/// <param name="hasSegMarkers">if true, input has segmentation markers</param> /// <param name="hasTags">if true, input has morphological analyses separated by tagDelimiter.</param> /// <param name="hasDomainLabels"> /// if true, input has a whitespace-terminated domain at the beginning /// of each line of text /// </param> /// <param name="stripRewrites"> /// if true, erase orthographical rewrites from the gold labels (for /// comparison purposes) /// </param> /// <param name="tokFactory">a TokenizerFactory for the input</param> public ArabicDocumentReaderAndWriter(bool hasSegMarkers, bool hasTags, bool hasDomainLabels, string domain, bool stripRewrites, ITokenizerFactory <CoreLabel> tokFactory) { tf = tokFactory; inputHasTags = hasTags; inputHasDomainLabels = hasDomainLabels; inputDomain = domain; shouldStripRewrites = stripRewrites; segMarker = hasSegMarkers ? DefaultSegMarker : null; factory = LineIterator.GetFactory(new _ISerializableFunction_131(this)); }
public virtual void Init(SeqClassifierFlags flags) { this.flags = flags; factory = LineIterator.GetFactory(new Sighan2005DocumentReaderAndWriter.CTBDocumentParser(this)); // pichuan : flags.normalizationTable is null --> i believe this is replaced by some java class?? // (Thu Apr 24 11:10:42 2008) cdtos = new ChineseDocumentToSentenceProcessor(flags.normalizationTable); if (flags.dictionary != null) { string[] dicts = flags.dictionary.Split(","); cdict = new ChineseDictionary(dicts, cdtos, flags.expandMidDot); } if (flags.serializedDictionary != null) { string dict = flags.serializedDictionary; cdict = new ChineseDictionary(dict, cdtos, flags.expandMidDot); } if (flags.dictionary2 != null) { string[] dicts2 = flags.dictionary2.Split(","); cdict2 = new ChineseDictionary(dicts2, cdtos, flags.expandMidDot); } }
public virtual void Init(SeqClassifierFlags flags) { verboseForTrueCasing = flags.verboseForTrueCasing; factory = LineIterator.GetFactory(new TrueCasingForNISTDocumentReaderAndWriter.LineToTrueCasesParser()); }