Author: jukka
Date: Fri Oct 16 09:32:02 2009
New Revision: 825821
URL: http://svn.apache.org/viewvc?rev=825821&view=rev
Log:
TIKA-287: HtmlParser should resolve relative paths in <a href="xxx"> elements
Move also the <title/> and <meta/> tag handling to the Body/HtmlHandler class
Added:
lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlHandler.java
- copied, changed from r825818,
lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/BodyHandler.java
Removed:
lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/BodyHandler.java
Modified:
lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlParser.java
Copied:
lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlHandler.java
(from r825818,
lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/BodyHandler.java)
URL:
http://svn.apache.org/viewvc/lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlHandler.java?p2=lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlHandler.java&p1=lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/BodyHandler.java&r1=825818&r2=825821&rev=825821&view=diff
==============================================================================
---
lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/BodyHandler.java
(original)
+++
lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlHandler.java
Fri Oct 16 09:32:02 2009
@@ -21,12 +21,14 @@
import java.util.Map;
import java.util.Set;
+import org.apache.tika.metadata.Metadata;
import org.apache.tika.sax.TextContentHandler;
import org.apache.tika.sax.XHTMLContentHandler;
import org.xml.sax.Attributes;
+import org.xml.sax.ContentHandler;
import org.xml.sax.SAXException;
-class BodyHandler extends TextContentHandler {
+class HtmlHandler extends TextContentHandler {
/**
* Set of safe mappings from incoming HTML elements to outgoing
@@ -71,54 +73,103 @@
private final XHTMLContentHandler xhtml;
+ private final Metadata metadata;
+
+ private int bodyLevel = 0;
+
private int discardLevel = 0;
- public BodyHandler(XHTMLContentHandler xhtml) {
+ private int titleLevel = 0;
+
+ private StringBuilder title = new StringBuilder();
+
+ private HtmlHandler(XHTMLContentHandler xhtml, Metadata metadata) {
super(xhtml);
this.xhtml = xhtml;
+ this.metadata = metadata;
+ }
+
+ public HtmlHandler(ContentHandler handler, Metadata metadata) {
+ this(new XHTMLContentHandler(handler, metadata), metadata);
}
@Override
public void startElement(
String uri, String local, String name, Attributes atts)
throws SAXException {
- if (discardLevel != 0) {
+ if ("TITLE".equals(name) || titleLevel > 0) {
+ titleLevel++;
+ }
+ if ("BODY".equals(name) || bodyLevel > 0) {
+ bodyLevel++;
+ }
+ if (DISCARD_ELEMENTS.contains(name) || discardLevel > 0) {
discardLevel++;
- } else if (DISCARD_ELEMENTS.contains(name)) {
- discardLevel = 1;
- } else if (SAFE_ELEMENTS.containsKey(name)) {
- xhtml.startElement(SAFE_ELEMENTS.get(name));
- } else if ("A".equals(name)) {
- String href = atts.getValue("href");
- if (href != null) {
- xhtml.startElement("a", "href", href);
- } else {
- String anchor = atts.getValue("name");
- if (anchor != null) {
- xhtml.startElement("a", "name", anchor);
+ }
+
+ if (bodyLevel == 0 && discardLevel == 0 && "META".equals(name)) {
+ String content = atts.getValue("content");
+ if (atts.getValue("http-equiv") != null && content != null) {
+ metadata.set(atts.getValue("http-equiv"), content);
+ }
+ if (atts.getValue("name") != null && content != null) {
+ metadata.set(atts.getValue("name"), content);
+ }
+ }
+
+ if (bodyLevel > 0 && discardLevel == 0) {
+ if (SAFE_ELEMENTS.containsKey(name)) {
+ xhtml.startElement(SAFE_ELEMENTS.get(name));
+ } else if ("A".equals(name)) {
+ String href = atts.getValue("href");
+ if (href != null) {
+ xhtml.startElement("a", "href", href);
} else {
- xhtml.startElement("a");
+ String anchor = atts.getValue("name");
+ if (anchor != null) {
