dwc_agent 1.4.6 → 1.4.11

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 1cf058ee8ca5956a4d05ca0122d5ee9a7497b304576bd4af92fd0f0597846a08
-  data.tar.gz: 3ef7bafb6a3bd9b9af826f161448b259739ec791b4158e5704f6d4da98af48fe
+  metadata.gz: b7f6e68d702ce0526d2e03c6299cf6b4683a85c0539aa1637b1d8d12e2e0bc05
+  data.tar.gz: c47fa534fec2f64c0b308971291a8c5d7fb637cb8487d782bbc8523ed9e5efa2
 SHA512:
-  metadata.gz: 0be4b1c78b8fbf61c1a5363ce32efa780372b394bf88e9fd5a5fc42be77eea564d865637feaca08b63712d8780c515db2408e01967519fc014a3b6f5468793c9
-  data.tar.gz: 37095ef2920f06e6152aa58aab26034ce5b00d15778a8ae3270c7672c7e53ba2853223f7f67bfbfc5953d06050b8f2a8c8a2a1f0c3e41ba002aa18411a57092c
+  metadata.gz: 782189bcf8e90a397ebcebb1c176bec521ef9df482e25cfdcd8084612ff45c3a644771088b58741317fd31b65241908045189988fe497e459a7ef66f12de54ca
+  data.tar.gz: 3be7a71b49471a2c7172ccc04d81947caf280f2d5621dc634fb7cc45ba853d2395126613d4fd5f3800251d31ce2bd6627fb1820de9cad5e76f34730522a2daf4

data/lib/dwc_agent/constants.rb

@@ -2,9 +2,10 @@ module DwcAgent
   STRIP_OUT = %r{
     ^[\[{(]|
     [\]})]\??$|
+    (?i:acc\s?\#)|
     \s*?\d+\.\d+|
     \b\d+\(?(?i:[[:alpha:]])\)?\b|
-    \b[,;]?\s*(?i:et\.?\s+al)\.?|
+    \b[,;]?\s*(?i:et\.?\s+al|&\s+al)\.?|
     \b\s+(bis|ter)(\b|\z)|
     \bu\.\s*a\.|
     \b[,;]?\s*(?i:and|&)?\s*(?i:others)\s*\b|
@@ -13,7 +14,7 @@ module DwcAgent
     \b[,;]?\s*(?i:unkn?own)\b|
     \b[,;]?\s*(?i:n/a)\b|
     \b[,;]?\s*(?i:ann?onymous)\b|
-    \b[,;]?\s*\(?(?i:undetermined|indeterminable|dummy|interim|accession|ill(eg|is)ible|scripsit)\)?\b|
+    \b[,;]?\s*\(?(?i:undetermined|indeterminable|dummy|interim|accession|ill(eg|is)ible|scripsit|presumed?)\)?\b|
     \b[,;]?\s*(?i:importer|gift)\:?\b|
     \b[,;]?\s*(?i:string)\b|
     \b[,;]?\s*(?i:person\s*string)\b|
@@ -53,6 +54,7 @@ module DwcAgent
     \b\s*(?i:prob)\.\s*\b|
     \(?[,]?\s*?(?i:(local)?\s?collector|data\s*recorder|netter|(oper|prepar)ator)\(?s?\)?\.?\:?|
     \b[.-–,;:]?\s*(?i:department|faculty)\s*?(?i:of)?\s*?(?i:entomology|biology|zoology)|
+    (?i:Engº|Agrº|Fcº|Drº|Mº|Profº|Dº|Fº)|
     (?i:fide)\:?\s*\b|
     (?i:game\s+dept)\.?\s*\b|
     (?i:see\s+notes?\s*(inside)?)|
@@ -71,7 +73,7 @@ module DwcAgent
     (?i:Field\s+Museum\s+of\s+Natural\s+History)|
     (?i:American\s+Museum\s+of\s+Natural\s+History)|
     (?i:The\s+Paleontological\s+Research\s+Institution)|
-    (?i:museum\s+victoria)|
+    (?i:museums?\s+victoria)|
     \b\s*(?i:United\s+States|Russia)\s*\b|
     (?i:revised|photograph|fruits\s+only)|
     -?\s*(?i:sight\s+(id|identifi?cation))\.?\s*\b|
@@ -79,7 +81,7 @@ module DwcAgent
     \b\s*\(?(?i:(fe)?male)\)?\s*\b|
     \b(?i:to\s+(sub)?spp?)\.?|
     (?i:nom\.?\s+rev\.?)|
-    FNA|DAO|HUH|FDNMB|MNHN|PNI|USNM|ZMUC|CSIRO|
+    FNA|DAO|HUH|FDNMB|MNHN|PNI|USNM|ZMUC|CSIRO|ACAD|USGS|NAWQA|
     AFSC\/POLISH\s+SORTING\s+CTR\.?|
     (?i:university|museum|exhibits?)|
     (?i:uqam)|
@@ -107,7 +109,7 @@ module DwcAgent
   }x
 
