text_hyphen 1.0.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/History.txt +6 -0
- data/LICENCE.txt +339 -0
- data/Manifest.txt +15 -0
- data/README.txt +21 -0
- data/Rakefile +28 -0
- data/ext/text/MANIFEST +1 -0
- data/ext/text/extconf.rb +10 -0
- data/ext/text/text_hyphen.c +67 -0
- data/install.rb +1098 -0
- data/lib/text/hyphen.rb +75 -0
- data/redist/libhnj-0.1.1.tar.gz +0 -0
- data/test/data/README_hyph_de_CH.txt +15 -0
- data/test/data/hyph_de_CH.dic +5799 -0
- data/test/rebuild.rb +16 -0
- data/test/test_dictionary.rb +159 -0
- metadata +101 -0
data/lib/text/hyphen.rb
ADDED
@@ -0,0 +1,75 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
# Text::Hyphen -- text-hyphen -- 16.03.2004 -- hwyss@ywesee.com
|
3
|
+
|
4
|
+
require 'text_hyphen.so'
|
5
|
+
|
6
|
+
VERSION = '1.0.0'
|
7
|
+
|
8
|
+
module Text
|
9
|
+
module Hyphen
|
10
|
+
class Dictionary
|
11
|
+
def best_chance(word)
|
12
|
+
weights = analyze(word)
|
13
|
+
best = weights.max
|
14
|
+
if(best > 0)
|
15
|
+
weights.index(best) + 1
|
16
|
+
end
|
17
|
+
end
|
18
|
+
def best_chance_before(pos, word)
|
19
|
+
weights = analyze(word)
|
20
|
+
valid = weights[0,pos]
|
21
|
+
best = valid.max
|
22
|
+
if(best > 0)
|
23
|
+
valid.rindex(best) + 1
|
24
|
+
end
|
25
|
+
end
|
26
|
+
def first_chance(word)
|
27
|
+
idx = 0
|
28
|
+
analyze(word).any? { |weight|
|
29
|
+
idx += 1
|
30
|
+
if(weight > 0)
|
31
|
+
return idx
|
32
|
+
end
|
33
|
+
}
|
34
|
+
nil
|
35
|
+
end
|
36
|
+
def hyphenate_best(word)
|
37
|
+
hyph = word.dup
|
38
|
+
if(pos = best_chance(word))
|
39
|
+
hyph[pos,0] = '-'
|
40
|
+
end
|
41
|
+
hyph
|
42
|
+
end
|
43
|
+
def hyphenate_best_before(pos, word)
|
44
|
+
hyph = word.dup
|
45
|
+
if(pos = best_chance_before(pos, word))
|
46
|
+
hyph[pos,0] = '-'
|
47
|
+
end
|
48
|
+
hyph
|
49
|
+
end
|
50
|
+
def hyphenate_first(word)
|
51
|
+
hyph = word.dup
|
52
|
+
if(pos = first_chance(word))
|
53
|
+
hyph[pos,0] = '-'
|
54
|
+
end
|
55
|
+
hyph
|
56
|
+
end
|
57
|
+
def hyphenate_before(pos, word)
|
58
|
+
hyph = word.dup
|
59
|
+
if(pos = last_chance_before(pos, word))
|
60
|
+
hyph[pos,0] = '-'
|
61
|
+
end
|
62
|
+
hyph
|
63
|
+
end
|
64
|
+
def last_chance_before(pos, word)
|
65
|
+
weights = analyze(word)
|
66
|
+
pos.downto(0) { |pos|
|
67
|
+
if(weights.at(pos - 1) > 0)
|
68
|
+
return pos
|
69
|
+
end
|
70
|
+
}
|
71
|
+
nil
|
72
|
+
end
|
73
|
+
end
|
74
|
+
end
|
75
|
+
end
|
Binary file
|
@@ -0,0 +1,15 @@
|
|
1
|
+
Hyphenation dictionary
|
2
|
+
----------------------
|
3
|
+
|
4
|
+
Language: German (de DE).
|
5
|
+
Origin: Based on the TeX hyphenation tables
|
6
|
+
License: GNU LGPL license.
|
7
|
+
Author: conversion author is Marco Huggenberger<marco@by-night.ch>
|
8
|
+
|
9
|
+
Please note, this dictionary is based on syllable matching patterns
|
10
|
+
and thus should be suitable under other variations of German
|
11
|
+
|
12
|
+
HYPH de DE hyph_de
|
13
|
+
HYPH de CH hyph_de
|
14
|
+
|
15
|
+
|