private Tree TransformTree(Tree tree, bool isRoot)
{
string label = tree.Label().Value();
// log.info("ChineseCollinizer: Node label is " + label);
if (tree.IsLeaf())
{
if (deletePunct && ctlp.IsPunctuationWord(label))
{
return(null);
}
else
{
return(tf.NewLeaf(new StringLabel(label)));
}
}
if (tree.IsPreTerminal() && deletePunct && ctlp.IsPunctuationTag(label))
{
// System.out.println("Deleting punctuation");
return(null);
}
IList <Tree> children = new List <Tree>();
if (label.Matches("ROOT.*") && tree.NumChildren() == 1)
{
// keep non-unary roots for now
return(TransformTree(tree.Children()[0], true));
}
//System.out.println("Enhanced label is " + label);
// remove all functional and machine-generated annotations
label = label.ReplaceFirst("[^A-Z].*$", string.Empty);
// merge parentheticals with adverb phrases
label = label.ReplaceFirst("PRN", "ADVP");
//System.out.println("New label is " + label);
for (int cNum = 0; cNum < tree.Children().Length; cNum++)
{
Tree child = tree.Children()[cNum];
Tree newChild = TransformTree(child, false);
if (newChild != null)
{
children.Add(newChild);
}
}
// We don't delete the root because there are trees in the
// Chinese treebank that only have punctuation in them!!!
if (children.IsEmpty() && !isRoot)
{
return(null);
}
return(tf.NewTreeNode(new StringLabel(label), children));
}
/// <summary>
/// The main method reads (segmented, whitespace delimited) words from a file
/// and prints them with their English translation(s).
/// </summary>
/// <remarks>
/// The main method reads (segmented, whitespace delimited) words from a file
/// and prints them with their English translation(s).
/// The path and filename of the CEDict Lexicon can be supplied via the
/// "-dictPath" flag; otherwise the default filename "cedict_ts.u8" in the
/// current directory is checked.
/// By default, only the first translation is printed. If the "-all" flag
/// is given, all translations are printed.
/// The input and output encoding can be specified using the "-encoding" flag.
/// Otherwise UTF-8 is assumed.
/// </remarks>
/// <exception cref="System.IO.IOException"/>
public static void Main(string[] args)
{
IDictionary <string, int> flagsToNumArgs = Generics.NewHashMap();
flagsToNumArgs["-dictPath"] = 1;
flagsToNumArgs["-encoding"] = 1;
IDictionary <string, string[]> argMap = StringUtils.ArgsToMap(args, flagsToNumArgs);
string[] otherArgs = argMap[null];
if (otherArgs.Length < 1)
{
log.Info("usage: ChineseEnglishWordMap [-all] [-dictPath path] [-encoding enc_string] inputFile");
System.Environment.Exit(1);
}
string filename = otherArgs[0];
bool allTranslations = argMap.Contains("-all");
string charset = defaultCharset;
if (argMap.Contains("-encoding"))
{
charset = argMap["-encoding"][0];
}
BufferedReader r = new BufferedReader(new InputStreamReader(new FileInputStream(filename), charset));
ITreebankLanguagePack tlp = new ChineseTreebankLanguagePack();
string[] dpString = argMap["-dictPath"];
ChineseEnglishWordMap cewm = (dpString == null) ? new ChineseEnglishWordMap() : new ChineseEnglishWordMap(dpString[0]);
int totalWords = 0;
int coveredWords = 0;
PrintWriter pw = new PrintWriter(new OutputStreamWriter(System.Console.Out, charset), true);
for (string line = r.ReadLine(); line != null; line = r.ReadLine())
{
string[] words = line.Split("\\s", 1000);
foreach (string word in words)
{
totalWords++;
if (word.Length == 0)
{
continue;
}
pw.Print(StringUtils.Pad(word + ':', 8));
if (tlp.IsPunctuationWord(word))
{
totalWords--;
pw.Print(word);
}
else
{
if (IsDigits(word))
{
pw.Print(word + " [NUMBER]");
}
else
{
if (cewm.ContainsKey(word))
{
coveredWords++;
if (allTranslations)
{
IList <string> trans = new List <string>(cewm.GetAllTranslations(word));
foreach (string s in trans)
{
pw.Print((trans.IndexOf(s) > 0 ? "|" : string.Empty) + s);
}
}
else
{
pw.Print(cewm.GetFirstTranslation(word));
}
}
else
{
pw.Print("[UNK]");
}
}
}
pw.Println();
}
pw.Println();
}
r.Close();
log.Info("Finished translating " + totalWords + " words (");
log.Info(coveredWords + " were in dictionary).");
}