+ xhtml.startElement("a", "name", anchor);
+ } else {
+ xhtml.startElement("a");
+ }
}
}
}
+
+ title.setLength(0);
}
@Override
public void endElement(
String uri, String local, String name) throws SAXException {
- if (discardLevel != 0) {
+ if (bodyLevel > 0 && discardLevel == 0) {
+ if (SAFE_ELEMENTS.containsKey(name)) {
+ xhtml.endElement(SAFE_ELEMENTS.get(name));
+ } else if ("A".equals(name)) {
+ xhtml.endElement("a");
+ }
+ }
+
+ if (titleLevel > 0) {
+ titleLevel--;
+ if (titleLevel == 0) {
+ metadata.set(Metadata.TITLE, title.toString().trim());
+ }
+ }
+ if (bodyLevel > 0) {
+ bodyLevel--;
+ }
+ if (discardLevel > 0) {
discardLevel--;
- } else if (SAFE_ELEMENTS.containsKey(name)) {
- xhtml.endElement(SAFE_ELEMENTS.get(name));
- } else if ("A".equals(name)) {
- xhtml.endElement("a");
}
}
@Override
public void characters(char[] ch, int start, int length)
throws SAXException {
- if (discardLevel == 0) {
+ if (titleLevel > 0 && bodyLevel == 0) {
+ title.append(ch, start, length);
+ }
+ if (bodyLevel > 0 && discardLevel == 0) {
super.characters(ch, start, length);
}
}
@@ -126,7 +177,7 @@
@Override
public void ignorableWhitespace(char[] ch, int start, int length)
throws SAXException {
- if (discardLevel == 0) {
+ if (bodyLevel > 0 && discardLevel == 0) {
super.ignorableWhitespace(ch, start, length);
}
}
Modified:
lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlParser.java
URL:
http://svn.apache.org/viewvc/lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlParser.java?rev=825821&r1=825820&r2=825821&view=diff
==============================================================================
---
lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlParser.java
(original)
+++
lucene/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlParser.java
Fri Oct 16 09:32:02 2009
@@ -25,13 +25,7 @@
import org.apache.tika.io.CloseShieldInputStream;
import org.apache.tika.metadata.Metadata;
import org.apache.tika.parser.Parser;
-import org.apache.tika.sax.TeeContentHandler;
-import org.apache.tika.sax.WriteOutContentHandler;
import org.apache.tika.sax.XHTMLContentHandler;
-import org.apache.tika.sax.xpath.Matcher;
-import org.apache.tika.sax.xpath.MatchingContentHandler;
-import org.apache.tika.sax.xpath.XPathParser;
-import org.xml.sax.Attributes;
import org.xml.sax.ContentHandler;
import org.xml.sax.InputSource;
import org.xml.sax.SAXException;
@@ -59,15 +53,7 @@
// Prepare the HTML content handler that generates proper
// XHTML events to records relevant document metadata
- XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
- XPathParser xpath = new XPathParser(null, "");
- Matcher body = xpath.parse("/HTML/BODY//node()");
- Matcher title = xpath.parse("/HTML/HEAD/TITLE//node()");
- Matcher meta = xpath.parse("/HTML/HEAD/META//node()");
- handler = new TeeContentHandler(
- new MatchingContentHandler(new BodyHandler(xhtml), body),
- new MatchingContentHandler(getTitleHandler(metadata), title),
- new MatchingContentHandler(getMetaHandler(metadata), meta));
+ handler = new HtmlHandler(handler, metadata);
// Parse the HTML document
org.ccil.cowan.tagsoup.Parser parser =
@@ -86,29 +72,4 @@
parse(stream, handler, metadata, context);
}
- private ContentHandler getTitleHandler(final Metadata metadata) {
- return new WriteOutContentHandler() {
- @Override
- public void endElement(String u, String l, String n) {
- metadata.set(Metadata.TITLE, toString());
- }
- };
- }
-
- private ContentHandler getMetaHandler(final Metadata metadata) {
- return new WriteOutContentHandler() {
- @Override
- public void startElement(
- String uri, String local, String name, Attributes atts)
- throws SAXException {
- if (atts.getValue("http-equiv") != null) {
- metadata.set(atts.getValue("http-equiv"),
atts.getValue("content"));
- }
- if (atts.getValue("name") != null) {
- metadata.set(atts.getValue("name"),
atts.getValue("content"));
- }
- }
- };
- }
-
}