### Eclipse Workspace Patch 1.0
#P tika-parsers
Index: src/main/java/org/apache/tika/parser/html/BoilerpipeContentHandler.java
===================================================================
--- src/main/java/org/apache/tika/parser/html/BoilerpipeContentHandler.java (revision 985009)
+++ src/main/java/org/apache/tika/parser/html/BoilerpipeContentHandler.java (working copy)
@@ -21,8 +21,10 @@
import org.apache.tika.metadata.Metadata;
import org.apache.tika.sax.WriteOutContentHandler;
import org.apache.tika.sax.XHTMLContentHandler;
+import org.xml.sax.Attributes;
import org.xml.sax.ContentHandler;
import org.xml.sax.SAXException;
+import org.xml.sax.helpers.AttributesImpl;
import de.l3s.boilerpipe.BoilerpipeExtractor;
import de.l3s.boilerpipe.BoilerpipeProcessingException;
@@ -96,11 +98,28 @@
throw new SAXException(e);
}
+ Attributes emptyAttrs = new AttributesImpl();
+
delegate.startDocument();
+ delegate.startPrefixMapping("", XHTMLContentHandler.XHTML);
+
+ delegate.startElement(XHTMLContentHandler.XHTML, "html", "html", emptyAttrs);
+ delegate.startElement(XHTMLContentHandler.XHTML, "head", "head", emptyAttrs);
+ delegate.startElement(XHTMLContentHandler.XHTML, "title", "title", emptyAttrs);
+
+ if (td.getTitle() != null) {
+ char[] titleChars = td.getTitle().toCharArray();
+ delegate.characters(titleChars, 0, titleChars.length);
+ }
+
+ delegate.endElement(XHTMLContentHandler.XHTML, "title", "title");
+ delegate.endElement(XHTMLContentHandler.XHTML, "head", "head");
+ delegate.startElement(XHTMLContentHandler.XHTML, "body", "body", emptyAttrs);
+
for (TextBlock block : td.getTextBlocks()) {
if (block.isContent()) {
- delegate.startElement(XHTMLContentHandler.XHTML, "p", "p", null);
+ delegate.startElement(XHTMLContentHandler.XHTML, "p", "p", emptyAttrs);
char[] chars = block.getText().toCharArray();
delegate.characters(chars, 0, chars.length);
delegate.endElement(XHTMLContentHandler.XHTML, "p", "p");
@@ -108,6 +127,11 @@
}
}
+ delegate.endElement(XHTMLContentHandler.XHTML, "body", "body");
+ delegate.endElement(XHTMLContentHandler.XHTML, "html", "html");
+
+ delegate.endPrefixMapping("");
+
delegate.endDocument();
}
}
Index: src/test/java/org/apache/tika/parser/html/HtmlParserTest.java
===================================================================
--- src/test/java/org/apache/tika/parser/html/HtmlParserTest.java (revision 986089)
+++ src/test/java/org/apache/tika/parser/html/HtmlParserTest.java (working copy)
@@ -555,6 +555,31 @@
}
/**
+ * Test case for TIKA-480: fix NPE when using BodyContentHandler or HtmlTransformer
+ * as delegate for BoilerpipeContentHandler
+ * @see TIKA-480
+ */
+ public void testBoilerplateDelegation() throws Exception {
+ String path = "/test-documents/boilerplate.html";
+
+ Metadata metadata = new Metadata();
+ StringWriter sw = new StringWriter();
+ new HtmlParser().parse(
+ HtmlParserTest.class.getResourceAsStream(path),
+ makeHtmlTransformer(sw), metadata, new ParseContext());
+
+ String content = sw.toString();
+
+ // Should have , , , <body> elements
+ assertTrue(Pattern.matches("(?s).*<html xmlns=\"http://www.w3.org/1999/xhtml\">.*</html>.*$", content));
+ assertTrue(Pattern.matches("(?s).*<head>.*</head>.*$", content));
+ assertTrue(Pattern.matches("(?s).*<title>Title.*$", content));
+ assertTrue(Pattern.matches("(?s).*.*.*$", content));
+ }
+
+
+
+ /**
* Create ContentHandler that transforms SAX events into textual HTML output,
* and writes it out to - typically this is a StringWriter.
*