nlpir 0.0.4-x86-mingw32

Sign up to get free protection for your applications and to get access to all the features.
Files changed (92) hide show
  1. checksums.yaml +7 -0
  2. data/.gitignore +17 -0
  3. data/LICENSE.txt +22 -0
  4. data/README.md +195 -0
  5. data/Rakefile +11 -0
  6. data/bin/NLPIR.dll +0 -0
  7. data/lib/Data/BIG2GBK.map +0 -0
  8. data/lib/Data/BIG5.pdat +0 -0
  9. data/lib/Data/BIG5.wordlist +0 -0
  10. data/lib/Data/BiWord.big +0 -0
  11. data/lib/Data/Configure.xml +15 -0
  12. data/lib/Data/CoreDict.pdat +0 -0
  13. data/lib/Data/CoreDict.pos +0 -0
  14. data/lib/Data/CoreDict.unig +0 -0
  15. data/lib/Data/FieldDict.pdat +0 -0
  16. data/lib/Data/FieldDict.pos +0 -0
  17. data/lib/Data/GBK.pdat +0 -0
  18. data/lib/Data/GBK.wordlist +0 -0
  19. data/lib/Data/GBK2BIG.map +0 -0
  20. data/lib/Data/GBK2GBKC.map +0 -0
  21. data/lib/Data/GBK2UTF.map +0 -0
  22. data/lib/Data/GBKA.pdat +0 -0
  23. data/lib/Data/GBKA.wordlist +0 -0
  24. data/lib/Data/GBKA2UTF.map +0 -0
  25. data/lib/Data/GBKC.pdat +0 -0
  26. data/lib/Data/GBKC.wordlist +0 -0
  27. data/lib/Data/GBKC2GBK.map +0 -0
  28. data/lib/Data/GranDict.pdat +3369 -8
  29. data/lib/Data/GranDict.pos +0 -0
  30. data/lib/Data/ICTPOS.map +96 -0
  31. data/lib/Data/NLPIR.ctx +0 -0
  32. data/lib/Data/NLPIR.user +0 -0
  33. data/lib/Data/NLPIR_First.map +96 -0
  34. data/lib/Data/NewWord.lst +25 -0
  35. data/lib/Data/PKU.map +96 -0
  36. data/lib/Data/PKU_First.map +96 -0
  37. data/lib/Data/UTF2GBK.map +0 -0
  38. data/lib/Data/UTF2GBKA.map +0 -0
  39. data/lib/Data/UTF8.pdat +0 -0
  40. data/lib/Data/UTF8.wordlist +0 -0
  41. data/lib/Data/UserDict.pdat +0 -0
  42. data/lib/Data/charset.type +0 -0
  43. data/lib/Data/nr.ctx +0 -0
  44. data/lib/Data/nr.fsa +0 -0
  45. data/lib/Data/nr.role +0 -0
  46. data/lib/nlpir/version.rb +3 -0
  47. data/lib/nlpir.rb +275 -0
  48. data/nlpir.gemspec +25 -0
  49. data/test/Data/BIG2GBK.map +0 -0
  50. data/test/Data/BIG5.pdat +0 -0
  51. data/test/Data/BIG5.wordlist +0 -0
  52. data/test/Data/BiWord.big +0 -0
  53. data/test/Data/Configure.xml +15 -0
  54. data/test/Data/CoreDict.pdat +0 -0
  55. data/test/Data/CoreDict.pos +0 -0
  56. data/test/Data/CoreDict.unig +0 -0
  57. data/test/Data/FieldDict.pdat +0 -0
  58. data/test/Data/FieldDict.pos +0 -0
  59. data/test/Data/GBK.pdat +0 -0
  60. data/test/Data/GBK.wordlist +0 -0
  61. data/test/Data/GBK2BIG.map +0 -0
  62. data/test/Data/GBK2GBKC.map +0 -0
  63. data/test/Data/GBK2UTF.map +0 -0
  64. data/test/Data/GBKA.pdat +0 -0
  65. data/test/Data/GBKA.wordlist +0 -0
  66. data/test/Data/GBKA2UTF.map +0 -0
  67. data/test/Data/GBKC.pdat +0 -0
  68. data/test/Data/GBKC.wordlist +0 -0
  69. data/test/Data/GBKC2GBK.map +0 -0
  70. data/test/Data/GranDict.pdat +3369 -8
  71. data/test/Data/GranDict.pos +0 -0
  72. data/test/Data/ICTPOS.map +96 -0
  73. data/test/Data/NLPIR.ctx +0 -0
  74. data/test/Data/NLPIR.user +0 -0
  75. data/test/Data/NLPIR_First.map +96 -0
  76. data/test/Data/NewWord.lst +73 -0
  77. data/test/Data/PKU.map +96 -0
  78. data/test/Data/PKU_First.map +96 -0
  79. data/test/Data/UTF2GBK.map +0 -0
  80. data/test/Data/UTF2GBKA.map +0 -0
  81. data/test/Data/UTF8.pdat +0 -0
