dwc_agent 0.2.1 → 0.2.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: c955cb02feb025b039612097594c1bbb2dc94647349dcf3dd566ed9111314d88
-  data.tar.gz: 37749c3f89b1d11334362e8be70ce5a996d66285eb637ae0a4a6bd9866674f02
+  metadata.gz: 5175936a78d6dc64f17a2827dccfaa80c66f07242b17639b3668e6403e9a19b7
+  data.tar.gz: a13ead08f2756db93e5abaff55d2cd9316b7ee0739926571045d52abfc4243c8
 SHA512:
-  metadata.gz: 956ab7dedbba9b87fd62294290ae3b14b8f35c19b7768d6318237cb61545e27dbe8cdd9d64000c72b3931b738446f03ef070ee6526876ae6e46cddf8b2a9f0bf
-  data.tar.gz: 15f9ae4819ea5180c3c046a4d763005d69999c20eab8e362051a4c5406bf0c1dc54389ede6f61462d6898911fc10ed2f5fc66942d30df0a716de10651ab33f9b
+  metadata.gz: 368ea5fd8755f19dafd4b651fc68ef89a0e34246deedd6261b4cd97ec3dbf6b3f4eb13bcc3e7552ce46d77f25b721e9269e4082b861681486dfb12a797e8c254
+  data.tar.gz: f61d5cc267f5b210a327431eec44da687d03d991455e6c7797aa6d1aac1fca47c2664e8579143e46951c1413cafb068acc5d8fb13ad9c0c76442ccc578581833

data/bin/dwcagent

@@ -7,7 +7,7 @@ require 'json'
 names = []
 DwcAgent.parse(ARGV[0]).each do |r|
   name = DwcAgent.clean(r)
-  if !name[:family].nil? && name[:family].length >= 3
+  if !name[:family].nil? && name[:family].length >= 2
     names << name
   end
 end

data/lib/dwc_agent/cleaner.rb

@@ -7,6 +7,9 @@ module DwcAgent
       end
     end
 
+    def initialize
+    end
+
     # Cleans the passed-in namae object from the parse method and
     # re-organizes it to better match expected Darwin Core output.
     #
@@ -45,7 +48,7 @@ module DwcAgent
         (parsed_namae.given == parsed_namae.given.upcase || 
         parsed_namae.given == parsed_namae.given.downcase) && 
         !parsed_namae.given.include?(".") &&
-        parsed_namae.given.gsub(".","").length >= 4
+        parsed_namae.given.tr(".","").length >= 4
           parsed_namae.given = NameCase(parsed_namae.given)
       end
 

data/lib/dwc_agent/constants.rb

@@ -50,7 +50,7 @@ module DwcAgent
     (?i:autres?\s+de|probab|likely|possibl(e|y)|doubtful)|
     \b\s*(?i:maybe)\s*\b|
     \b\s*(?i:prob)\.\s*\b|
-    \(?(?i:collector|data\s*recorder|netter|(oper|prepar)ator)\(?s?\)?\.?\:?|
+    \(?[,;]?\s*?(?i:(local)?\s?collector|data\s*recorder|netter|(oper|prepar)ator)\(?s?\)?\.?\:?|
     (?i:fide)\:?\s*\b|
     (?i:game\s+dept)\.?\s*\b|
     (?i:see\s+notes?\s*(inside)?)|

data/lib/dwc_agent/parser.rb

@@ -7,6 +7,15 @@ module DwcAgent
       end
     end
 
+    def initialize
+      options = { 
+        prefer_comma_as_separator: true,
+        separator: SPLIT_BY,
+        title: TITLE
+      }
+      @namae = Namae::Parser.new(options)
+    end
+      
     # Parses the passed-in string and returns a list of names.
     #
     # @param names [String] the name or names to be parsed
@@ -14,20 +23,15 @@ module DwcAgent
     def parse(name)
       return [] if name.nil? || name == ""
       residual_terminators_regex = Regexp.new SPLIT_BY.to_s + %r{\s*\z}.to_s
-      cleaned = name.gsub(STRIP_OUT, ' ')
-                    .gsub(/[#{CHAR_SUBS.keys.join('\\')}]/, CHAR_SUBS)
-                    .gsub(/(#{PHRASE_SUBS.keys.join('|')})/, PHRASE_SUBS)
-                    .gsub(/([A-Z]{1}\.)([[:alpha:]]{2,})/, '\1 \2')
-                    .gsub(COMPLEX_SEPARATORS, '\1 | \2')
-                    .gsub(residual_terminators_regex, '')
-                    .squeeze(' ').strip
-      options = { 
-        prefer_comma_as_separator: true,
-        separator: SPLIT_BY,
-        title: TITLE
-      }
-      namae = Namae::Parser.new(options)
-      namae.parse(cleaned)
+      name.gsub!(STRIP_OUT, ' ')
+      name.gsub!(/[#{CHAR_SUBS.keys.join('\\')}]/, CHAR_SUBS)
+      name.gsub!(/(#{PHRASE_SUBS.keys.join('|')})/, PHRASE_SUBS)
+      name.gsub!(/([A-Z]{1}\.)([[:alpha:]]{2,})/, '\1 \2')
+      name.gsub!(COMPLEX_SEPARATORS, '\1 | \2')
+      name.gsub!(residual_terminators_regex, '')
+      name.squeeze!(' ')
+      name.strip!
+      @namae.parse(name)
     end
 
   end

data/lib/dwc_agent/similarity.rb

@@ -7,6 +7,9 @@ module DwcAgent
       end
     end
 
+    def initialize
+    end
+
     # Produces a similarity score of two given names
     # Logic inspired by R.D.M. Page, https://orcid.org/0000-0002-7101-9767
     # At https://linen-baseball.glitch.me/
@@ -15,10 +18,12 @@ module DwcAgent
     # @param given2 [String] a second given name
     # @return [Float] the similarity score
     def similarity_score(given1, given2)
-      given1_parts = given1.gsub(/\.\s+/,".").split(/[\.\s]/)
-      given2_parts = given2.gsub(/\.\s+/,".").split(/[\.\s]/)
-      largest = [given1_parts,given2_parts].max
-      smallest = [given1_parts,given2_parts].min
+      given1.gsub!(/\.\s+/,".")
+      g1_arr = given1.split(/[\.\s]/)
+      given2.gsub!(/\.\s+/,".")
+      g2_arr = given2.split(/[\.\s]/)
+      largest = [g1_arr,g2_arr].max
+      smallest = [g1_arr,g2_arr].min
 
       score = 0
       largest.each_with_index do |val,index|

data/lib/dwc_agent/version.rb

@@ -3,7 +3,7 @@ module DwcAgent
 
     MAJOR = 0
     MINOR = 2
-    PATCH = 1
+    PATCH = 2
     BUILD = nil
 
     def self.version

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: dwc_agent
 version: !ruby/object:Gem::Version
-  version: 0.2.1
+  version: 0.2.2
 platform: ruby
 authors:
 - David P. Shorthouse
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-06-16 00:00:00.000000000 Z
+date: 2019-07-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: namae