Examples with Term com.hankcs.hanlp.seg.common.Term used on opensource projects

Example 41 with Term

use of com.hankcs.hanlp.seg.common.Term in project HanLP by hankcs.

the class CRFSegment method toVertexList.

private static List<Vertex> toVertexList(List<Term> termList, boolean appendStart) {
    ArrayList<Vertex> vertexList = new ArrayList<Vertex>(termList.size() + 1);
    if (appendStart)
        vertexList.add(Vertex.B);
    for (Term term : termList) {
        CoreDictionary.Attribute attribute = CoreDictionary.get(term.word);
        if (attribute == null) {
            if (term.word.trim().length() == 0)
                attribute = new CoreDictionary.Attribute(Nature.x);
            else
                attribute = new CoreDictionary.Attribute(Nature.nz);
        } else
            term.nature = attribute.nature[0];
        Vertex vertex = new Vertex(term.word, attribute);
        vertexList.add(vertex);
    }
    return vertexList;
}

Example 42 with Term

use of com.hankcs.hanlp.seg.common.Term in project HanLP by hankcs.

the class CRFSegment method segSentence.

@Override
protected List<Term> segSentence(char[] sentence) {
    if (sentence.length == 0)
        return Collections.emptyList();
    char[] sentenceConverted = CharTable.convert(sentence);
    Table table = new Table();
    table.v = atomSegmentToTable(sentenceConverted);
    crfModel.tag(table);
    List<Term> termList = new LinkedList<Term>();
    if (HanLP.Config.DEBUG) {
        System.out.println("CRF标注结果");
        System.out.println(table);
    }
    int offset = 0;
    OUTER: for (int i = 0; i < table.v.length; offset += table.v[i][1].length(), ++i) {
        String[] line = table.v[i];
        switch(line[2].charAt(0)) {
            case 'B':
                {
                    int begin = offset;
                    while (table.v[i][2].charAt(0) != 'E') {
                        offset += table.v[i][1].length();
                        ++i;
                        if (i == table.v.length) {
                            break;
                        }
                    }
                    if (i == table.v.length) {
                        termList.add(new Term(new String(sentence, begin, offset - begin), null));
                        break OUTER;
                    } else
                        termList.add(new Term(new String(sentence, begin, offset - begin + table.v[i][1].length()), null));
                }
                break;
            default:
                {
                    termList.add(new Term(new String(sentence, offset, table.v[i][1].length()), null));
                }
                break;
        }
    }
    if (config.speechTagging) {
        List<Vertex> vertexList = toVertexList(termList, true);
        Viterbi.compute(vertexList, CoreDictionaryTransformMatrixDictionary.transformMatrixDictionary);
        int i = 0;
        for (Term term : termList) {
            if (term.nature != null)
                term.nature = vertexList.get(i + 1).guessNature();
            ++i;
        }
    }
    if (config.useCustomDictionary) {
        List<Vertex> vertexList = toVertexList(termList, false);
        combineByCustomDictionary(vertexList);
        termList = toTermList(vertexList, config.offset);
    }
    return termList;
}

Example 43 with Term

use of com.hankcs.hanlp.seg.common.Term in project HanLP by hankcs.

the class DemoPlaceRecognition method main.

public static void main(String[] args) {
    String[] testCase = new String[] { "蓝翔给宁夏固原市彭阳县红河镇黑牛沟村捐赠了挖掘机" };
    Segment segment = HanLP.newSegment().enablePlaceRecognize(true);
    for (String sentence : testCase) {
        List<Term> termList = segment.seg(sentence);
        System.out.println(termList);
    }
}

Example 44 with Term

use of com.hankcs.hanlp.seg.common.Term in project HanLP by hankcs.

the class DemoStopWord method main.

public static void main(String[] args) {
    String text = "小区居民有的反对喂养流浪猫，而有的居民却赞成喂养这些小宝贝";
    // 可以动态修改停用词词典
    CoreStopWordDictionary.add("居民");
    System.out.println(NotionalTokenizer.segment(text));
    CoreStopWordDictionary.remove("居民");
    System.out.println(NotionalTokenizer.segment(text));
    // 可以对任意分词器的结果执行过滤
    List<Term> termList = BasicTokenizer.segment(text);
    System.out.println(termList);
    CoreStopWordDictionary.apply(termList);
    System.out.println(termList);
    // 还可以自定义过滤逻辑
    CoreStopWordDictionary.FILTER = new Filter() {

        @Override
        public boolean shouldInclude(Term term) {
            switch(term.nature) {
                case nz:
                    return !CoreStopWordDictionary.contains(term.word);
            }
            return false;
        }
    };
    System.out.println(NotionalTokenizer.segment(text));
}

Example 45 with Term

use of com.hankcs.hanlp.seg.common.Term in project HanLP by hankcs.

the class DemoTranslatedNameRecognition method main.

public static void main(String[] args) {
    String[] testCase = new String[] { "一桶冰水当头倒下，微软的比尔盖茨、Facebook的扎克伯格跟桑德博格、亚马逊的贝索斯、苹果的库克全都不惜湿身入镜，这些硅谷的科技人，飞蛾扑火似地牺牲演出，其实全为了慈善。", "世界上最长的姓名是简森·乔伊·亚历山大·比基·卡利斯勒·达夫·埃利奥特·福克斯·伊维鲁莫·马尔尼·梅尔斯·帕特森·汤普森·华莱士·普雷斯顿。" };
    Segment segment = HanLP.newSegment().enableTranslatedNameRecognize(true);
    for (String sentence : testCase) {
        List<Term> termList = segment.seg(sentence);
        System.out.println(termList);
    }
}