public void InteractiveTranslate(IReadOnlyList <string> segment, out WordGraph smtWordGraph, out TranslationResult ruleResult)
{
Load();
smtWordGraph = _smtEngine.GetWordGraph(segment);
ruleResult = _ruleEngine?.Translate(segment);
_lastUsedTime = DateTime.Now;
}
private static WordGraphDto CreateDto(WordGraph wordGraph, IReadOnlyList <string> sourceSegment)
{
return(new WordGraphDto
{
InitialStateScore = (float)wordGraph.InitialStateScore,
FinalStates = wordGraph.FinalStates.ToArray(),
Arcs = wordGraph.Arcs.Select(a => CreateDto(a, sourceSegment)).ToArray()
});
}
public void GetWordGraph_EmptySegment_ReturnsEmptyGraph()
{
using (var smtModel = new ThotSmtModel(TestHelpers.ToyCorpusConfigFileName))
using (ISmtEngine engine = smtModel.CreateEngine())
{
WordGraph wordGraph = engine.GetWordGraph(new string[0]);
Assert.That(wordGraph.IsEmpty, Is.True);
}
}
public WordGraph GetOrCreateNode(char character)
{
if (!Nodes.TryGetValue(character, out WordGraph wordGraph))
{
Nodes[character] = wordGraph = new WordGraph(Length - 1, character);
}
return(wordGraph);
}
private static void Initialize(string dictionaryFilename)
{
Console.WriteLine("Parsing dictionary...");
var allWords = GetAllWords(dictionaryFilename);
Console.WriteLine("Building word graph (this may take a few moments)...");
var wordGraph = new WordGraph(allWords);
Transformer = new WordTransformer(wordGraph);
}
public ThotInteractiveTranslationSession(ThotSmtEngine engine, int n, IReadOnlyList <string> sourceSegment,
WordGraph wordGraph)
{
_engine = engine;
_sourceSegment = sourceSegment;
_n = n;
_prefix = new List <string>();
_isLastWordComplete = true;
_wordGraphProcessor = new ErrorCorrectionWordGraphProcessor(_engine.ErrorCorrectionModel, _sourceSegment,
wordGraph);
UpdateInteractiveResults();
}
public static void Main(string[] args)
{
DebugForm debugWindow = new DebugForm();
WordGraph graph = new WordGraph();
ImportXml.XmlToWordGraph("vocals.xml", out graph);
List <Vocable> vocs = graph.getContent();
foreach (Vocable v in vocs)
{
System.Console.WriteLine(v);
}
debugWindow.ShowItems(vocs);
Application.EnableVisualStyles();
Application.Run(debugWindow);
}
/// <summary>
/// 填充所有字符轮廓
/// </summary>
/// <param name="x">起始X</param>
/// <param name="y">Y值</param>
/// <param name="spacing">字符间隔</param>
private void FillWordOutlines(int x, int y, int spacing, Font font)
{
string text = tbWords.Text;
if (!string.IsNullOrEmpty(text))
{
foreach (char t in text)
{
TreeNode wordNode = new TreeNode(t.ToString());
tvList.Nodes.Add(wordNode);
//获取字符编码
uint ch = GetGB2312Coding(t);
//获取轮廓数据
DOutline outline = WordGraph.GetOutline(ch, font);
//构建轮廓实例
WordOutlineDrawing word = BuildWordOutline(outline, new PointF(x, y), wordNode);
wordOutlines.Add(word);
//下个字的起始位置＝当前起始位置＋宽度＋间隔
x += (int)outline.Width + spacing;
}
}
}
public HybridInteractiveTranslationResult(WordGraph smtWordGraph, TranslationResult ruleResult)
{
SmtWordGraph = smtWordGraph;
RuleResult = ruleResult;
}
public static bool XmlToWordGraph(string pathToXml, out WordGraph graph)
{
string exeLoc = Path.GetDirectoryName(Assembly.GetExecutingAssembly().Location);
try {
XmlTextReader reader = new XmlTextReader(Path.Combine(exeLoc, "..\\..\\xml\\vocals.xml"));
graph = new WordGraph();
Vocable voc = default (Vocable);
Alphabet type = default (Alphabet);
while (reader.Read()) {
if (reader.NodeType == XmlNodeType.Element) {
switch (reader.Name) {
case "Vocal":
voc = new Vocable(null);
continue;
case "Romaji":
if (reader.Read())
if (reader.NodeType == XmlNodeType.Text && reader.Value.Trim() != "")
voc.Romaji = reader.Value.Trim();
continue;
case "Type":
if (reader.Read())
if (reader.NodeType == XmlNodeType.Text && reader.Value.Trim() != "") {
type = (Alphabet)int.Parse(reader.Value.Trim());
}
continue;
case "Japanese":
if (reader.Read())
if (reader.NodeType == XmlNodeType.Text && reader.Value.Trim() != "")
foreach (char s in reader.Value.Trim())
voc.Word.Add(new Syllable(""+s, type));
continue;
case "Kanji":
if (reader.Read())
if (reader.NodeType == XmlNodeType.Text && reader.Value.Trim() != "")
foreach (string obj in reader.Value.Split('\n'))
if (!obj.Trim().Equals(""))
voc.Kanji.AddLast(obj.Trim());
continue;
case "English":
if (reader.Read())
if (reader.NodeType == XmlNodeType.Text && reader.Value.Trim() != "")
foreach (string obj in reader.Value.Split('\n'))
if (!obj.Trim().Equals(""))
voc.EnWords.AddLast(obj.Trim());
continue;
case "German":
