Home » lucene-3.0.1-src » org.apache » lucene » xmlparser » [javadoc | source]

    1   package org.apache.lucene.xmlparser;
    2   import java.io.Reader;
    3   
    4   import javax.xml.parsers.DocumentBuilder;
    5   import javax.xml.parsers.DocumentBuilderFactory;
    6   
    7   import org.w3c.dom.Document;
    8   import org.w3c.dom.Element;
    9   import org.w3c.dom.Node;
   10   import org.xml.sax.InputSource;
   11   /**
   12    * Licensed to the Apache Software Foundation (ASF) under one or more
   13    * contributor license agreements.  See the NOTICE file distributed with
   14    * this work for additional information regarding copyright ownership.
   15    * The ASF licenses this file to You under the Apache License, Version 2.0
   16    * (the "License"); you may not use this file except in compliance with
   17    * the License.  You may obtain a copy of the License at
   18    *
   19    *     http://www.apache.org/licenses/LICENSE-2.0
   20    *
   21    * Unless required by applicable law or agreed to in writing, software
   22    * distributed under the License is distributed on an "AS IS" BASIS,
   23    * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
   24    * See the License for the specific language governing permissions and
   25    * limitations under the License.
   26    */
   27   
   28   /**
   29    * 
   30    */
   31   public class DOMUtils
   32   {
   33       public static Element getChildByTagOrFail(Element e, String name)	throws ParserException
   34   	{
   35   		Element kid = getChildByTagName(e, name);
   36   		if (null == kid)
   37   		{
   38   			throw new ParserException(e.getTagName() + " missing \"" + name
   39   					+ "\" child element");
   40   		}
   41   		return kid;
   42   	}
   43        
   44        public static Element getFirstChildOrFail(Element e) throws ParserException
   45   	{
   46   		Element kid = getFirstChildElement(e);
   47   		if (null == kid)
   48   		{
   49   			throw new ParserException(e.getTagName()
   50   					+ " does not contain a child element");
   51   		}
   52   		return kid;
   53   	}
   54        
   55   	public static String getAttributeOrFail(Element e, String name)	throws ParserException
   56   	{
   57   		String v = e.getAttribute(name);
   58   		if (null == v)
   59   		{
   60   			throw new ParserException(e.getTagName() + " missing \"" + name
   61   					+ "\" attribute");
   62   		}
   63   		return v;
   64   	}
   65       public static String getAttributeWithInheritanceOrFail(Element e,	String name) throws ParserException
   66   	{
   67   		String v = getAttributeWithInheritance(e, name);
   68   		if (null == v)
   69   		{
   70   			throw new ParserException(e.getTagName() + " missing \"" + name
   71   					+ "\" attribute");
   72   		}
   73   		return v;
   74   	}
   75       public static String getNonBlankTextOrFail(Element e) throws ParserException
   76   	{
   77   		String v = getText(e);
   78   		if (null != v)
   79   			v = v.trim();
   80   		if (null == v || 0 == v.length())
   81   		{
   82   			throw new ParserException(e.getTagName() + " has no text");
   83   		}
   84   		return v;
   85   	}
   86    
   87        
   88        
   89        
   90   	
   91   	
   92   	/* Convenience method where there is only one child Element of a given name */
   93   	public static Element getChildByTagName(Element e, String name)
   94   	{
   95   	       for (Node kid = e.getFirstChild(); kid != null; kid = kid.getNextSibling())
   96   		{
   97   			if( (kid.getNodeType()==Node.ELEMENT_NODE) && (name.equals(kid.getNodeName())) )
   98   			{
   99   				return (Element)kid;
  100   			}
  101   		}
  102   		return null;
  103   	}
  104   
  105   	/**
  106   	 * Returns an attribute value from this node, or first parent node with this attribute defined
  107   	 * @param element 
  108   	 * @param attributeName
  109   	 * @return A non-zero-length value if defined, otherwise null
  110   	 */
  111   	public static String getAttributeWithInheritance(Element element, String attributeName)
  112   	{
  113   		String result=element.getAttribute(attributeName);
  114   		if( (result==null)|| ("".equals(result) ) )
  115   		{
  116   			Node n=element.getParentNode();
  117   			if((n==element)||(n==null))
  118   			{
  119   				return null;
  120   			}
  121   			if(n instanceof Element)
  122   			{
  123   				Element parent=(Element) n;
  124   				return getAttributeWithInheritance(parent,attributeName);
  125   			}
  126   			return null; //we reached the top level of the document without finding attribute
  127   		}
