Class: Text::Hyphen::Dictionary

Inherits:
Object
  • Object
show all
Defined in:
lib/text/hyphen.rb,
ext/text/text_hyphen.c

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

.new(filename) ⇒ Object 



8
9
10
11
12
13
14
15
16
17
18
19
 
# File 'ext/text/text_hyphen.c', line 8

VALUE th_dict_new(VALUE class, VALUE filename) {
  VALUE rb_dict;
  HyphenDict * dictionary; 
  const char * fname;
  fname = StringValuePtr(filename);
  if(!(dictionary = hnj_hyphen_load(fname))) {
    rb_sys_fail(fname);
  }
  rb_dict = Data_Wrap_Struct(class, 0, hnj_hyphen_free, dictionary);
  rb_obj_call_init(rb_dict, 0, 0);
  return rb_dict;
}

Instance Method Details

#analyze(string) ⇒ Object Also known as: analyse 



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
 
# File 'ext/text/text_hyphen.c', line 21

VALUE th_dict_analyze(VALUE self, VALUE string) {
  HyphenDict * dict;
  const char * word;
  int word_size, hyph_status, hyph_weight, idx;
  char * hyphens;
  VALUE rb_hyphens, downcase;
  ID s_downcase;
  s_downcase = rb_intern("downcase");
  downcase = rb_funcall(string, s_downcase, 0);
  word = StringValuePtr(downcase);
  word_size = strlen(word);
  hyphens = malloc(word_size * 2);
  Data_Get_Struct(self, HyphenDict, dict);
  hnj_hyphen_hyphenate(dict, word, word_size, hyphens);
  /*
   * This used to work with another version of libhyphen - but which?
   *
  if(hnj_hyphen_hyphenate(dict, word, word_size, hyphens)) {
    rb_funcall(rb_mKernel, rb_intern("puts"), 1, rb_str_new2("bailing out due to hyphenate error..."));
    free(hyphens);
    rb_raise(rb_eRuntimeError, "Hyphenation Error");
  }
  */
  rb_hyphens = rb_ary_new();
  for(idx = 0; idx < word_size; idx++) {
    hyph_status = hyphens[idx] - 48;
    if(hyph_status & 1) {
      hyph_weight = (hyph_status - 1) / 2 + 1;
    } else {
      hyph_weight = (hyph_status / 2) * -1;  
    }
    rb_ary_push(rb_hyphens, INT2FIX(hyph_weight));
  }
  free(hyphens);
  return rb_hyphens;
}

#best_chance(word) ⇒ Object 



11
12
13
14
15
16
17
 
# File 'lib/text/hyphen.rb', line 11

def best_chance(word)
  weights = analyze(word)
  best = weights.max
  if(best > 0)
    weights.index(best) + 1
  end
end

#best_chance_before(pos, word) ⇒ Object 



18
19
20
21
22
23
24
25
 
# File 'lib/text/hyphen.rb', line 18

def best_chance_before(pos, word)
  weights = analyze(word)
  valid = weights[0,pos]
  best = valid.max
  if(best > 0)
    valid.rindex(best) + 1
  end
end

#first_chance(word) ⇒ Object 



26
27
28
29
30
31
32
33
34
35
 
# File 'lib/text/hyphen.rb', line 26

def first_chance(word)
  idx = 0
  analyze(word).any? { |weight|
    idx += 1
    if(weight > 0)
      return idx
    end
  }
  nil
end

#hyphenate_before(pos, word) ⇒ Object 



57
58
59
60
61
62
63
 
# File 'lib/text/hyphen.rb', line 57

def hyphenate_before(pos, word)
  hyph = word.dup
  if(pos = last_chance_before(pos, word))
    hyph[pos,0] = '-'
  end
  hyph
end

#hyphenate_best(word) ⇒ Object 



36
37
38
39
40
41
42
 
# File 'lib/text/hyphen.rb', line 36

def hyphenate_best(word)
  hyph = word.dup
  if(pos = best_chance(word))
    hyph[pos,0] = '-'
  end
  hyph
end

#hyphenate_best_before(pos, word) ⇒ Object 



43
44
45
46
47
48
49
 
# File 'lib/text/hyphen.rb', line 43

def hyphenate_best_before(pos, word)
  hyph = word.dup
  if(pos = best_chance_before(pos, word))
    hyph[pos,0] = '-'
  end
  hyph
end

#hyphenate_first(word) ⇒ Object 



50
51
52
53
54
55
56
 
# File 'lib/text/hyphen.rb', line 50

def hyphenate_first(word)
  hyph = word.dup
  if(pos = first_chance(word))
    hyph[pos,0] = '-'
  end
  hyph
end

#last_chance_before(pos, word) ⇒ Object 



64
65
66
67
68
69
70
71
72
 
# File 'lib/text/hyphen.rb', line 64

def last_chance_before(pos, word)
  weights = analyze(word)
  pos.downto(0) { |pos|
    if(weights.at(pos - 1) > 0)
      return pos
    end
  }
  nil
end