if (reader.Read())
if (reader.NodeType == XmlNodeType.Text && reader.Value.Trim() != "")
foreach (string obj in reader.Value.Split('\n'))
if (!obj.Trim().Equals(""))
voc.DeWords.AddLast(obj.Trim());
continue;
default:
continue;
}
}
if (reader.NodeType == XmlNodeType.EndElement && reader.Name == "Vocal")
graph.Add(voc);
}
} catch (FileNotFoundException ex) {
Console.WriteLine(ex.Message);
graph = null;
return false;
}
return true;
}
public Result <WordGraph.T> BuildWordGraph(string s)
{
return(Try(() => WordGraph.Build(s, dict.Root)));
}
public WordGraphPath(WordGraph firstWordGraph, WordGraph secondWordGraph)
{
_firstWordGraph = firstWordGraph;
_secondWordGraph = secondWordGraph;
_stringBuilder = new StringBuilder();
}
public static List <string> GraphCompute(string[] words)
{
// Requested length = 8? We'll only keep of words of length 2-6
var wordGraphsPerLength = new WordGraph[requestedLength - 3];
for (int i = 0; i < wordGraphsPerLength.Length; i++)
{
var length = i + 2;
wordGraphsPerLength[i] = new WordGraph(length, ' ');
}
WordGraph GetWordGraph(int length)
{
return(wordGraphsPerLength[length - 2]);
}
var wordsWithRequestedLength = new List <string>();
for (int wordIndex = 0; wordIndex < words.Length; wordIndex++)
{
var word = words[wordIndex].ToLower();
var wordLength = word.Length;
if (wordLength == requestedLength)
{
wordsWithRequestedLength.Add(word);
continue;
}
if (wordLength < 2 || wordLength > requestedLength - 2)
{
continue;
}
var wordGraph = GetWordGraph(wordLength);
for (var characterIndex = 0; characterIndex < wordLength; characterIndex++)
{
var character = word[characterIndex];
wordGraph = wordGraph.GetOrCreateNode(character);
}
}
//stopwatch.Reset();
//stopwatch.Start();
var paths = new List <WordGraphPath>(8);
for (int lengthOfFirstWord = 2; lengthOfFirstWord < requestedLength - 1; lengthOfFirstWord++)
{
var lengthOfSecondWord = requestedLength - lengthOfFirstWord;
var firstWordGraph = GetWordGraph(lengthOfFirstWord);
var secondWordGraph = GetWordGraph(lengthOfSecondWord);
var path = new WordGraphPath(firstWordGraph, secondWordGraph);
paths.Add(path);
}
var combinations = new List <string>(512);
for (var w = 0; w < wordsWithRequestedLength.Count; w++)
{
var word = wordsWithRequestedLength[w];
for (var p = 0; p < paths.Count; p++)
{
var path = paths[p];
var wordCombination = path.Walk(word);
if (wordCombination != null)
{
combinations.Add(wordCombination);
}
}
}
return(combinations);
}
public static bool XmlToWordGraph(string pathToXml, out WordGraph graph)
{
string exeLoc = Path.GetDirectoryName(Assembly.GetExecutingAssembly().Location);
try {
XmlTextReader reader = new XmlTextReader(Path.Combine(exeLoc, "..\\..\\xml\\vocals.xml"));
graph = new WordGraph();
Vocable voc = default(Vocable);
Alphabet type = default(Alphabet);
while (reader.Read())
{
if (reader.NodeType == XmlNodeType.Element)
{
switch (reader.Name)
{
case "Vocal":
voc = new Vocable(null);
continue;
case "Romaji":
if (reader.Read())
{
if (reader.NodeType == XmlNodeType.Text && reader.Value.Trim() != "")
{
voc.Romaji = reader.Value.Trim();
}
}
continue;
case "Type":
if (reader.Read())
{
if (reader.NodeType == XmlNodeType.Text && reader.Value.Trim() != "")
{
type = (Alphabet)int.Parse(reader.Value.Trim());
}
}
continue;
case "Japanese":
if (reader.Read())
{
if (reader.NodeType == XmlNodeType.Text && reader.Value.Trim() != "")
{
foreach (char s in reader.Value.Trim())
{
voc.Word.Add(new Syllable("" + s, type));
}
}
}
continue;
case "Kanji":
if (reader.Read())
{
if (reader.NodeType == XmlNodeType.Text && reader.Value.Trim() != "")
{
foreach (string obj in reader.Value.Split('\n'))
{
if (!obj.Trim().Equals(""))
{
voc.Kanji.AddLast(obj.Trim());
}
}
}
}
continue;
case "English":
if (reader.Read())
{
if (reader.NodeType == XmlNodeType.Text && reader.Value.Trim() != "")
{
foreach (string obj in reader.Value.Split('\n'))
{
if (!obj.Trim().Equals(""))
{
voc.EnWords.AddLast(obj.Trim());
}
}
}
}
continue;
case "German":
if (reader.Read())
{
if (reader.NodeType == XmlNodeType.Text && reader.Value.Trim() != "")
{
foreach (string obj in reader.Value.Split('\n'))
{
if (!obj.Trim().Equals(""))
{
voc.DeWords.AddLast(obj.Trim());
}
}
}
}
continue;
default:
continue;
}
}
if (reader.NodeType == XmlNodeType.EndElement && reader.Name == "Vocal")
{
graph.Add(voc);
}
}
} catch (FileNotFoundException ex) {
Console.WriteLine(ex.Message);
graph = null;
return(false);
}
return(true);
}