[Htmlparser-cvs] htmlparser/src/org/htmlparser/filters AndFilter.java,1.2,1.3 CssSelectorNodeFilter.

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 422-6466

Update of /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters
In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv30655/htmlparser/src/org/htmlparser/filters

Modified Files:
	AndFilter.java CssSelectorNodeFilter.java 
	HasAttributeFilter.java HasChildFilter.java 
	HasParentFilter.java HasSiblingFilter.java 
	LinkRegexFilter.java LinkStringFilter.java 
	NodeClassFilter.java NotFilter.java OrFilter.java 
	RegexFilter.java TagNameFilter.java 
Log Message:
Documentation revamp part one.
Deprecated node decorators.
Added doSemanticAction for Text and Comment nodes.
Added missing sitecapturer scripts.
Fixed DOS batch files to work when called from any location.

Index: AndFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/AndFilter.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -C2 -d -r1.2 -r1.3
*** AndFilter.java	13 Feb 2005 20:36:00 -0000	1.2
--- AndFilter.java	10 Apr 2005 23:20:43 -0000	1.3
***************
*** 95,98 ****
--- 95,100 ----
       * Accept nodes that are acceptable to all of it's predicate filters.
       * @param node The node to check.
+      * @return <code>true</code> if all the predicate filters find the node
+      * is acceptable, <code>false</code> otherwise.
       */
      public boolean accept (Node node)

Index: RegexFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/RegexFilter.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -C2 -d -r1.2 -r1.3
*** RegexFilter.java	13 Feb 2005 20:36:00 -0000	1.2
--- RegexFilter.java	10 Apr 2005 23:20:43 -0000	1.3
***************
*** 54,58 ****
   * </pre>
   * which matches a date in yyyy-mm-dd format between 1900-01-01 and 2099-12-31,
!  * with a choice of five separators, dash, space, either slash or a period.
   * The year is matched by (19|20)\d\d which uses alternation to allow the
   * either 19 or 20 as the first two digits. The round brackets are mandatory.
--- 54,59 ----
   * </pre>
   * which matches a date in yyyy-mm-dd format between 1900-01-01 and 2099-12-31,
!  * with a choice of five separators, either a dash, a space, either kind of
!  * slash or a period.
   * The year is matched by (19|20)\d\d which uses alternation to allow the
   * either 19 or 20 as the first two digits. The round brackets are mandatory.
***************
*** 174,177 ****
--- 175,180 ----
       * Accept string nodes that match the regular expression.
       * @param node The node to check.
+      * @return <code>true</code> if the regular expression matches the 
+      * text of the node, <code>false</code> otherwise.
       */
      public boolean accept (Node node)

Index: HasParentFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/HasParentFilter.java,v
retrieving revision 1.6
retrieving revision 1.7
diff -C2 -d -r1.6 -r1.7
*** HasParentFilter.java	13 Feb 2005 20:36:00 -0000	1.6
--- HasParentFilter.java	10 Apr 2005 23:20:43 -0000	1.7
***************
*** 128,131 ****
--- 128,133 ----
       * filter.
       * @param node The node to check.
+      * @return <code>true</code> if the node has an acceptable parent,
+      * <code>false</code> otherwise.
       */
      public boolean accept (Node node)

Index: LinkStringFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/LinkStringFilter.java,v
retrieving revision 1.1
retrieving revision 1.2
diff -C2 -d -r1.1 -r1.2
*** LinkStringFilter.java	6 Apr 2005 10:20:23 -0000	1.1
--- LinkStringFilter.java	10 Apr 2005 23:20:43 -0000	1.2
***************
*** 38,42 ****
--- 38,49 ----
  public class LinkStringFilter implements NodeFilter
  {
+     /**
+      * The pattern to search for in the link.
+      */
      protected String mPattern;
+ 
+     /**
+      * Flag indicating case sensitive/insensitive search.
+      */
      protected boolean mCaseSensitive;
      

