dwc_agent 1.0.0 → 1.4.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 69d6e0f6c0f9d59801d0e53a89dd73896f0ac5a8230156879227ea6efe14bb7d
-  data.tar.gz: 1330921589cbcad22273c4b5cae821eba7266cc4357756498ae3b6d4f029e3f8
+  metadata.gz: e2819869504dd7262508c63978dae9b03cd6380838bf20f1e5ffecb8464b610a
+  data.tar.gz: 3d1afff77e4bea9995cec5a3f72687646cdb6c9f7637bef8444f7a45df4ab556
 SHA512:
-  metadata.gz: 735057f853b259f4d7fcb9f4c73d2f193f9a0e363d439a8bf894bbd9f878e8667cd14979c10cf138c7cc007f83cd09b7d5f9f21e0482b6303f73a40c788a957a
-  data.tar.gz: 8760dc0d9976543fe6ce448e2129b94e2d86d5a17f8e0b4b297f4dd23e8c4e4d73b4c7af4863e5edfa6abbbdb83181e6364e8064b0c73f185b0370c3ea1f597c
+  metadata.gz: 223cdf7a80a54e102da1f35d0b98d420d5a3b4ca1a11494ac3feca73a80d86f83a092b3c8b5dac1e5079a56cb5c69c3771876a091770cb4c9e45de0fc905d177
+  data.tar.gz: 042d702364e285c9e629180802ecd2f1898af1c50ec6fbdcacefa8f117d734db7843fd17e0f816771fcef47c7839ccfa8014d963f6798eaec25e7a3854b88289

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

@@ -22,10 +22,6 @@ module DwcAgent
         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

data/lib/dwc_agent/constants.rb

@@ -159,7 +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{
@@ -198,11 +198,13 @@ module DwcAgent
     (?i:no\s+(agent)?\s?(data|disponible)(\s+available)?)|
     (?i:not?\s+(entered|stated))|
     (?i:nomenclatur(e|al)\s+adjustment)|
+    (?i:not\s+available)|
     (?i:ontario|qu(e|é)bec|saskatchewan|new brunswick|sault|newfoundland|assurance|vancouver|u\.?s\.?s\.?r\.?)|
     (?i:recreation|culture)|
     (?i:shaped|dark|pale|areas|phase|spotting|interior|between|closer)|
     (?i:soci(e|é)t(y|é)|cent(er|re)|community|history|conservation|conference|assoc|class|commission|consortium|council|club|exposit|alliance|protective|circle)|
     (?i:commercial|company|control|product)|
+    (?i:sequence\s+data)|
     (?i:size|large|colou?r)\s+|
     (?i:skeleton)|
     (?i:survey|assessment|station|monitor|stn\.|index|project|bureau|engine|exchange|ex(c|k)ursi(e|o|ó)n?|exped\.?|exp(e|i)di(c|t)i(e|o|ó)n?|experiment|explora(d|t)|festival|generation|inventory|marine|service)|
@@ -213,6 +215,7 @@ module DwcAgent
     \s*(?i:too)\s+|\s*(?i:the)\s+|
     (?i:taxiderm(ies|y))|
     (?i:though)|
+    (?i:texas\s+instruments?)\s*?(for)?|
     (?:tropical)|
     (?i:toward|seen at)|
     (?i:unidentified|unspecified|unk?nown|unnamed|unread|unmistak|no agent)|
@@ -230,7 +233,12 @@ module DwcAgent
     "of",
     "curators",
     "nomenclatural",
-    "adjustment"
+    "adjustment",
+    "available",
+    "data",
+    "orig",
+    "science",
+    "catalogue"
   ]
 
   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

@@ -2,8 +2,8 @@ module DwcAgent
   class Version
 
     MAJOR = 1
-    MINOR = 0
-    PATCH = 0
+    MINOR = 4
+    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: 1.0.0
+  version: 1.4.1
 platform: ruby
 authors:
 - David P. Shorthouse
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-09-29 00:00:00.000000000 Z
+date: 2019-11-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: namae