Reputation: 1827
I'm new in TIKA. I try to convert Microsoft word documents to HTML by using Tika. I'm using TikaOnDotNet wrapper to used TIKA on .Net framework. My conversion code is like following:
byte[] file = Files.toByteArray(new File(@"myPath\document.doc"));
AutoDetectParser tikaParser = new AutoDetectParser();
ByteArrayOutputStream output = new ByteArrayOutputStream();
SAXTransformerFactory factory = (SAXTransformerFactory)TransformerFactory.newInstance();
TransformerHandler handler = factory.newTransformerHandler();
handler.getTransformer().setOutputProperty(OutputKeys.METHOD, "html");
handler.getTransformer().setOutputProperty(OutputKeys.INDENT, "yes");
handler.getTransformer().setOutputProperty(OutputKeys.ENCODING, "UTF-8");
handler.setResult(new StreamResult(output));
ExpandedTitleContentHandler handler1 = new ExpandedTitleContentHandler(handler);
tikaParser.parse(new ByteArrayInputStream(file), handler1, new Metadata());
File ofile = new File(@"C:\toHtml\text.html");
ofile.createNewFile();
DataOutputStream stream = new DataOutputStream(new FileOutputStream(ofile));
output.writeTo(stream);
everything working well except the embedded images. The generated HTML contains image tag like:
<img src="embedded:image2.wmf" alt="image2.wmf"/>
but the image source does not exists. Please advise me
Upvotes: 3
Views: 2842
Reputation: 1827
Credits goes to @Gagravarr.
please note that this is a simple implementation of code, the original codes are available in comment of the questions.
This implementation is based on TikaOnDotNet wrapper.....
public class DocToHtml
{
private TikaConfig config = TikaConfig.getDefaultConfig();
public void Convert()
{
byte[] file = Files.toByteArray(new File(@"filename.doc"));
AutoDetectParser tikaParser = new AutoDetectParser();
ByteArrayOutputStream output = new ByteArrayOutputStream();
SAXTransformerFactory factory = (SAXTransformerFactory)TransformerFactory.newInstance();
var inputStream = new ByteArrayInputStream(file);
// ToHTMLContentHandler handler = new ToHTMLContentHandler();
var metaData = new Metadata();
EncodingDetector encodingDetector = new UniversalEncodingDetector();
var encode = encodingDetector.detect(inputStream, metaData) ?? new UTF_32();
TransformerHandler handler = factory.newTransformerHandler();
handler.getTransformer().setOutputProperty(OutputKeys.METHOD, "html");
handler.getTransformer().setOutputProperty(OutputKeys.INDENT, "yes");
handler.getTransformer().setOutputProperty(OutputKeys.ENCODING, encode.toString());
handler.setResult(new StreamResult(output));
ContentHandler imageRewriting = new ImageRewritingContentHandler(handler);
// ExpandedTitleContentHandler handler1 = new ExpandedTitleContentHandler(handler);
ParseContext context = new ParseContext();
context.set(typeof(EmbeddedDocumentExtractor), new FileEmbeddedDocumentEtractor());
tikaParser.parse(inputStream, imageRewriting, new Metadata(), context);
byte[] array = output.toByteArray();
System.IO.File.WriteAllBytes(@"C:\toHtml\text.html", array);
}
private class ImageRewritingContentHandler : ContentHandlerDecorator
{
public ImageRewritingContentHandler(ContentHandler handler) : base(handler)
{
}
public override void startElement(string uri, string localName, string name, Attributes origAttrs)
{
if ("img".Equals(localName))
{
AttributesImpl attrs;
if (origAttrs is AttributesImpl)
attrs = (AttributesImpl)origAttrs;
else
attrs = new AttributesImpl(origAttrs);
for (int i = 0; i < attrs.getLength(); i++)
{
if ("src".Equals(attrs.getLocalName(i)))
{
String src = attrs.getValue(i);
if (src.StartsWith("embedded:"))
{
var newSrc = src.Replace("embedded:", @"images\");
attrs.setValue(i, newSrc);
}
}
}
attrs.addAttribute(null, "width", "width","width", "100px");
base.startElement(uri, localName, name, attrs);
}
else
base.startElement(uri, localName, name, origAttrs);
}
}
private class FileEmbeddedDocumentEtractor : EmbeddedDocumentExtractor
{
private int count = 0;
public bool shouldParseEmbedded(Metadata m)
{
return true;
}
public void parseEmbedded(InputStream inputStream, ContentHandler contentHandler, Metadata metadata, bool outputHtml)
{
Detector detector = new DefaultDetector();
string name = metadata.get("resourceName");
MediaType contentType = detector.detect(inputStream, metadata);
if (contentType.getType() != "image") return;
var embeddedFile = name;
File outputFile = new File(@"C:\toHtml\images", embeddedFile);
try
{
using (FileOutputStream os = new FileOutputStream(outputFile))
{
var tin = inputStream as TikaInputStream;
if (tin != null)
{
if (tin.getOpenContainer() != null && tin.getOpenContainer() is DirectoryEntry)
{
POIFSFileSystem fs = new POIFSFileSystem();
fs.writeFilesystem(os);
}
else
{
IOUtils.copy(inputStream, os);
}
}
}
}
catch (Exception ex)
{
throw;
}
}
}
}
Upvotes: 2