public static LemmatizerModel TrainModel(string path)
{
FileStream fs = new FileStream(path, FileMode.Open, FileAccess.Read);
TrainingParameters trainParams = new TrainingParameters();
trainParams.Set(Parameters.Iterations, "1");
trainParams.Set(Parameters.Cutoff, "0");
LemmatizerFactory lemmatizerFactory = new LemmatizerFactory();
LemmaSampleStream sampleStream = new LemmaSampleStream(new PlainTextByLineStream(fs));
return(LemmatizerME.Train(TRAINING_LANGUAGE, sampleStream, trainParams, lemmatizerFactory));
}
static void Main(string[] args)
{
var rmlPath = System.Environment.GetEnvironmentVariable("RML");
Console.WriteLine("For test LemmatizerNET you need Lemmatizer dictionaries (RML)");
Console.Write("\tRML directory (" + rmlPath + "): ");
if (string.IsNullOrEmpty(rmlPath))
{
var newRmlPath = Console.ReadLine();
if (!string.IsNullOrEmpty(newRmlPath))
{
rmlPath = newRmlPath;
}
}
//Console.Write("Select language 'R'-Russian, 'G'-German, 'E'-English (default - R): ");
var langStr = "R"; // Console.ReadLine().ToUpper(CultureInfo.InvariantCulture);
MorphLanguage lang;
switch (langStr)
{
case "":
case "R":
lang = MorphLanguage.Russian;
break;
case "G":
lang = MorphLanguage.German;
break;
case "E":
lang = MorphLanguage.English;
break;
default:
Console.WriteLine("Wrong selection. Using default language Russian");
lang = MorphLanguage.Russian;
break;
}
ILemmatizer lem = LemmatizerFactory.Create(lang);
string rgt = "";
try
{
StreamReader r = new StreamReader(rmlPath + @"\Dicts\Morph\" + langStr.ToLower() + "gramtab.tab", Encoding.GetEncoding(1251));
rgt = r.ReadToEnd(); r.Close();
}
catch (Exception e)
{
}
try
{
var manager = FileManager.GetFileManager(rmlPath);
lem.LoadDictionariesRegistry(manager);
}
catch (IOException e)
{
Console.WriteLine("Unable to load dictionaries due to the following:\r\n\t");
Console.WriteLine(e.Message);
return;
}
while (true)
{
Console.Write("\nInput word (q - exit): ");
var word = Console.ReadLine().Replace("\"", "").Replace("'", "").Trim();
if (word.Equals("q", StringComparison.InvariantCultureIgnoreCase))
{
return;
}
//Позволяет декодировать грамкоды
if (word.ToLower().Contains("g") || word.Contains("\t") || word.Contains("\a")) //m_gramcodes = 0x0322630c "абажай"
{
string gc = Regex.Match(word, "[А-Яа-яёЁ]+").Groups[0].Value;
string r = "";
for (int i = 0; i < gc.Length / 2; i++)
{
Console.WriteLine(Regex.Match(rgt, "^" + gc.Substring(2 * i, 2) + "[^а-яА-яЕё]*(.*)", RegexOptions.Multiline).Groups[1].Value.Replace("\r", ""));
}
Console.WriteLine("");
continue;
}
var paradigmList = lem.CreateParadigmCollectionFromForm(word, false, false);
if (paradigmList.Count == 0)
{
try //Позволяет декодировать граммемы, если число вместо слова
{
string[] g = Grammems;
if (word.StartsWith("f"))
{
word = word.Substring(1); g = Flags;
} //декодируем флаги
UInt64 gr = Convert.ToUInt64(word);
for (int i = g.Length - 1; i > -1; i--)
{
if (((1uL << i) & gr) > 0)
{
Console.Write(g[i] + ",");
}
}
Console.WriteLine("");
}
catch (Exception)
{
}
Console.WriteLine("Paradigms not found");
continue;
}
string ancodes = "";
for (var i = 0; i < paradigmList.Count; i++)
{
var paradigm = paradigmList[i];
Console.Write("Paradigm: ");
rmlPath = paradigm.SrcAncode;
Console.WriteLine(paradigm.Norm);
int k = paradigm.GetAccent(0);
k = paradigm.SrcAccentedVowel;
Console.Write("\tFounded: ");
Console.WriteLine(paradigm.Founded);
Console.Write("\tParadigmID: ");
Console.WriteLine(paradigm.ParadigmID);
Console.Write("\tAccentModelNo: ");
Console.WriteLine(paradigm.AccentModelNo);
Console.WriteLine("=====");
Console.WriteLine("$type_grm = " + (paradigm.TypeAncode == "??" ? "" : Regex.Match(rgt, "^" + paradigm.TypeAncode + "[^а-яА-яёЁ]*([^\r]*)", RegexOptions.Multiline).Groups[1].Value));
ancodes += paradigm.TypeAncode;
for (var j = 0; j < paradigm.Count; j++)
{
ancodes += paradigm.GetAncode(j);
Console.Write("\t\t");
Console.Write(paradigm.GetAccent(j) == 255 ? paradigm.GetForm(j) : paradigm.GetForm(j).Insert(paradigm.GetAccent(j) + 1, "'"));
Console.Write("\t");
Console.WriteLine(Regex.Match(rgt, "^" + paradigm.GetAncode(j) + "[^а-яА-яЕё]*(.*)", RegexOptions.Multiline).Groups[1].Value.Replace("\r", ""));
}
}
}
}