  128   		return result;		
  129   	}
  130   
  131   
  132   
  133   	/* Convenience method where there is only one child Element of a given name */
  134   	public static String getChildTextByTagName(Element e, String tagName)
  135   	{
  136   		Element child=getChildByTagName(e,tagName);
  137   		if(child!=null)
  138   		{
  139   			return getText(child);
  140   		}
  141   		return null;
  142   	}
  143   
  144   	/* Convenience method to append a new child with text*/
  145   	public static Element insertChild(Element parent, String tagName, String text)
  146   	{
  147   	  	Element child = parent.getOwnerDocument().createElement(tagName);
  148   		parent.appendChild(child);
  149   		if(text!=null)
  150   		{
  151   		  	child.appendChild(child.getOwnerDocument().createTextNode(text));
  152   		}
  153   		return child;
  154   	}
  155   
  156   	public static String getAttribute(Element element, String attributeName, String deflt)
  157   	{
  158   		String result=element.getAttribute(attributeName);
  159   		if( (result==null)|| ("".equals(result) ) )
  160   		{
  161   			return deflt;
  162   		}
  163   		return result;
  164   	}
  165   	public static float getAttribute(Element element, String attributeName, float deflt)
  166   	{
  167   		String result=element.getAttribute(attributeName);
  168   		if( (result==null)|| ("".equals(result) ) )
  169   		{
  170   			return deflt;
  171   		}
  172   		return Float.parseFloat(result);
  173   	}	
  174   
  175   	public static int getAttribute(Element element, String attributeName, int deflt)
  176   	{
  177   		String result=element.getAttribute(attributeName);
  178   		if( (result==null)|| ("".equals(result) ) )
  179   		{
  180   			return deflt;
  181   		}
  182   		return Integer.parseInt(result);
  183   	}
  184   	
  185   	public static boolean getAttribute(Element element, String attributeName,
  186   			boolean deflt)
  187   	{
  188   		String result = element.getAttribute(attributeName);
  189   		if ((result == null) || ("".equals(result)))
  190   		{
  191   			return deflt;
  192   		}
  193   		return Boolean.valueOf(result).booleanValue();
  194   	}	
  195   
  196   	/* Returns text of node and all child nodes - without markup */
  197   	//MH changed to Node from Element 25/11/2005
  198   	public static String getText(Node e)
  199   	{
  200   		StringBuilder sb=new StringBuilder();
  201   		getTextBuffer(e, sb);
  202   		return sb.toString();
  203   	}
  204   	
  205   	public static Element getFirstChildElement(Element element)
  206   	{
  207   		for (Node kid = element.getFirstChild(); kid != null; kid = kid
  208   				.getNextSibling())
  209   		{
  210   			if (kid.getNodeType() == Node.ELEMENT_NODE) 
  211   			{
  212   				return (Element) kid;
  213   			}
  214   		}
  215   		return null;
  216   	}	
  217   
  218   	private static void getTextBuffer(Node e, StringBuilder sb)
  219   	{
  220   	    for (Node kid = e.getFirstChild(); kid != null; kid = kid.getNextSibling())
  221   		{
  222   			switch(kid.getNodeType())
  223   			{
  224   				case Node.TEXT_NODE:
  225   				{
  226   					sb.append(kid.getNodeValue());
  227   					break;
  228   				}
  229   				case Node.ELEMENT_NODE:
  230   				{
  231   					getTextBuffer(kid, sb);
  232   					break;
  233   				}
  234   				case Node.ENTITY_REFERENCE_NODE:
  235   				{
  236   					getTextBuffer(kid, sb);
  237   					break;
  238   				}
  239   			}
  240   		}
  241   	}
  242   
  243   	/**
  244   	* Helper method to parse an XML file into a DOM tree, given a reader.
  245   	* @param is reader of the XML file to be parsed
  246   	* @return an org.w3c.dom.Document object
  247   	*/
  248   	public static Document loadXML(Reader is)
  249   	{
  250   
  251   		DocumentBuilderFactory dbf = DocumentBuilderFactory.newInstance();
  252   		DocumentBuilder db = null;
  253   		
  254   		try
  255   		{
  256   			db = dbf.newDocumentBuilder();
  257   		}
  258   		catch (Exception se)
  259   		{
  260   			throw new RuntimeException("Parser configuration error", se);
  261   		}
  262   
  263   		// Step 3: parse the input file
  264   		org.w3c.dom.Document doc = null;
  265   		try
  266   		{
  267   			doc = db.parse(new InputSource(is));
  268   			//doc = db.parse(is);
  269   		}
  270   		catch (Exception se)
  271   		{
  272   			throw new RuntimeException("Error parsing file:" + se, se);
  273   		}
  274   
  275   		return doc;
  276   	}	
  277   }
  278   
  279   
  280   

Home » lucene-3.0.1-src » org.apache » lucene » xmlparser » [javadoc | source]