  82. data/test/Data/UTF8.wordlist +0 -0
  83. data/test/Data/UserDict.pdat +0 -0
  84. data/test/Data/charset.type +0 -0
  85. data/test/Data/nr.ctx +0 -0
  86. data/test/Data/nr.fsa +0 -0
  87. data/test/Data/nr.role +0 -0
  88. data/test/test.txt +52 -0
  89. data/test/test_nlpir.rb +158 -0
  90. data/test/test_result.txt +87 -0
  91. data/test/userdict.txt +5 -0
  92. metadata +206 -0
data/nlpir.gemspec ADDED
@@ -0,0 +1,25 @@
1
+ # coding: utf-8
2
+ lib = File.expand_path('../lib', __FILE__)
3
+ $LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
4
+ require 'nlpir/version'
5
+
6
+ Gem::Specification.new do |spec|
7
+ spec.name = "nlpir"
8
+ spec.version = Nlpir::VERSION
9
+ spec.authors = ["JoeWoo"]
10
+ spec.email = ["0wujian0@gmail.com"]
11
+ spec.description = %q{"A rubygem wrapper of chinese segment tools ICTCLAS2013"}
12
+ spec.summary = %q{"ICTCLAS是由中国科学院计算技术研究所研发。中文分词；词性标注；命名实体识别；新词识别；同时支持用户词典。"}
13
+ spec.homepage = "https://github.com/JoeWoo/nlpir_win"
14
+ spec.license = "MIT"
15
+
16
+ spec.files = `git ls-files`.split($/)
17
+ spec.executables = spec.files.grep(%r{^bin/}) { |f| File.basename(f) }
18
+ spec.test_files = spec.files.grep(%r{^(test|spec|features)/})
19
+ spec.require_paths = ["lib"]
20
+ spec.required_ruby_version = '>= 2.0.0'
21
+ spec.platform = Gem::Platform::local
22
+
23
+ spec.add_development_dependency "bundler", "~> 1.3"
24
+ spec.add_development_dependency "rake"
25
+ end
data/test/Data/BIG2GBK.map ADDED
Binary file
data/test/Data/BIG5.pdat ADDED
Binary file
data/test/Data/BIG5.wordlist ADDED
Binary file
data/test/Data/BiWord.big ADDED
Binary file
data/test/Data/Configure.xml ADDED
@@ -0,0 +1,15 @@
1
+ <?xmlversion="1.0"encoding="GB2312"?>
2
+ <NLPIR>
3
+ <TagSet>ICTPOS.map</TagSet>//���Ա�ע��ӳ���ļ�
4
+ <UserDict>on</UserDict>//On��UserDictionaryapplied;Off:notapplied��
5
+ <UserDictPrior>On</UserDictPrior>//�û��ʵ�����,Addedin2006-03-16,requiredbyNECOn���û��ʵ�ͺ��Ĵʵ���ͬʱ�еĴʻ㣬�û��ʵ����ȣ������ܲ�Ҫ���ã���������Ĵʵ��еĴʶ�����Ϊ�û��ʵ䣬��Ч���ʵ��䷴
6
+ <FieldDict>off</FieldDict>//On��FieldDictionaryapplied;Off:notapplied��
7
+ <GranularityContorl>off</GranularityContorl>
8
+ <Log>On</Log>//On,Off�����磺Off:�ر���־���ܣ�On:����־����
9
+ <version>2013</version>//ϵͳ�汾��
10
+ <Modify>2012-11-14</Modify>//ϵͳ����޶�ʱ��
11
+ <Lexicon>2012-11-14</Lexicon>//�ʵ�����޶�ʱ��
12
+ <adaptive>true</adaptive>//����Ӧ�ִʣ�Ĭ��Ϊfalse������Ӧ�ִʵ�Ч�ʻ�ϵ�
13
+ <author>�Ż�ƽ��ʿ</author>//����
14
+ <Contact>pipy_zhang@msn.com</Contact>//������ϵ��ʽ
15
+ </NLPIR>
data/test/Data/CoreDict.pdat ADDED
Binary file
data/test/Data/CoreDict.pos ADDED
Binary file
data/test/Data/CoreDict.unig ADDED
Binary file
data/test/Data/FieldDict.pdat ADDED
Binary file
data/test/Data/FieldDict.pos ADDED
Binary file
data/test/Data/GBK.pdat ADDED
Binary file
data/test/Data/GBK.wordlist ADDED
Binary file
data/test/Data/GBK2BIG.map ADDED
Binary file
data/test/Data/GBK2GBKC.map ADDED
Binary file
data/test/Data/GBK2UTF.map ADDED
Binary file
data/test/Data/GBKA.pdat ADDED
Binary file
data/test/Data/GBKA.wordlist ADDED
Binary file
data/test/Data/GBKA2UTF.map ADDED
Binary file
data/test/Data/GBKC.pdat ADDED
Binary file
data/test/Data/GBKC.wordlist ADDED
Binary file
data/test/Data/GBKC2GBK.map ADDED
Binary file