[Htmlparser-cvs] htmlparser/src/org/htmlparser/visitors HtmlPage.java,1.35,1.36 NodeVisitor.java,1.3
Brought to you by:
derrickoswald
Menu
▾
▴
[Htmlparser-cvs] htmlparser/src/org/htmlparser/visitors HtmlPage.java,1.35,1.36 NodeVisitor.java,1.30,1.31 TagFindingVisitor.java,1.36,1.37 TextExtractingVisitor.java,1.34,1.35 UrlModifyingVisitor.java,1.33,1.34
Update of /cvsroot/htmlparser/htmlparser/src/org/htmlparser/visitors
In directory sc8-pr-cvs1:/tmp/cvs-serv11047/visitors
Modified Files:
HtmlPage.java NodeVisitor.java TagFindingVisitor.java
TextExtractingVisitor.java UrlModifyingVisitor.java
Log Message:
Fixed up the broken visitor logic.
Added some docos on NodeVisitor.
Index: HtmlPage.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/visitors/HtmlPage.java,v
retrieving revision 1.35
retrieving revision 1.36
diff -C2 -d -r1.35 -r1.36
*** HtmlPage.java 28 Sep 2003 15:34:00 -0000 1.35
--- HtmlPage.java 28 Sep 2003 19:30:04 -0000 1.36
***************
*** 64,85 ****
public void visitTag(Tag tag)
{
! if (tag.isEndTag ())
! {
if (isBodyTag(tag))
! bodyTagBegin = false;
! addTagToBodyIfApplicable(tag);
}
! else
! {
! addTagToBodyIfApplicable(tag);
! if (isTable(tag)) {
! tables.add(tag);
! }
! else {
! if (isBodyTag(tag))
! bodyTagBegin = true;
! }
! }
}
--- 64,83 ----
public void visitTag(Tag tag)
{
! addTagToBodyIfApplicable(tag);
!
! if (isTable(tag)) {
! tables.add(tag);
! }
! else {
if (isBodyTag(tag))
! bodyTagBegin = true;
}
! }
! public void visitEndTag(Tag tag)
! {
! if (isBodyTag(tag))
! bodyTagBegin = false;
! addTagToBodyIfApplicable(tag);
}
Index: NodeVisitor.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/visitors/NodeVisitor.java,v
retrieving revision 1.30
retrieving revision 1.31
diff -C2 -d -r1.30 -r1.31
*** NodeVisitor.java 28 Sep 2003 15:34:00 -0000 1.30
--- NodeVisitor.java 28 Sep 2003 19:30:04 -0000 1.31
***************
*** 36,90 ****
import org.htmlparser.tags.TitleTag;
! public abstract class NodeVisitor {
! private boolean recurseChildren;
! private boolean recurseSelf;
!
! public NodeVisitor() {
! this(true);
}
!
! public NodeVisitor(boolean recurseChildren) {
! this.recurseChildren = recurseChildren;
! this.recurseSelf = true;
}
!
! public NodeVisitor(boolean recurseChildren,boolean recurseSelf) {
! this.recurseChildren = recurseChildren;
! this.recurseSelf = recurseSelf;
}
! public void visitTag(Tag tag) {
!
}
! public void visitStringNode(StringNode stringNode) {
}
!
! public void visitLinkTag(LinkTag linkTag) {
}
!
! public void visitImageTag(ImageTag imageTag) {
}
!
! public void visitTitleTag(TitleTag titleTag) {
!
}
! public void visitRemarkNode(RemarkNode remarkNode) {
!
}
! public boolean shouldRecurseChildren() {
! return recurseChildren;
}
!
! public boolean shouldRecurseSelf() {
! return recurseSelf;
}
! /**
! * Override this method if you wish to do special
! * processing upon completion of parsing
! */
! public void finishedParsing() {
}
}
--- 36,154 ----
import org.htmlparser.tags.TitleTag;
! /**
! * The base class for the 'Visitor' pattern.
! * Classes that wish to use <code>visitAllNodesWith()</code> will subclass
! * this class and provide implementations for methods they are interested in
! * processing.<p>
! * The operation of <code>visitAllNodesWith()</code> is to call
! * <code>beginParsing()</code>, then <code>visitXXX()</code> according to the
! * types of nodes encountered in depth-first order and finally
! * <code>finishedParsing()</code>.<p>
! * There are currently three specialized <code>visitXXX()</code> calls for
! * titles, images and links. Thes call their specialized visit, and then
! * perform the generic processing.
! * Typical code to print all the link tags:
! * <pre>
! * import org.htmlparser.Parser;
! * import org.htmlparser.tags.LinkTag;
! * import org.htmlparser.util.ParserException;
! * import org.htmlparser.visitors.NodeVisitor;
! *
! * public class Visitor extends NodeVisitor
! * {
! * public Visitor ()
! * {
! * }
! * public void visitLinkTag (LinkTag linkTag)
! * {
! * System.out.println (linkTag);
! * }
! * public static void main (String[] args) throws ParserException
! * {
! * Parser parser = new Parser ("http://cbc.ca");
! * parser.registerScanners ();
! * Visitor visitor = new Visitor ();
! * parser.visitAllNodesWith (visitor);
! * }
! * }
! * </pre>
! */
! public abstract class NodeVisitor
! {
! private boolean mRecurseChildren;
! private boolean mRecurseSelf;
!
