Examples with CoNLLWord com.hankcs.hanlp.corpus.dependency.CoNll.CoNLLWord used on opensource projects

Example 11 with CoNLLWord

use of com.hankcs.hanlp.corpus.dependency.CoNll.CoNLLWord in project HanLP by hankcs.

the class MaxEntDependencyModelMaker method makeModel.

public static boolean makeModel(String corpusLoadPath, String modelSavePath) throws IOException {
    BufferedWriter bw = new BufferedWriter(new OutputStreamWriter(IOUtil.newOutputStream(modelSavePath)));
    LinkedList<CoNLLSentence> sentenceList = CoNLLLoader.loadSentenceList(corpusLoadPath);
    int id = 1;
    for (CoNLLSentence sentence : sentenceList) {
        System.out.printf("%d / %d...", id++, sentenceList.size());
        String[][] edgeArray = sentence.getEdgeArray();
        CoNLLWord[] word = sentence.getWordArrayWithRoot();
        for (int i = 0; i < word.length; ++i) {
            for (int j = 0; j < word.length; ++j) {
                if (i == j)
                    continue;
                // 这就是一个边的实例，从i出发，到j，当然它可能存在也可能不存在，不存在取null照样是一个实例
                List<String> contextList = new LinkedList<String>();
                // 先生成i和j的原子特征
                contextList.addAll(generateSingleWordContext(word, i, "i"));
                contextList.addAll(generateSingleWordContext(word, j, "j"));
                // 然后生成二元组的特征
                contextList.addAll(generateUniContext(word, i, j));
                // 将特征字符串化
                for (String f : contextList) {
                    bw.write(f);
                    bw.write(' ');
                }
                // 事件名称为依存关系
                bw.write("" + edgeArray[i][j]);
                bw.newLine();
            }
        }
        System.out.println("done.");
    }
    bw.close();
    return true;
}

Example 12 with CoNLLWord

use of com.hankcs.hanlp.corpus.dependency.CoNll.CoNLLWord in project HanLP by hankcs.

the class WordNatureWeightModelMaker method makeModel.

public static boolean makeModel(String corpusLoadPath, String modelSavePath) {
    Set<String> posSet = new TreeSet<String>();
    DictionaryMaker dictionaryMaker = new DictionaryMaker();
    for (CoNLLSentence sentence : CoNLLLoader.loadSentenceList(corpusLoadPath)) {
        for (CoNLLWord word : sentence.word) {
            addPair(word.NAME, word.HEAD.NAME, word.DEPREL, dictionaryMaker);
            addPair(word.NAME, wrapTag(word.HEAD.POSTAG), word.DEPREL, dictionaryMaker);
            addPair(wrapTag(word.POSTAG), word.HEAD.NAME, word.DEPREL, dictionaryMaker);
            addPair(wrapTag(word.POSTAG), wrapTag(word.HEAD.POSTAG), word.DEPREL, dictionaryMaker);
            posSet.add(word.POSTAG);
        }
    }
    for (CoNLLSentence sentence : CoNLLLoader.loadSentenceList(corpusLoadPath)) {
        for (CoNLLWord word : sentence.word) {
            addPair(word.NAME, word.HEAD.NAME, word.DEPREL, dictionaryMaker);
            addPair(word.NAME, wrapTag(word.HEAD.POSTAG), word.DEPREL, dictionaryMaker);
            addPair(wrapTag(word.POSTAG), word.HEAD.NAME, word.DEPREL, dictionaryMaker);
            addPair(wrapTag(word.POSTAG), wrapTag(word.HEAD.POSTAG), word.DEPREL, dictionaryMaker);
            posSet.add(word.POSTAG);
        }
    }
    StringBuilder sb = new StringBuilder();
    for (String pos : posSet) {
        sb.append("case \"" + pos + "\":\n");
    }
    IOUtil.saveTxt("data/model/dependency/pos-thu.txt", sb.toString());
    return dictionaryMaker.saveTxtTo(modelSavePath);
}