/** Uses a DiskTreebank with a CHTBTokenizer and a BobChrisTreeNormalizer. */
@Override
public DiskTreebank diskTreebank() {
String encoding = inputEncoding;
if (!java.nio.charset.Charset.isSupported(encoding)) {
printlnErr("Warning: desired encoding " + encoding + " not accepted. ");
printlnErr("Using UTF-8 to construct DiskTreebank");
encoding = "UTF-8";
}
return new DiskTreebank(treeReaderFactory(), encoding);
}
/** Uses a MemoryTreebank with a CHTBTokenizer and a BobChrisTreeNormalizer */
@Override
public MemoryTreebank memoryTreebank() {
String encoding = inputEncoding;
if (!java.nio.charset.Charset.isSupported(encoding)) {
System.out.println("Warning: desired encoding " + encoding + " not accepted. ");
System.out.println("Using UTF-8 to construct MemoryTreebank");
encoding = "UTF-8";
}
return new MemoryTreebank(treeReaderFactory(), encoding);
}
/**
* The PrintWriter used to print output. It's the responsibility of pw to deal properly with
* character encodings for the relevant treebank.
*/
public PrintWriter pw(OutputStream o) {
String encoding = outputEncoding;
if (!java.nio.charset.Charset.isSupported(encoding)) {
System.out.println("Warning: desired encoding " + encoding + " not accepted. ");
System.out.println("Using UTF-8 to construct PrintWriter");
encoding = "UTF-8";
}
try {
return new PrintWriter(new OutputStreamWriter(o, encoding), true);
} catch (UnsupportedEncodingException e) {
System.out.println("Warning: desired encoding " + outputEncoding + " not accepted. " + e);
try {
return new PrintWriter(new OutputStreamWriter(o, "UTF-8"), true);
} catch (UnsupportedEncodingException e1) {
System.out.println(
"Something is really wrong. Your system doesn't even support UTF-8!" + e1);
return new PrintWriter(o, true);
}
}
}