kokugo_tagger 0.0.2 → 0.0.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: b609be62c8aea30eda82811044d1d7ecc6eeb1b4
-  data.tar.gz: 733d6a86cea778492416df55f6e79de5c072a4c3
+  metadata.gz: 38c7ab0d9feb7ce099075914af22947354187b40
+  data.tar.gz: b240e3e012dce85475716a16ed6b1ff738d65327
 SHA512:
-  metadata.gz: 02a682b673b2103320ced6836cb3784c8b9543628840d9aec1033de7a509869449818be930c6c2f06b22ccbfeb4748a2172916a2e849db836dd51b74837ad46d
-  data.tar.gz: b2ba4fe4e4aeeeb83d9a709243479c70298426e86ab771a19eaa28420ddd8dee6b18dcafcedf3b4c5e648a513b7b06cbbeff97352b13b1a0c7b09cb0361eafec
+  metadata.gz: 28c3d0f941e530245655e9704d675343fdabb9f94123bf8065aad4044492b9de0b50e2cead2cd99b1ceecceb8a6d89a2417a8d52566d39a8aed3250aec50d93f
+  data.tar.gz: d2a488f3efdfeaabcaf72d610052db4327ba8af401b9e974f2e191e0579efe730d80dc5d9086a9db1a3eeaf4a4a5f89ab1b6e5136ced2cdfa75a66bda4a25797

data/lib/kokugo_tagger/parser.rb

@@ -29,10 +29,10 @@ module CabochaParser
 		return excab
 	end
 	def parse_chunk(line)
-		null, id, rel, part, score = line.chomp.split("\s")
-		link, dep = rel[0..-2], rel[-1]
+		null, id, dep, part, score = line.chomp.split("\s")
+		link, rel = dep[0..-2], dep[-1]
 		head, func = part.split('/')
-		chunk = {type: 'CHUNK', id: id, link: link, dep: dep, head: head, func: func, score: score}
+		chunk = {type: 'CHUNK', id: id, link: link, rel: rel, head: head, func: func, score: score}
 		return chunk
 	end
 	def parse_token(line)

data/lib/kokugo_tagger/tagger.rb

@@ -25,8 +25,18 @@ module KokugoTagger
 		cform data
 	end
 	def segment_s(data)
+		@segments ||= []
+		@segments << data
+		@last_item = data
+	end
+	def group_s(data)
+		@groups ||= []
+		@groups << data
+		@last_item = data
 	end
 	def attr(data)
+		@last_item[:attributes] ||= []
+		@last_item[:attributes] << data
 	end
 	def eos(data)
 		before_eos
@@ -35,7 +45,7 @@ module KokugoTagger
 			puts '#! ATTR bccwj-kok:pred "%s述語"' % chunk[:pos] if chunk[:pred]
 			puts '#! ATTR bccwj-kok:conj "%s"' % chunk[:conj] if chunk[:conj]
 		end
-		@chunks, @chunk, @lpos, @segments = nil
+		@chunks, @chunk, @lpos, @segments, @groups = nil
 	end
 	def pos(token)
 		case token[:pos]
@@ -61,10 +71,12 @@ module KokugoTagger
 			case token[:text]
 			when 'が'
 				@chunk.update conj:'主語'
+			when 'を', 'に'
+				@chunk.update conj:'補語'
 			when 'の', 'との', 'という', 'といった'
 				@chunk.update conj:'修飾(連体)'
 			else
-				@chunk.update conj:'補語'
+				@chunk.update conj:'修飾(連用)'
 			end
 		when /^(助詞-副助詞|助詞-係助詞)/
 			@chunk.update conj:'修飾(連用)'
@@ -88,6 +100,30 @@ module KokugoTagger
 		end			
 	end
 	def before_eos
+		# BCCWJ-DepPara
+		@chunks.each do |chunk|
+			chunk[:conj] = [chunk[:conj], '断片'].compact.join(':') if chunk[:rel] == 'F'
+			chunk[:conj] = [chunk[:conj], '文節内'].compact.join(':') if chunk[:rel] == 'B'
+			chunk[:conj] = '文末' if chunk[:rel] == 'Z'
+		end
+		# 並列・同格関係
+		@groups ||= []
+		@segments ||= []
+		@groups.each do |group|
+			next unless group[:name] =~ /^(Parallel|Apposition)$/
+			members = group[:member].map{|n| n.to_i}
+			members = @segments.values_at(*members)
+			chunk_ids = members.map do |segment|
+				_end = segment[:end].to_i
+				chunk = @chunks.find{|c| c[:start] < _end and c[:end] >= _end}
+				chunk[:id].to_i if chunk
+			end
+			chunk_ids = chunk_ids.compact.uniq.sort
+			if chunk_ids.size > 1
+				conj = {'Parallel' => '並立', 'Apposition' => '同格'}[group[:name]]
+				chunk_ids[0..-2].each{|cid| @chunks[cid][:conj] = conj}
+			end
+		end
 		# 属性を付与できなかった文節に対して、係り受けを利用して属性を補完
 		# 連用成分を受ける文節を述語とみなす
 		@chunks.each do |chunk|

data/lib/kokugo_tagger/version.rb

@@ -1,3 +1,3 @@
 module KokugoTagger
-  VERSION = "0.0.2"
+  VERSION = "0.0.3"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: kokugo_tagger
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.0.3
 platform: ruby
 authors:
 - Mizuho IMADA
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-11-17 00:00:00.000000000 Z
+date: 2015-06-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -77,7 +77,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.4.3
+rubygems_version: 2.2.2
 signing_key: 
 specification_version: 4
 summary: Write a short summary. Required.