dwc_agent 3.0.8.0 → 3.0.9.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: '087733d2b0fb769363b4075c7f41c2f66029babdd94369754d948fd7ea79a918'
-  data.tar.gz: 4ac264b58351da0cebf5d0e5139e61fafb0c3a2ce7230dc360c5d75498251c03
+  metadata.gz: c0d83fd4dba1ddb6b9976bb1dea4c95a5365cf31b9c1afc335e035f70317a40f
+  data.tar.gz: 2fc6de4a6de283d9cf8d813306bb034ca67e20e684ce6136d6f2366512633d34
 SHA512:
-  metadata.gz: 5df9637901a40e47c22872811af6a71a2e482042f3615bf263897945fdadff0254bd373b549f63a0a989e58dfee1d9ffc1e7189da271720badf51596188fc3ae
-  data.tar.gz: d07083d5a2a439eabd9bb1615de318f482ce67b15bce86e32c823a52f6e0c9615373f587f1f66c9b14ebd5bb91028d18b21b95c9771724d6bf7570037981f723
+  metadata.gz: d6ca8f0a7507c8a8d09b183f59ca47d97ff30a0de3e017da155e0b1e57ae5d34df4328ae9bdbb2fd4682b0d57afac7e251baebaa305fea5bc46b2886b22e5385
+  data.tar.gz: b9ad4777ade9052a3cd54926173c77973f899e76b3f65f43f31b45481de746e70daa4ecdcbb0b190567c27f3496d69cb5093d35522243b284ac5d4931fbe09a0

data/lib/dwc_agent/constants.rb

@@ -161,6 +161,10 @@ module DwcAgent
     \b(?i:via|from)\s*\b
   }x
 
+  POST_STRIP_TIDY = %r{
+    ^\s*[&,;]\s*
+  }x
+
   CHAR_SUBS = {
     '"' => '\'',
     '|' => ' | ',
@@ -210,7 +214,7 @@ module DwcAgent
     "^(\\S{4,},\\s+(?:\\S\\.\\s*){1,})\\s+(\\S{4,},\\s+(?:\\S\.\\s*){1,})$" => "\\1 | \\2",
     "(\\S{1}\\.)([[:alpha:]]{2,})" => "\\1 \\2",
     "^([[:alpha:]]{2,})(?:\\s+)((?:\\S{1}\\.\\s?){1,})$" => "\\1, \\2",
-    "([[:alpha:]]*),?\\s+(.*)\\s+(van|von)$" => "\\3 \\1, \\2",
+    "([[:alpha:]]*),?\\s*(.*)\\s+(van|von)$" => "\\3 \\1, \\2",
     "^([A-Z.\\s]+)\\s+(?:and|&|et|e)\\s+([A-Z.\\s]+)\\s+([[:alpha:]]{2,})\\s+([[:alpha:]]{2,})$" => "\\1 \\4 | \\2 \\3 \\4",
     "^([A-Z.\\s]+)\\s+(?:and|&|et|e)\\s+([A-Z.\\s]+)\\s+([[:alpha:]]{2,})(.*)$" => "\\1 \\3 | \\2 \\3 | \\4",
     "^([A-Z][[:alpha:]]{2,}),\\s*?([A-Z][[:alpha:]]{2,})\\s*?(?i:and|&|et|e|,)\\s+([A-Z][[:alpha:]]{2,})$" => "\\1 | \\2 | \\3",
@@ -222,7 +226,7 @@ module DwcAgent
     (?i:abundant)|
     (?i:adult|juvenile)|
     (?i:administra(d|t)or)|
-    (?i:anon)|
+    ^(?i:anon)$|
     (?i:australian?)|
     (?i:average)|
     (?i:believe|unclear|ill?egible|none|suggested|(dis)?agrees?)|approach|

data/lib/dwc_agent/parser.rb

@@ -18,6 +18,7 @@ module DwcAgent
       }
       @namae = Namae::Parser.new(options)
       @strip_out_regex = Regexp.new STRIP_OUT.to_s
+      @tidy_remains_regex = Regexp.new POST_STRIP_TIDY.to_s
       @char_subs_regex = Regexp.new [CHAR_SUBS.keys.join].to_s
       @phrase_subs_regex = Regexp.new PHRASE_SUBS.keys.map{|a| Regexp.escape a }.join('|').to_s
       @residual_terminators_regex = Regexp.new SPLIT_BY.to_s + %r{\s*\z}.to_s
@@ -31,6 +32,7 @@ module DwcAgent
     def parse(name)
       return [] if name.nil? || name == ""
       name.gsub!(@strip_out_regex, ' ')
+      name.gsub!(@tidy_remains_regex, '')
       name.gsub!(Regexp.union(@char_subs_regex, @phrase_subs_regex), CHAR_SUBS.merge(PHRASE_SUBS))
       @separators.each{|k| name.gsub!(k[0], k[1])}
       name.gsub!(@residual_terminators_regex, '')

data/lib/dwc_agent/version.rb

@@ -4,7 +4,7 @@ module DwcAgent
 
     MAJOR = 3
     MINOR = 0
-    PATCH = 8
+    PATCH = 9
     BUILD = 0
 
     def self.version

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: dwc_agent
 version: !ruby/object:Gem::Version
-  version: 3.0.8.0
+  version: 3.0.9.0
 platform: ruby
 authors:
 - David P. Shorthouse
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-08-07 00:00:00.000000000 Z
+date: 2023-09-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: namae