wordcuta 0.1.0

data/lib/wordcuta.rb

@@ -0,0 +1,65 @@
+# coding: utf-8
+require "ffi"
+
+module WordcutFFI
+  class TextRange < FFI::Struct
+    layout :s, :size_t,
+           :e, :size_t
+  end
+
+  extend FFI::Library
+
+  ffi_lib "wordcutw"
+  
+  attach_function :wordcut_new_with_dict, [:string], :pointer
+  attach_function :wordcut_into_text_ranges, [:pointer, :string, :pointer], :pointer
+  attach_function :wordcut_into_strings, [:pointer, :string, :pointer], :pointer
+  attach_function :wordcut_put_delimiters, [:pointer, :string, :string], :string
+  attach_function :delete_wordcut, [:pointer], :void
+  attach_function :delete_strings, [:pointer, :size_t], :void
+end
+
+module WordcutA
+  TextRange = Struct.new(:s, :e)
+
+  DEFAULT_THAI_DICT_PATH = File.expand_path('../../data/thai-dix.txt', __FILE__)
+  
+  class Wordcut
+    def initialize(dict_path)
+      @wordcut_p = FFI::AutoPointer.new(WordcutFFI.wordcut_new_with_dict(dict_path),
+                                        WordcutFFI.method(:delete_wordcut))
+    end
+
+    def into_ranges(text)
+      ranges = []
+      FFI::MemoryPointer.new(:size_t, 1) do |range_cnt|
+        ranges_p = WordcutFFI.wordcut_into_text_ranges(@wordcut_p, text, range_cnt)
+        ranges = range_cnt.get_uint(0)
+                   .times
+                   .map {WordcutFFI::TextRange.new(ranges_p + _1 * WordcutFFI::TextRange.size)}
+                   .map {TextRange.new(_1[:s], _1[:e])}
+      end
+      return ranges
+    end
+
+    def into_strings(text)
+      words = []
+      FFI::MemoryPointer.new(:size_t, 1) do |words_cnt_p|
+        words_p = WordcutFFI::wordcut_into_strings(@wordcut_p, text, words_cnt_p)
+        words_cnt = words_cnt_p.get_uint(0)
+        words = words_p.get_array_of_string(0, words_cnt)
+        WordcutFFI::delete_strings(words_p, words_cnt)
+      end
+      return words.map {_1.force_encoding("UTF-8")}
+    end
+
+    def put_delimiters(text, delim)
+      WordcutFFI::wordcut_put_delimiters(@wordcut_p, text, delim).force_encoding("UTF-8")
+    end
+  end
+end
+
+# wordcut = WordcutA::Wordcut.new("thai.txt")
+# p wordcut.into_ranges("กากาม")
+# p wordcut.into_strings("กากาขาม")
+# p wordcut.put_delimiters("กากาขามมาก", "---")

data/wordcuta.gemspec

@@ -0,0 +1,12 @@
+Gem::Specification.new do |s|
+  s.name = 'wordcuta'
+  s.version = '0.1.0'
+  s.authors = ['Vee Satayamas']
+  s.email = ['5ssgdxltv@relay.firefox.com']
+  s.licenses = ['LGPL-3.0']
+  s.description = "A word segmentation tools for ASEAN languages wrapper for Ruby"
+  s.homepage = "https://github.com/veer66/wordcuta"
+  s.required_ruby_version = Gem::Requirement.new(">= 2.3.0")
+  s.summary = "A word segmentation tools for ASEAN languages wrapper for Ruby"
+  s.files = %w(README.md LICENSE Gemfile wordcuta.gemspec data/thai-dix.txt lib/wordcuta.rb)
+end

metadata

@@ -0,0 +1,49 @@
+--- !ruby/object:Gem::Specification
+name: wordcuta
+version: !ruby/object:Gem::Version
+  version: 0.1.0
+platform: ruby
+authors:
+- Vee Satayamas
+autorequire:
+bindir: bin
+cert_chain: []
+date: 2021-04-23 00:00:00.000000000 Z
+dependencies: []
+description: A word segmentation tools for ASEAN languages wrapper for Ruby
+email:
+- 5ssgdxltv@relay.firefox.com
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- Gemfile
+- LICENSE
+- README.md
+- data/thai-dix.txt
+- lib/wordcuta.rb
+- wordcuta.gemspec
+homepage: https://github.com/veer66/wordcuta
+licenses:
+- LGPL-3.0
+metadata: {}
+post_install_message:
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: 2.3.0
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubygems_version: 3.2.3
+signing_key:
+specification_version: 4
+summary: A word segmentation tools for ASEAN languages wrapper for Ruby
+test_files: []