! public NodeVisitor ()
! {
! this (true);
}
!
! public NodeVisitor (boolean recurseChildren)
! {
! this (recurseChildren, true);
}
!
! public NodeVisitor (boolean recurseChildren, boolean recurseSelf)
! {
! mRecurseChildren = recurseChildren;
! mRecurseSelf = recurseSelf;
}
! /**
! * Override this method if you wish to do special
! * processing prior to the start of parsing.
! */
! public void beginParsing ()
! {
}
! public void visitTag (Tag tag)
! {
!
}
!
! public void visitEndTag (Tag tag)
! {
!
}
!
! public void visitStringNode (StringNode stringNode)
! {
}
!
! public void visitRemarkNode (RemarkNode remarkNode)
! {
!
}
!
! /**
! * Override this method if you wish to do special
! * processing upon completion of parsing.
! */
! public void finishedParsing ()
! {
}
! public void visitLinkTag (LinkTag linkTag)
! {
}
!
! public void visitImageTag (ImageTag imageTag)
! {
! }
!
! public void visitTitleTag (TitleTag titleTag)
! {
!
}
! public boolean shouldRecurseChildren ()
! {
! return (mRecurseChildren);
! }
!
! public boolean shouldRecurseSelf ()
! {
! return (mRecurseSelf);
}
}
Index: TagFindingVisitor.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/visitors/TagFindingVisitor.java,v
retrieving revision 1.36
retrieving revision 1.37
diff -C2 -d -r1.36 -r1.37
*** TagFindingVisitor.java 28 Sep 2003 15:34:00 -0000 1.36
--- TagFindingVisitor.java 28 Sep 2003 19:30:04 -0000 1.37
***************
*** 69,86 ****
public void visitTag(Tag tag)
{
- if (tag.isEndTag ())
- {
- if (!endTagCheck) return;
- for (int i=0;i<tagsToBeFound.length;i++)
- if (tag.getTagName().substring (1).equalsIgnoreCase(tagsToBeFound[i]))
- {
- endTagCount[i]++;
- endTags[i].add(tag);
- }
- }
for (int i=0;i<tagsToBeFound.length;i++)
if (tag.getTagName().equalsIgnoreCase(tagsToBeFound[i])) {
count[i]++;
tags[i].add(tag);
}
}
--- 69,87 ----
public void visitTag(Tag tag)
{
for (int i=0;i<tagsToBeFound.length;i++)
if (tag.getTagName().equalsIgnoreCase(tagsToBeFound[i])) {
count[i]++;
tags[i].add(tag);
+ }
+ }
+
+ public void visitEndTag(Tag tag)
+ {
+ if (!endTagCheck) return;
+ for (int i=0;i<tagsToBeFound.length;i++)
+ if (tag.getTagName().equalsIgnoreCase(tagsToBeFound[i]))
+ {
+ endTagCount[i]++;
+ endTags[i].add(tag);
}
}
Index: TextExtractingVisitor.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/visitors/TextExtractingVisitor.java,v
retrieving revision 1.34
retrieving revision 1.35
diff -C2 -d -r1.34 -r1.35
*** TextExtractingVisitor.java 28 Sep 2003 15:34:00 -0000 1.34
--- TextExtractingVisitor.java 28 Sep 2003 19:30:04 -0000 1.35
***************
*** 67,74 ****
}
- public void visitTitleTag(TitleTag titleTag) {
- textAccumulator.append(titleTag.getTitle ());
- }
-
private String replaceNonBreakingSpaceWithOrdinarySpace(String text) {
return text.replace('\u00a0',' ');
--- 67,70 ----
***************
*** 77,90 ****
public void visitTag(Tag tag)
{
! if (tag.isEndTag ())
! {
! if (isPreTag(tag))
! preTagBeingProcessed = false;
! }
! else
! {
! if (isPreTag(tag))
! preTagBeingProcessed = true;
! }
}
--- 73,84 ----
public void visitTag(Tag tag)
{
! if (isPreTag(tag))
! preTagBeingProcessed = true;
! }
!
! public void visitEndTag(Tag tag)
! {
! if (isPreTag(tag))
! preTagBeingProcessed = false;
}
Index: UrlModifyingVisitor.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/visitors/UrlModifyingVisitor.java,v
retrieving revision 1.33
retrieving revision 1.34
diff -C2 -d -r1.33 -r1.34
*** UrlModifyingVisitor.java 28 Sep 2003 15:34:00 -0000 1.33
--- UrlModifyingVisitor.java 28 Sep 2003 19:30:05 -0000 1.34
***************
*** 62,66 ****
public void visitImageTag(ImageTag imageTag) {
imageTag.setImageURL(linkPrefix + imageTag.getImageURL());
- modifiedResult.append(imageTag.toHtml());
}
--- 62,65 ----
***************
*** 71,78 ****
public void visitTag(Tag tag)
{
! if (tag.isEndTag ())
! modifiedResult.append(tag.toHtml());
! else
! modifiedResult.append(tag.toHtml());
}
--- 70,79 ----
public void visitTag(Tag tag)
{
! modifiedResult.append(tag.toHtml());
! }
!
! public void visitEndTag(Tag tag)
! {
! modifiedResult.append(tag.toHtml());
}
|