Index: LinkRegexFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/LinkRegexFilter.java,v
retrieving revision 1.1
retrieving revision 1.2
diff -C2 -d -r1.1 -r1.2
*** LinkRegexFilter.java	6 Apr 2005 10:20:23 -0000	1.1
--- LinkRegexFilter.java	10 Apr 2005 23:20:43 -0000	1.2
***************
*** 40,43 ****
--- 40,46 ----
  public class LinkRegexFilter implements NodeFilter
  {
+     /**
+      * The regular expression to use on the link.
+      */ 
      protected Pattern mRegex;
      
***************
*** 47,51 ****
       * @param regexPattern The pattern to match.
       */
!     public LinkRegexFilter (String regexPattern) throws Exception
      {
          this (regexPattern, true);
--- 50,54 ----
       * @param regexPattern The pattern to match.
       */
!     public LinkRegexFilter (String regexPattern)
      {
          this (regexPattern, true);
***************
*** 58,62 ****
       * @param caseSensitive Specifies case sensitivity for the matching process.
       */
!     public LinkRegexFilter (String regexPattern, boolean caseSensitive) throws Exception
      {
          if (caseSensitive)
--- 61,65 ----
       * @param caseSensitive Specifies case sensitivity for the matching process.
       */
!     public LinkRegexFilter (String regexPattern, boolean caseSensitive)
      {
          if (caseSensitive)

Index: OrFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/OrFilter.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -C2 -d -r1.2 -r1.3
*** OrFilter.java	13 Feb 2005 20:36:00 -0000	1.2
--- OrFilter.java	10 Apr 2005 23:20:43 -0000	1.3
***************
*** 93,96 ****
--- 93,98 ----
       * Accept nodes that are acceptable to any of it's predicate filters.
       * @param node The node to check.
+      * @return <code>true</code> if any of the predicate filters find the node
+      * is acceptable, <code>false</code> otherwise.
       */
      public boolean accept (Node node)

Index: HasChildFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/HasChildFilter.java,v
retrieving revision 1.3
retrieving revision 1.4
diff -C2 -d -r1.3 -r1.4
*** HasChildFilter.java	13 Feb 2005 20:36:00 -0000	1.3
--- HasChildFilter.java	10 Apr 2005 23:20:43 -0000	1.4
***************
*** 34,37 ****
--- 34,40 ----
  /**
   * This class accepts all tags that have a child acceptable to the filter.
+  * It can be set to operate recursively, that is perform a scan down
+  * through the node heirarchy in a breadth first traversal looking for any
+  * descendant that matches the predicate filter (which stops the search). 
   */
  public class HasChildFilter
***************
*** 123,126 ****
--- 126,131 ----
       * Accept tags with children acceptable to the filter.
       * @param node The node to check.
+      * @return <code>true</code> if the node has an acceptable child,
+      * <code>false</code> otherwise.
       */
      public boolean accept (Node node)

Index: HasSiblingFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/HasSiblingFilter.java,v
retrieving revision 1.1
retrieving revision 1.2
diff -C2 -d -r1.1 -r1.2
*** HasSiblingFilter.java	13 Feb 2005 20:36:00 -0000	1.1
--- HasSiblingFilter.java	10 Apr 2005 23:20:43 -0000	1.2
***************
*** 85,88 ****
--- 85,90 ----
       * Accept tags with a sibling acceptable to the filter.
       * @param node The node to check.
+      * @return <code>true</code> if the node has an acceptable sibling,
+      * <code>false</code> otherwise.
       */
      public boolean accept (Node node)

Index: CssSelectorNodeFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/CssSelectorNodeFilter.java,v
retrieving revision 1.4
retrieving revision 1.5
diff -C2 -d -r1.4 -r1.5
*** CssSelectorNodeFilter.java	17 Jul 2004 13:45:04 -0000	1.4
--- CssSelectorNodeFilter.java	10 Apr 2005 23:20:43 -0000	1.5
***************
*** 72,75 ****
--- 72,79 ----
      private NodeFilter therule;
  
+     /**
+      * Create a Cascading Style Sheet node filter.
+      * @param selector The selector expression.
+      */
      public CssSelectorNodeFilter(String selector)
      {
***************
*** 79,85 ****
      }
  
!     public boolean accept(Node n)
      {
!         return therule.accept(n);
      }
  
--- 83,95 ----
      }
  
!     /**
!      * Accept nodes that match the selector expression.
!      * @param node The node to check.
!      * @return <code>true</code> if the node matches,
!      * <code>false</code> otherwise.
!      */
!     public boolean accept(Node node)
      {
!         return therule.accept(node);
      }
  
