From: David ‘Bombe’ Roden <bombe@pterodactylus.net>
Date: Mon, 9 Mar 2009 20:24:39 +0000 (+0100)
Subject: Basic implementation of HTML parser.
X-Git-Url: https://git.pterodactylus.net/?a=commitdiff_plain;h=80488ec32fe623903f8717a969af2c2089974c57;p=arachne.git

Basic implementation of HTML parser.
---

diff --git a/src/net/pterodactylus/arachne/parser/HtmlEditorKitParser.java b/src/net/pterodactylus/arachne/parser/HtmlEditorKitParser.java
index 45ba774..47e2d52 100644
--- a/src/net/pterodactylus/arachne/parser/HtmlEditorKitParser.java
+++ b/src/net/pterodactylus/arachne/parser/HtmlEditorKitParser.java
@@ -3,16 +3,51 @@
  */
 package net.pterodactylus.arachne.parser;
 
+import java.io.BufferedReader;
+import java.io.IOException;
 import java.io.InputStream;
+import java.io.InputStreamReader;
 
+import javax.swing.text.MutableAttributeSet;
+import javax.swing.text.html.HTML;
 import javax.swing.text.html.HTMLEditorKit;
+import javax.swing.text.html.HTML.Tag;
+import javax.swing.text.html.HTMLEditorKit.ParserCallback;
+import javax.swing.text.html.parser.ParserDelegator;
+
+import de.ina.util.io.Closer;
 
 /**
- * {@link Parser} implementation based on Swing’s {@link HTMLEditorKit}.
+ * {@link Parser} implementation based on Swing’s {@link HTMLEditorKit}. This
+ * parser can be re-used but is not thread-safe!
  *
  * @author David ‘Bombe’ Roden <bombe@pterodactylus.net>
  */
-public class HtmlEditorKitParser implements Parser {
+public class HtmlEditorKitParser extends HTMLEditorKit.ParserCallback implements Parser {
+
+	/** The parser listener. */
+	private ParserListener parserListener;
+
+	/** The current input stream. */
+	private InputStream inputStream;
+
+	/** Whether we’re currently parsing the title. */
+	private boolean inTitle;
+
+	/** The title text. */
+	private String titleText;
+
+	/** Whether we’re currently parsing a link. */
+	private boolean inLink;
+
+	/** The target of the link. */
+	private String linkTarget;
+
+	/** The title attribute of the link. */
+	private String linkTitle;
+
+	/** The text of the link. */
+	private StringBuilder linkText;
 
 	/**
 	 * {@inheritdoc}
@@ -21,6 +56,77 @@ public class HtmlEditorKitParser implements Parser {
 	 *      InputStream, String)
 	 */
 	public void parse(ParserListener parserListener, InputStream inputStream, String charset) throws IOException {
+		this.parserListener = parserListener;
+		this.inputStream = inputStream;
+		InputStreamReader inputStreamReader = null;
+		BufferedReader htmlReader = null;
+		try {
+			inputStreamReader = new InputStreamReader(inputStream, charset);
+			htmlReader = new BufferedReader(inputStreamReader);
+			new ParserDelegator().parse(htmlReader, this, true);
+		} finally {
+			Closer.close(htmlReader);
+			Closer.close(inputStreamReader);
+		}
+	}
+
+	//
+	// METHODS FROM ParserCallback
+	//
+
+	/**
+	 * {@inheritDoc}
+	 */
+	@Override
+	public void handleStartTag(Tag tag, MutableAttributeSet attributeSet, int position) {
+		if (tag == Tag.TITLE) {
+			if (!attributeSet.containsAttribute(ParserCallback.IMPLIED, Boolean.TRUE)) {
+				inTitle = true;
+			}
+		} else if (tag == Tag.A) {
+			String href = (String) attributeSet.getAttribute(HTML.Attribute.HREF);
+			if (href != null) {
+				linkTarget = href;
+				inLink = true;
+				linkText = new StringBuilder();
+				linkTitle = null;
+				String title = (String) attributeSet.getAttribute(HTML.Attribute.TITLE);
+				if (title != null) {
+					linkTitle = title;
+				}
+			}
+		}
+		if (inLink) {
+			if (tag.breaksFlow()) {
+				linkText.append(' ');
+			}
+		}
+	}
+
+	/**
+	 * {@inheritDoc}
+	 */
+	@Override
+	public void handleText(char[] data, int pos) {
+		if (inTitle) {
+			titleText = new String(data);
+		} else if (inLink) {
+			linkText.append(data);
+		}
+	}
+
+	/**
+	 * {@inheritDoc}
+	 */
+	@Override
+	public void handleEndTag(Tag tag, int position) {
+		if (tag == Tag.TITLE) {
+			inTitle = false;
+			parserListener.parsedTitle(inputStream, titleText);
+		} else if (tag == Tag.A) {
+			inLink = false;
+			parserListener.parsedLink(inputStream, linkTarget, linkTitle, linkText.toString());
+		}
 	}
 
 }