Package org.ictclas4j.bean

Examples of org.ictclas4j.bean.Atom

41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
      result = new ArrayList<Atom>();
      String[] ss = GFString.atomSplit(str);

      int index = str.indexOf(Utility.SENTENCE_BEGIN);
      if (index == 0) {
        Atom atom = new Atom();
        atom.setWord(Utility.SENTENCE_BEGIN);
        atom.setLen(Utility.SENTENCE_BEGIN.length());
        atom.setPos(Utility.CT_SENTENCE_BEGIN);
        result.add(atom);
        index += Utility.SENTENCE_BEGIN.length();
      }

      if (index == -1)
        index = 0;
      for (int i = index; i < ss.length; i++) {
        if (Utility.SENTENCE_END.equals(str.substring(i))) {
          Atom atom = new Atom();
          atom.setWord(Utility.SENTENCE_END);
          atom.setLen(Utility.SENTENCE_END.length());
          atom.setPos(Utility.CT_SENTENCE_END);
          result.add(atom);
          break;
        }

        String s = ss[i];
        sAtom += s;
        int curType = Utility.charType(s);
        if (".".equals(s)
            && (i + 1 < ss.length && (Utility.charType(ss[i + 1]) == Utility.CT_NUM || GFString
                .isNumeric(ss[i+1]))))
          curType = Utility.CT_NUM;

        // ����Ǻ��֡��ָ�����
        if (curType == Utility.CT_CHINESE || curType == Utility.CT_INDEX || curType == Utility.CT_DELIMITER
            || curType == Utility.CT_OTHER) {

          Atom atom = new Atom();
          atom.setWord(s);
          atom.setLen(s.length());
          atom.setPos(curType);
          result.add(atom);
          sAtom = "";
        }
        // ��������֡���ĸ�����ֽڷ��ţ�������ڵ���Щ��Ϊһ��ԭ�ӡ����磺����SHX-123�ͺŵ��ֻ��������е�SHX-123����һ��ԭ��
        else {
          int nextType = 255;// ��һ���ַ�������
          if (i < ss.length - 1)
            nextType = Utility.charType(ss[i + 1]);
          if (nextType != curType || i == ss.length - 1) {
            Atom atom = new Atom();
            atom.setWord(sAtom);
            atom.setLen(sAtom.length());
            atom.setPos(curType);
            result.add(atom);
            sAtom = "";
          }
        }
      }
View Full Code Here
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
   * @return
   */
  public static SegGraph generate(ArrayList<Atom> atoms, DictLib dictLib) {
    SegGraph segGraph = null;
    SegNode sn = null;
    Atom atom = null;

    if (atoms != null && atoms.size() > 0 && dictLib != null) {
      segGraph = new SegGraph();
      Dictionary dict = dictLib.getCoreDict();

      // �ȰѷǺ����ַ��Ĵ���ʶ�����
      for (int i = 0; i < atoms.size(); i++) {
        atom = atoms.get(i);
        String word = atom.getWord();
        if (atom.getPos() == Utility.CT_CHINESE)
          sn = new SegNode(i, i + 1, 0, 0, atom.getWord());
        else {
          int pos = 0;
          double value = Utility.MAX_FREQUENCE;

          switch (atom.getPos()) {
          case Utility.CT_INDEX:
          case Utility.CT_NUM:
            pos = -POSTag.NUM;// 'm'*256
            word = Utility.UNKNOWN_NUM;
            value = 0;
            break;
          case Utility.CT_DELIMITER:
            pos = POSTag.PUNC;// 'w'*256;
            break;
          case Utility.CT_LETTER:
            pos = -POSTag.NOUN_LETTER;//
            value = 0;
            word = Utility.UNKNOWN_LETTER;
            break;
          case Utility.CT_SINGLE:// 12021-2129-3121
            if (Utility.getCharCount("+-1234567890", atom.getWord()) == atom.getLen()) {
              pos = -POSTag.NUM;// 'm'*256
              word = Utility.UNKNOWN_NUM;
            } else {
              pos = -POSTag.NOUN_LETTER;//
              word = Utility.UNKNOWN_LETTER;
            }
            value = 0;
            break;
          default:
            pos = atom.getPos();// '?'*256;
            break;
          }

          int gbkID = dictLib.getGBKID(word);
          sn = new SegNode(i, i + 1, pos, value, word);
          sn.setGbkID(gbkID);
        }

        sn.setSrcWord(atom.getWord());
        segGraph.insert(sn, true);
      }

      StringBuffer words = new StringBuffer();
      for (int i = 0; i < atoms.size(); i++) {
        int j = i + 1;
        words.delete(0, words.length());
        words.append(atoms.get(i).getWord());

        // ����ǡ��·ݡ�����Ҫ�ָ�
        boolean flag = false;
        if (j < atoms.size()) {
          Atom a2 = atoms.get(j);
          if ("��".equals(words.toString()) && "��".equals(a2.getWord())) {
            segGraph.delete(i, j);
            segGraph.delete(i + 1, j + 1);
            words.append(a2.getWord());
            flag = true;
            j++;
          }
        }