***************
*** 235,241 ****
                  if ("~=".equals(rel) && val != null)
                      n = new AttribMatchFilter(unescape(attrib),
!                                                                         "\\b"
!                                                                         + val.replaceAll("([^a-zA-Z0-9])", "\\\\$1")
!                                                                         + "\\b");
                  else if ("=".equals(rel) && val != null)
                      n = new HasAttributeFilter(attrib, val);
--- 245,251 ----
                  if ("~=".equals(rel) && val != null)
                      n = new AttribMatchFilter(unescape(attrib),
!                         "\\b"
!                         + val.replaceAll("([^a-zA-Z0-9])", "\\\\$1")
!                         + "\\b");
                  else if ("=".equals(rel) && val != null)
                      n = new HasAttributeFilter(attrib, val);
***************
*** 249,252 ****
--- 259,268 ----
      }
  
+     /**
+      * Replace escape sequences in a string.
+      * @param escaped The string to examine.
+      * @return The argument with escape sequences replaced by their
+      * equivalent character.
+      */
      public static String unescape(String escaped)
      {
***************
*** 258,262 ****
              if (m.group(1) != null)
                  m.appendReplacement(result,
!                                                         String.valueOf((char)Integer.parseInt(m.group(1), 16)));
              else if (m.group(2) != null)
                  m.appendReplacement(result, m.group(2));
--- 274,278 ----
              if (m.group(1) != null)
                  m.appendReplacement(result,
!                     String.valueOf((char)Integer.parseInt(m.group(1), 16)));
              else if (m.group(2) != null)
                  m.appendReplacement(result, m.group(2));

Index: TagNameFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/TagNameFilter.java,v
retrieving revision 1.4
retrieving revision 1.5
diff -C2 -d -r1.4 -r1.5
*** TagNameFilter.java	13 Feb 2005 20:36:00 -0000	1.4
--- TagNameFilter.java	10 Apr 2005 23:20:43 -0000	1.5
***************
*** 87,90 ****
--- 87,92 ----
       * The end tags are available on the enclosing non-end tag.
       * @param node The node to check.
+      * @return <code>true</code> if the tag name matches,
+      * <code>false</code> otherwise.
       */
      public boolean accept (Node node)

Index: HasAttributeFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/HasAttributeFilter.java,v
retrieving revision 1.5
retrieving revision 1.6
diff -C2 -d -r1.5 -r1.6
*** HasAttributeFilter.java	13 Feb 2005 20:36:00 -0000	1.5
--- HasAttributeFilter.java	10 Apr 2005 23:20:43 -0000	1.6
***************
*** 119,122 ****
--- 119,124 ----
       * Accept tags with a certain attribute.
       * @param node The node to check.
+      * @return <code>true</code> if the node has the attribute
+      * (and value if that is being checked too), <code>false</code> otherwise.
       */
      public boolean accept (Node node)

Index: NodeClassFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/NodeClassFilter.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -C2 -d -r1.2 -r1.3
*** NodeClassFilter.java	13 Feb 2005 20:36:00 -0000	1.2
--- NodeClassFilter.java	10 Apr 2005 23:20:43 -0000	1.3
***************
*** 78,81 ****
--- 78,83 ----
       * Accept nodes that are assignable from the class provided in the constructor.
       * @param node The node to check.
+      * @return <code>true</code> if the node is the right class,
+      * <code>false</code> otherwise.
       */
      public boolean accept (Node node)

Index: NotFilter.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/filters/NotFilter.java,v
retrieving revision 1.3
retrieving revision 1.4
diff -C2 -d -r1.3 -r1.4
*** NotFilter.java	13 Feb 2005 22:45:47 -0000	1.3
--- NotFilter.java	10 Apr 2005 23:20:43 -0000	1.4
***************
*** 85,88 ****
--- 85,90 ----
       * Accept nodes that are not acceptable to the predicate filter.
       * @param node The node to check.
+      * @return <code>true</code> if the node is not acceptable to the
+      * predicate filter, <code>false</code> otherwise.
       */
      public boolean accept (Node node)