greeb 0.1.0.rc6 → 0.1.0.rc7
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/lib/greeb/tokenizer.rb +10 -2
- data/lib/greeb/version.rb +1 -1
- metadata +3 -3
data/lib/greeb/tokenizer.rb
CHANGED
|
@@ -1,6 +1,9 @@
|
|
|
1
1
|
# encoding: utf-8
|
|
2
2
|
|
|
3
|
-
# Greeb's tokenization facilities. Use 'em with love.
|
|
3
|
+
# Greeb's tokenization facilities. Use 'em with love.
|
|
4
|
+
#
|
|
5
|
+
# Unicode character categories been obtained from
|
|
6
|
+
# <http://www.fileformat.info/info/unicode/category/index.htm>.
|
|
4
7
|
#
|
|
5
8
|
class Greeb::Tokenizer
|
|
6
9
|
# This runtime error appears when {Greeb::Tokenizer} tries to recognize
|
|
@@ -43,12 +46,16 @@ class Greeb::Tokenizer
|
|
|
43
46
|
|
|
44
47
|
# In-subsentence seprator (i.e.: "*" or "=").
|
|
45
48
|
#
|
|
46
|
-
SEPARATORS = /[ \p{
|
|
49
|
+
SEPARATORS = /[ \p{Nl}\p{No}\p{Pd}\p{Pc}\p{Po}\p{Sm}\p{So}\p{Sc}\p{Z}]+/u
|
|
47
50
|
|
|
48
51
|
# Line breaks.
|
|
49
52
|
#
|
|
50
53
|
BREAKS = /(\r\n|\n|\r)+/u
|
|
51
54
|
|
|
55
|
+
# Residuals.
|
|
56
|
+
#
|
|
57
|
+
RESIDUALS = /[\p{C}\p{M}\p{Sk}]+/u
|
|
58
|
+
|
|
52
59
|
attr_reader :text, :scanner
|
|
53
60
|
protected :scanner
|
|
54
61
|
|
|
@@ -86,6 +93,7 @@ class Greeb::Tokenizer
|
|
|
86
93
|
split_parse! PUNCTUATIONS, :punct or
|
|
87
94
|
split_parse! SEPARATORS, :separ or
|
|
88
95
|
split_parse! BREAKS, :break or
|
|
96
|
+
parse! RESIDUALS, :residual or
|
|
89
97
|
raise UnknownEntity.new(text, scanner.char_pos)
|
|
90
98
|
end
|
|
91
99
|
ensure
|
data/lib/greeb/version.rb
CHANGED
metadata
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: greeb
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 0.1.0.
|
|
4
|
+
version: 0.1.0.rc7
|
|
5
5
|
prerelease: 6
|
|
6
6
|
platform: ruby
|
|
7
7
|
authors:
|
|
@@ -9,7 +9,7 @@ authors:
|
|
|
9
9
|
autorequire:
|
|
10
10
|
bindir: bin
|
|
11
11
|
cert_chain: []
|
|
12
|
-
date: 2012-12-
|
|
12
|
+
date: 2012-12-14 00:00:00.000000000 Z
|
|
13
13
|
dependencies:
|
|
14
14
|
- !ruby/object:Gem::Dependency
|
|
15
15
|
requirement: !ruby/object:Gem::Requirement
|
|
@@ -109,7 +109,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
|
|
|
109
109
|
requirements:
|
|
110
110
|
- - ! '>='
|
|
111
111
|
- !ruby/object:Gem::Version
|
|
112
|
-
hash:
|
|
112
|
+
hash: 2716089438708653231
|
|
113
113
|
version: '0'
|
|
114
114
|
segments:
|
|
115
115
|
- 0
|