View Full Code Here
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
   * @return
   */
  public static SegGraph generate(ArrayList<Atom> atoms,Dictionary dict) {
    SegGraph segGraph = null;
    SegNode sn = null;
    Atom atom = null;

    if (atoms != null && atoms.size() > 0 && dict != null) {
      segGraph = new SegGraph();
      for (int i = 0; i < atoms.size(); i++) {
        atom = atoms.get(i);
        String word = atom.getWord();
        if (atom.getPos() == Utility.CT_CHINESE)
          sn = new SegNode(i, i + 1, 0,0, atom.getWord());
        else {
          double value = Utility.MAX_FREQUENCE;
          int pos = 0;

          switch (atom.getPos()) {
          case Utility.CT_INDEX:
          case Utility.CT_NUM:
            pos = -POSTag.NUM;// 'm'*256
            word = Utility.UNKNOWN_NUM;
            value = 0;
            break;
          case Utility.CT_DELIMITER:
            pos = POSTag.PUNC;// 'w'*256;
            break;
          case Utility.CT_LETTER:
            pos = -POSTag.NOUN_LETTER;//
            value = 0;
            word = Utility.UNKNOWN_LETTER;
            break;
          case Utility.CT_SINGLE:// 12021-2129-3121
            if (Utility.getCharCount("+-1234567890", atom.getWord()) == atom.getLen()) {
              pos = -POSTag.NUM;// 'm'*256
              word = Utility.UNKNOWN_NUM;
            } else {
              pos = -POSTag.NOUN_LETTER;//
              word = Utility.UNKNOWN_LETTER;
            }
            value = 0;
            break;
          default:
            pos = atom.getPos();// '?'*256;
            break;
          }

          sn = new SegNode(i, i + 1,pos, value , word);
        }

        sn.setSrcWord(atom.getWord());
        segGraph.insert(sn, true);
      }

      String word = null;
      for (int i = 0; i < atoms.size(); i++) {
        int j = i + 1;
        word = atoms.get(i).getWord();
        // ����ǡ��·ݡ�����Ҫ�ָ�
        boolean flag = false;
        if (j < atoms.size()) {
          Atom a2 = atoms.get(j);
          if ("��".equals(word) && "��".equals(a2.getWord())) {
            segGraph.delete(i, j);
            segGraph.delete(i + 1, j + 1);
            word += a2.getWord();
            flag = true;
            j++;
          }
        }

View Full Code Here
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
      result = new ArrayList<Atom>();
      String[] ss = GFString.atomSplit(str);

      int index = str.indexOf(Utility.SENTENCE_BEGIN);
      if (index == 0) {
        Atom atom = new Atom();
        atom.setWord(Utility.SENTENCE_BEGIN);
        atom.setLen(Utility.SENTENCE_BEGIN.length());
        atom.setPos(Utility.CT_SENTENCE_BEGIN);
        result.add(atom);
        index += Utility.SENTENCE_BEGIN.length();
      }

      if (index == -1)
        index = 0;
      for (int i = index; i < ss.length; i++) {
        if (Utility.SENTENCE_END.equals(str.substring(i))) {
          Atom atom = new Atom();
          atom.setWord(Utility.SENTENCE_END);
          atom.setLen(Utility.SENTENCE_END.length());
          atom.setPos(Utility.CT_SENTENCE_END);
          result.add(atom);
          break;
        }

        String s = ss[i];
        sAtom += s;
        int curType = Utility.charType(s);
        if (".".equals(s)
            && (i + 1 < ss.length && (Utility.charType(ss[i + 1]) == Utility.CT_NUM || GFString
                .isNumeric(ss[i+1]))))
          curType = Utility.CT_NUM;

        // ����Ǻ��֡��ָ�����
        if (curType == Utility.CT_CHINESE || curType == Utility.CT_INDEX || curType == Utility.CT_DELIMITER
            || curType == Utility.CT_OTHER) {

          Atom atom = new Atom();
          atom.setWord(s);
          atom.setLen(s.length());
          atom.setPos(curType);
          result.add(atom);
          sAtom = "";
        }
        // ��������֡���ĸ�����ֽڷ��ţ�������ڵ���Щ��Ϊһ��ԭ�ӡ����磺����SHX-123�ͺŵ��ֻ��������е�SHX-123����һ��ԭ��
        else {
          int nextType = 255;// ��һ���ַ�������
          if (i < ss.length - 1)
            nextType = Utility.charType(ss[i + 1]);
          if (nextType != curType || i == ss.length - 1) {
            Atom atom = new Atom();
            atom.setWord(sAtom);
            atom.setLen(sAtom.length());
            atom.setPos(curType);
            result.add(atom);
            sAtom = "";
          }
        }
      }
View Full Code Here
TOP

Related Classes of org.ictclas4j.bean.Atom

  • org.ictclas4j.segment.AtomSeg
  • org.ictclas4j.segment.GraphGenerate
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.