wordcuta 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/Gemfile +5 -0
- data/LICENSE +165 -0
- data/README.md +1 -0
- data/data/thai-dix.txt +15890 -0
- data/lib/wordcuta.rb +65 -0
- data/wordcuta.gemspec +12 -0
- metadata +49 -0
data/lib/wordcuta.rb
ADDED
@@ -0,0 +1,65 @@
|
|
1
|
+
# coding: utf-8
|
2
|
+
require "ffi"
|
3
|
+
|
4
|
+
module WordcutFFI
|
5
|
+
class TextRange < FFI::Struct
|
6
|
+
layout :s, :size_t,
|
7
|
+
:e, :size_t
|
8
|
+
end
|
9
|
+
|
10
|
+
extend FFI::Library
|
11
|
+
|
12
|
+
ffi_lib "wordcutw"
|
13
|
+
|
14
|
+
attach_function :wordcut_new_with_dict, [:string], :pointer
|
15
|
+
attach_function :wordcut_into_text_ranges, [:pointer, :string, :pointer], :pointer
|
16
|
+
attach_function :wordcut_into_strings, [:pointer, :string, :pointer], :pointer
|
17
|
+
attach_function :wordcut_put_delimiters, [:pointer, :string, :string], :string
|
18
|
+
attach_function :delete_wordcut, [:pointer], :void
|
19
|
+
attach_function :delete_strings, [:pointer, :size_t], :void
|
20
|
+
end
|
21
|
+
|
22
|
+
module WordcutA
|
23
|
+
TextRange = Struct.new(:s, :e)
|
24
|
+
|
25
|
+
DEFAULT_THAI_DICT_PATH = File.expand_path('../../data/thai-dix.txt', __FILE__)
|
26
|
+
|
27
|
+
class Wordcut
|
28
|
+
def initialize(dict_path)
|
29
|
+
@wordcut_p = FFI::AutoPointer.new(WordcutFFI.wordcut_new_with_dict(dict_path),
|
30
|
+
WordcutFFI.method(:delete_wordcut))
|
31
|
+
end
|
32
|
+
|
33
|
+
def into_ranges(text)
|
34
|
+
ranges = []
|
35
|
+
FFI::MemoryPointer.new(:size_t, 1) do |range_cnt|
|
36
|
+
ranges_p = WordcutFFI.wordcut_into_text_ranges(@wordcut_p, text, range_cnt)
|
37
|
+
ranges = range_cnt.get_uint(0)
|
38
|
+
.times
|
39
|
+
.map {WordcutFFI::TextRange.new(ranges_p + _1 * WordcutFFI::TextRange.size)}
|
40
|
+
.map {TextRange.new(_1[:s], _1[:e])}
|
41
|
+
end
|
42
|
+
return ranges
|
43
|
+
end
|
44
|
+
|
45
|
+
def into_strings(text)
|
46
|
+
words = []
|
47
|
+
FFI::MemoryPointer.new(:size_t, 1) do |words_cnt_p|
|
48
|
+
words_p = WordcutFFI::wordcut_into_strings(@wordcut_p, text, words_cnt_p)
|
49
|
+
words_cnt = words_cnt_p.get_uint(0)
|
50
|
+
words = words_p.get_array_of_string(0, words_cnt)
|
51
|
+
WordcutFFI::delete_strings(words_p, words_cnt)
|
52
|
+
end
|
53
|
+
return words.map {_1.force_encoding("UTF-8")}
|
54
|
+
end
|
55
|
+
|
56
|
+
def put_delimiters(text, delim)
|
57
|
+
WordcutFFI::wordcut_put_delimiters(@wordcut_p, text, delim).force_encoding("UTF-8")
|
58
|
+
end
|
59
|
+
end
|
60
|
+
end
|
61
|
+
|
62
|
+
# wordcut = WordcutA::Wordcut.new("thai.txt")
|
63
|
+
# p wordcut.into_ranges("กากาม")
|
64
|
+
# p wordcut.into_strings("กากาขาม")
|
65
|
+
# p wordcut.put_delimiters("กากาขามมาก", "---")
|
data/wordcuta.gemspec
ADDED
@@ -0,0 +1,12 @@
|
|
1
|
+
Gem::Specification.new do |s|
|
2
|
+
s.name = 'wordcuta'
|
3
|
+
s.version = '0.1.0'
|
4
|
+
s.authors = ['Vee Satayamas']
|
5
|
+
s.email = ['5ssgdxltv@relay.firefox.com']
|
6
|
+
s.licenses = ['LGPL-3.0']
|
7
|
+
s.description = "A word segmentation tools for ASEAN languages wrapper for Ruby"
|
8
|
+
s.homepage = "https://github.com/veer66/wordcuta"
|
9
|
+
s.required_ruby_version = Gem::Requirement.new(">= 2.3.0")
|
10
|
+
s.summary = "A word segmentation tools for ASEAN languages wrapper for Ruby"
|
11
|
+
s.files = %w(README.md LICENSE Gemfile wordcuta.gemspec data/thai-dix.txt lib/wordcuta.rb)
|
12
|
+
end
|
metadata
ADDED
@@ -0,0 +1,49 @@
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
2
|
+
name: wordcuta
|
3
|
+
version: !ruby/object:Gem::Version
|
4
|
+
version: 0.1.0
|
5
|
+
platform: ruby
|
6
|
+
authors:
|
7
|
+
- Vee Satayamas
|
8
|
+
autorequire:
|
9
|
+
bindir: bin
|
10
|
+
cert_chain: []
|
11
|
+
date: 2021-04-23 00:00:00.000000000 Z
|
12
|
+
dependencies: []
|
13
|
+
description: A word segmentation tools for ASEAN languages wrapper for Ruby
|
14
|
+
email:
|
15
|
+
- 5ssgdxltv@relay.firefox.com
|
16
|
+
executables: []
|
17
|
+
extensions: []
|
18
|
+
extra_rdoc_files: []
|
19
|
+
files:
|
20
|
+
- Gemfile
|
21
|
+
- LICENSE
|
22
|
+
- README.md
|
23
|
+
- data/thai-dix.txt
|
24
|
+
- lib/wordcuta.rb
|
25
|
+
- wordcuta.gemspec
|
26
|
+
homepage: https://github.com/veer66/wordcuta
|
27
|
+
licenses:
|
28
|
+
- LGPL-3.0
|
29
|
+
metadata: {}
|
30
|
+
post_install_message:
|
31
|
+
rdoc_options: []
|
32
|
+
require_paths:
|
33
|
+
- lib
|
34
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
35
|
+
requirements:
|
36
|
+
- - ">="
|
37
|
+
- !ruby/object:Gem::Version
|
38
|
+
version: 2.3.0
|
39
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
40
|
+
requirements:
|
41
|
+
- - ">="
|
42
|
+
- !ruby/object:Gem::Version
|
43
|
+
version: '0'
|
44
|
+
requirements: []
|
45
|
+
rubygems_version: 3.2.3
|
46
|
+
signing_key:
|
47
|
+
specification_version: 4
|
48
|
+
summary: A word segmentation tools for ASEAN languages wrapper for Ruby
|
49
|
+
test_files: []
|