pubannotation_evaluator 1.0.2 → 1.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6fa609f2eaa40072aaf68f19d943908cf959c68c951cfabe7c530fae96efdbe8
-  data.tar.gz: 66e92d37b5db14676152c17d777ff25071844e3e4297ec8bb5cd4633cc867470
+  metadata.gz: 806a1cccb5be0bc733f7deb05fe647d03816235a28bffec791aeb450a33cca46
+  data.tar.gz: a10d449746dd75e7923a4c08e82e1c6e8ac7310ddf9524b876c61327d88e8936
 SHA512:
-  metadata.gz: ca65f506b81afaa8eb2fc8af8acebda766db0a66ec8d317796df7934ed6f92227ba7364004f6c7bfe507c790c067bea67d4887c7137b2cf980f35fe6efcdca7d
-  data.tar.gz: '0607055296bb34c586a7d8c03e91a71d5781a76478ae1673501dc29fe85eeda0f55e6e10584b906d9b2b8268bbf08cefef7815328723dee20cce16c8b5cf912a'
+  metadata.gz: 27486eff982a8a8972a03aa42d9eeca6f7598d9bf76631ba2881251be786bde75254b02b9502701e5601a81e7b6edb052d852ef3d1f006477a213bde602b3631
+  data.tar.gz: 6ea8c8f982c327a86c5cfa61a51a8f5bc4d3d03edd75d407ca720bf3494ab572d4d78f86c114d87211d533cb8172984d467f9267ef75b137168510207b35e623

data/lib/pubannotation_evaluator/pubannotation_evaluator.rb

@@ -3,17 +3,17 @@ class PubannotationEvaluator
 	SOFT_MATCH_WORDS = 2
 	EXACT_TYPE_MATCH = 'study_type == reference_type ? 1 : 0'
 
-	def initialize(soft_match_chatacters = SOFT_MATCH_CHARACTERS, soft_match_words = SOFT_MATCH_WORDS, denotation_type_match = EXACT_TYPE_MATCH, relation_type_match = EXACT_TYPE_MATCH)
-		@soft_match_chatacters = soft_match_chatacters
-		@soft_match_words = soft_match_words
+	def initialize(soft_match_chatacters, soft_match_words, denotation_type_match, relation_type_match)
+		@soft_match_chatacters = soft_match_chatacters || SOFT_MATCH_CHARACTERS
+		@soft_match_words = soft_match_words || SOFT_MATCH_WORDS
 		@denotation_type_match = eval <<-HEREDOC
 			Proc.new do |study_type, reference_type|
-				#{denotation_type_match}
+				#{denotation_type_match || EXACT_TYPE_MATCH}
 			end
 		HEREDOC
 		@relation_type_match = eval <<-HEREDOC
 			Proc.new do |study_type, reference_type|
-				#{relation_type_match}
+				#{relation_type_match || EXACT_TYPE_MATCH}
 			end
 		HEREDOC
 	end
@@ -41,7 +41,6 @@ class PubannotationEvaluator
 								 comparison_modifications.collect{|a| a.merge(type: :modification)}
 
 		docspec = {sourcedb:study_annotations[:sourcedb], sourceid:study_annotations[:sourceid]}
-		docspec[:divid] = study_annotations[:divid] if study_annotations.has_key?(:divid)
 		comparison.collect{|d| d.merge(docspec)}
 	end
 
@@ -71,6 +70,8 @@ class PubannotationEvaluator
 		matches  = find_denotation_matches(mmatches)
 		false_positives = study_denotations - matches.collect{|r| r[:study]}
 		false_negatives = reference_denotations - matches.collect{|r| r[:reference]}
+		false_positives.each{|r| r.merge!(text:text[r[:span][:begin]...r[:span][:end]])}
+		false_negatives.each{|r| r.merge!(text:text[r[:span][:begin]...r[:span][:end]])}
 		comparison = matches + false_positives.collect{|s| {study:s}} + false_negatives.collect{|r| {reference:r}}
 		[comparison, mmatches]
 	end
@@ -83,8 +84,11 @@ class PubannotationEvaluator
 		mmatches = []
 		study_denotations.each do |s|
 			r_begin = reference_denotations.bsearch_index{|r| r[:span][:end] > s[:span][:begin]}
+			break if r_begin.nil?
+
 			r_end = reference_denotations.bsearch_index{|r| r[:span][:begin] > s[:span][:end]}
 			r_end = r_end.nil? ? -1 : r_end - 1
+
 			reference_denotations[r_begin .. r_end].each do |r|
 				relatedness = get_relatedness_of_denotations(s, r, text)
 				mmatches << {study:s, reference:r, weight:relatedness} if relatedness > 0
@@ -165,8 +169,7 @@ class PubannotationEvaluator
 
 		# predicate match
 		match_pred_weight = @relation_type_match.call(s[:pred], r[:pred])
-
-		return (match_subj[:weight] + match_obj[:weight] + match_pred_weight).to_f / 3
+		# return (match_subj[:weight] + match_obj[:weight] + match_pred_weight).to_f / 3
 	end
 
 	def find_relation_matches(matches)
@@ -192,11 +195,11 @@ class PubannotationEvaluator
 		r_matched = []
 		matches_group_by_s = matches.group_by{|m| m[:study]}
 		matches_group_by_s.each_value do |m|
+			m.delete_if{|i| r_matched.include?(i[:reference])}
 			if m.length == 1
 				s_matched << m[0][:study]
 				r_matched << m[0][:reference]
-			else
-				m.delete_if{|i| r_matched.include?(i[:reference])}
+			elsif m.length > 1
 				m_sel = m.max{|a, b| comp.call(a, b)}
 				m.replace([m_sel])
 				s_matched << m_sel[:study]

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: pubannotation_evaluator
 version: !ruby/object:Gem::Version
-  version: 1.0.2
+  version: 1.1.0
 platform: ruby
 authors:
 - Jin-Dong Kim
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-04-30 00:00:00.000000000 Z
+date: 2021-12-19 00:00:00.000000000 Z
 dependencies: []
 description: A tool to evaluate the accuracy of a set of annotations.
 email: jdkim@dbcls.rois.ac.jp
@@ -39,8 +39,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.7.9
+rubygems_version: 3.0.9
 signing_key: 
 specification_version: 4
 summary: It compares a set of annotations (study annotations) against another set