Basically, i parse several xml output from Apache Tika to get metadata (via meta tags) and list of embedded files using <div class="embedded" id="content">
. However, i found my map had several key Unknown tag (0x...)
. I wonder if it caused by Tika's incomplete tag output because the error i get only related to unclosed tag - which i suspect within the body of XML instead of the output i want (meta, div). However, it is rather illogical where the only code that writes into the map are meta tags and divs (with embedded class) - which is only a small part of the document.
public class Parse {
private class internalXMLReader extends DefaultHandler{
public final Map<String, Object> entityList = new HashMap<>();
@Override
public void startElement(String uri, String localname, String qName, Attributes attributes) throws SAXException{
String key, content;
if(qName.equalsIgnoreCase("meta")){
key = attributes.getValue("name");
content = attributes.getValue("content");
if(key.contains("Content-Type")){
String tmp[] = attributes.getValue("content").replace(' ', '\0').split(";");
if(tmp.length > 1){
content = tmp[0];
}
}
entityList.put(key, content);
}
else if(qName.equalsIgnoreCase("div")){
if((attributes.getValue("class") != null) && (attributes.getValue("class").equalsIgnoreCase("embedded"))){
key = "embedded";
List<String> inlist;
if(entityList.containsKey("embedded") && (entityList.get("embedded") instanceof List)){
inlist = (List) entityList.get(key);
}
else{
inlist = new LinkedList<>();
entityList.put(key, inlist);
}
inlist.add(attributes.getValue("id"));
}
}
}
@Override
public void endElement(String uri, String localname, String qName) throws SAXException{
//no, i just did not want to validate or such..
}
@Override
public void characters(char ch[], int start, int length) throws SAXException{
//no, we don't actually read <something>this</something> yet
}
}
public Entity parse(String xml, Entity in){
try{
InputSource xmlinput = new InputSource(new StringReader(xml));
SAXParserFactory factory = SAXParserFactory.newInstance();
SAXParser parser = factory.newSAXParser();
internalXMLReader handler = new internalXMLReader();
parser.parse(xmlinput, handler);
in.addMeta(handler.entityList);
}
catch(IOException | ParserConfigurationException | SAXException ex){
Logger.getLogger(TikaParseNCluste.class.getName()).log(Level.SEVERE, null, ex);
}
return in;
}
}
Perhaps i should take a look at my 800+ xml files.
Google and javadocs had no information regarding this thing and i'm rather impatient. However runs grep -i -l -r "name=\"unknown" .
i got several jpg files had <meta name="Unknown..." contents="..."/>
perhaps this is why. I don't expect ApacheTika would give such outputs. So, i changed my code to:
...
if(qName.equalsIgnoreCase("meta") && (attributes.getValue("name") != null)){
key = attributes.getValue("name");
if((key != null) && (!key.contains("Unknown"))){
content = attributes.getValue("content");
if(key.contains("Content-Type")){
String tmp[] = attributes.getValue("content").replace(' ', '\0').split(";");
if(tmp.length > 1){
content = tmp[0];
}
}
entityList.put(key, content);
}
}
...
I wonder if it's a bug or something else. So far, quick query to Google Search with keyword apache tika unknown tag only lead me here.