[ifaceName]);  // if fbs fails to load, the error can be _CC[cName] has no properties
	else
		alert("Can't get the components class name: " + cName);
};

this.CCIN = function(cName, ifaceName)
{
    return _CC[cName].createInstance(_CI[ifaceName]);
};

this.QI = function(obj, iface)
{
    return obj.QueryInterface(iface);
};

// ************************************************************************************************

}).apply(XPCOMUtils);