dwc_agent 0.4.3 → 1.3.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: c4ea46a2cca2719aebba4a99251aaf02f6d2fb36f21f3e6ea28b76584fc7345a
-  data.tar.gz: 72cde7bbdf5c8f93923710f887f299a4618e32c5d129e8cac0bbcc1a285492fd
+  metadata.gz: 89ee1173b349dee8f3ef6fad429449acf1fb3816ae49972b0fb35e7eb9a37720
+  data.tar.gz: c4dc07d7b1e113835ca065f17cb86e394a313b5b1fc335c2bdbca7ec8bafbdb3
 SHA512:
-  metadata.gz: 77c1027c302f5b853641266a833d197c1d81045ffad0ca0a2b3f4161d1fc4443fff944e6493d9c1089e710ac5aaff33bd0075698c7b70ff0ddf90d79602c8442
-  data.tar.gz: 9b92955bf9421e4b5f7a0c5cc5fa04265a393f04de6d7f23217cea739732b81bd8f9fb9042c02db20fe13118f2c9f3b75e1a374787b83d8e0ee1350e3e2a6c7b
+  metadata.gz: b5c450fcb300b84bf942643ed527658840c8c2981cc4d3766a0ff9c3d4f4f5e27c940342f7a344ca502029f0b51e61b1580a08b8c6a68a19db55279cfe52fba6
+  data.tar.gz: 02a17d7b41313802dc6e14efbb0f54604768a53c7d5637114204d058112748189809fc20b557b772cf1650045c4856ec6ba49bac162a28d3768e729cf872092c

data/bin/dwcagent

@@ -5,7 +5,7 @@ require 'dwc_agent'
 require 'json'
 
 names = []
-DwcAgent.parse(ARGV[0]).each do |r|
+DwcAgent.parse(ARGV[0].dup).each do |r|
   name = DwcAgent.clean(r)
   if !name[:family].nil? && name[:family].length >= 2
     names << name

data/bin/dwcagent-similarity

@@ -3,4 +3,4 @@
 
 require 'dwc_agent'
 
-puts DwcAgent.similarity_score(ARGV[0],ARGV[1])
+puts DwcAgent.similarity_score(ARGV[0].dup,ARGV[1].dup)

data/lib/dwc_agent/cleaner.rb

@@ -16,16 +16,12 @@ module DwcAgent
     # @param parsed_namae [Object] the namae object
     # @return [Hash] the given, family hash
     def clean(parsed_namae)
-      blank_name = { given: nil, family: nil }
+      blank_name = { given: nil, family: nil, particle: nil }
 
       if parsed_namae.family && FAMILY_BLACKLIST.any?{ |s| s.casecmp(parsed_namae.family) == 0 }
         return blank_name
       end
 
-      if parsed_namae.family && parsed_namae.family.length < 2 && parsed_namae.family.count('.') == 0
-        return blank_name
-      end
-
       if parsed_namae.family && parsed_namae.family.length == 3 && parsed_namae.family.count('.') == 1
         return blank_name
       end
@@ -99,6 +95,10 @@ module DwcAgent
         particle = nil
       end
 
+      if !particle.nil? && particle.include?(".")
+        particle = nil
+      end
+
       if !family.nil? && (family == family.upcase || family == family.downcase)
         family = NameCase(family)
       end
@@ -115,7 +115,7 @@ module DwcAgent
         return blank_name
       end
 
-      { given: given, family: family }
+      { given: given, family: family, particle: particle }
     end
 
   end

data/lib/dwc_agent/constants.rb

@@ -81,6 +81,7 @@ module DwcAgent
     AFSC\/POLISH\s+SORTING\s+CTR\.?|
     (?i:university|museum|exhibits?)|
     (?i:uqam)|
+    (?i:sem\s+(colec?tor|data))|
     \b[,;]\s+\d+\z|
     ["!@?]|
     [,]?\d+|
@@ -158,7 +159,7 @@ module DwcAgent
   }
 
   COMPLEX_SEPARATORS = %r{
-    ^([A-Za-z]{4,},\s+(?:[A-Z]\.\s*){1,})\s+([A-Za-z]{4,},\s+(?:[A-Z]\.\s*){1,})$
+    ^(\S{4,},\s+(?:\S\.\s*){1,})\s+(\S{4,},\s+(?:\S\.\s*){1,})$
   }x
 
   BLACKLIST = %r{
@@ -227,7 +228,9 @@ module DwcAgent
     "von",
     "the",
     "of",
-    "curators"
+    "curators",
+    "nomenclatural",
+    "adjustment"
   ]
 
   TITLE = /\s*\b(sir|count(ess)?|(gen|adm|col|maj|capt|cmdr|lt|sgt|cpl|pvt|prof|dr|md|ph\.?d|rev|docteur|mme|abbé|ptre)\.?|frère|frere|père|pere|professor|esq\.?)(\s+|$)/i

data/lib/dwc_agent/parser.rb

@@ -20,7 +20,7 @@ module DwcAgent
       @char_subs_regex = Regexp.new [CHAR_SUBS.keys.join('\\')].to_s
       @phrase_subs_regex = Regexp.new (PHRASE_SUBS.keys.join('|')).to_s
       @complex_separators_regex = Regexp.new COMPLEX_SEPARATORS.to_s
-      @add_separators_regex = Regexp.new %r{([A-Z]{1}\.)([[:alpha:]]{2,})}.to_s
+      @add_separators_regex = Regexp.new %r{(\S{1}\.)([[:alpha:]]{2,})}.to_s
     end
       
     # Parses the passed-in string and returns a list of names.

data/lib/dwc_agent/version.rb

@@ -1,9 +1,9 @@
 module DwcAgent
   class Version
 
-    MAJOR = 0
-    MINOR = 4
-    PATCH = 3
+    MAJOR = 1
+    MINOR = 3
+    PATCH = 1
     BUILD = nil
 
     def self.version

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: dwc_agent
 version: !ruby/object:Gem::Version
-  version: 0.4.3
+  version: 1.3.1
 platform: ruby
 authors:
 - David P. Shorthouse
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-09-27 00:00:00.000000000 Z
+date: 2019-11-05 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: namae