   SPLIT_BY = %r{
-    [–|&+/;:]|
+    [–|ǀ∣｜│&+\/;:]|
     \s+-\s+|
     \s+a\.\s+|
     \b(e|y|i|en|et|or|per|for)\s*\b|
@@ -134,6 +136,9 @@ module DwcAgent
 
   CHAR_SUBS = {
     '|' => ' | ',
+    'ǀ' => ' | ',
+    '∣' => ' | ',
+    '│' => ' | ',
     '(' => ' ',
     ')' => ' ',
     '?' => '',
@@ -156,9 +161,13 @@ module DwcAgent
     'dr\.' => 'Dr. ',
     'mr\.' => 'Mr. ',
     'mrs\.' => 'Mrs. ',
+    'ms\.' => 'Ms. ',
     'prof\.' => 'Prof. ',
     '\, ph\.d\.' => ' Ph.D.',
-    '\, bro\.' => ' Bro.'
+    '\, bro\.' => ' Bro.',
+    ' jr\.,' => ' Jr.;',
+    ' jr,' => ' Jr.;',
+    '-Jr' => ' Jr.'
   }
 
   COMPLEX_SEPARATORS = %r{
@@ -174,6 +183,7 @@ module DwcAgent
     (?i:average)|
     (?i:believe|unclear|ill?egible|none|suggested|(dis)?agrees?)|approach|
     (?i:barcod)|
+    (?i:BgWd)|
     (?i:biolog|botan|zoo|ecolog|mycol|(in)?vertebrate|fisheries|genetic|animal|mushroom|wildlife|plumage|flower|agriculture)|
     (?i:bris?tish|canadi?an?|chinese|arctic|japan|russian|north\s+america)|
     (?i:carex|salix)|
@@ -196,7 +206,7 @@ module DwcAgent
     (?i:geographic)|
     (?i:mus(eum|ée)|universit(y|é|e|at)|college|institute?|acad(e|é)m|school|écol(e|iers?)|laboratoi?r|projec?t|polytech|dep(t|art?ment)|research|clinic|hospital|cientifica|sanctuary|safari)|
     (?i:univ\.)|
-    (?i:graduate|student|élèves?|éleveur|étudiants|estudi?antes?|labo\.|storekeep|supervisor|superint|rcmp|coordinator|minority|fishermen|police|taxonomist|consultant|participante?s?|team|(é|e)quipe|memb(er|re)|crew|group|staff|personnel|family|captain|friends|assistant|worker)|
+    (?i:graduate|student|élèves?|éleveur|étudiants|estudi?antes?|labo\.|storekeep|supervisor|superint|rcmp|coordinator|minority|fishermen|police|taxonomist|consultant|participante?s?|team|(é|e)quipe|memb(er|re)|crew|group|staff|personnel|family|captain|friends|assistant|worker|gamekeeper)|
     (?i:non\s+pr(é|e)cis(é|e))|
     (?i:no\s+consta)|
     (?i:no\s+(agent)?\s?(data|disponible)(\s+available)?)|
@@ -223,7 +233,7 @@ module DwcAgent
     (?i:texas\s+instruments?)\s*?(for)?|
     (?:tropical)|
     (?i:toward|seen at)|
-    (?i:unidentified|unspecified|unk?nown|unnamed|unread|unmistak|no agent)|
+    (?i:unidentified|unspecified|unk?nown?|unnamed|unread|unmistak|no agent)|
     (?i:urn\:)|
     (?i:usda|ucla)|
     (?i:workshop|garden|farm|jardin|public)|
@@ -261,6 +271,6 @@ module DwcAgent
     "has not"
   ]
 
-  TITLE = /\s*\b(sir|count(ess)?|(gen|adm|col|maj|capt|cmdr|lt|sgt|cpl|pvt|prof|dr|md|ph\.?d|rev|mme|abbé|ptre|bro|esq)\.?|docteur|father|cantor|vicar|père|pastor|rabbi|reverend|pere|soeur|sister|professor)(\s+|$)/i
+  TITLE = /\s*\b(sir|count(ess)?|colonel|(gen|adm|col|maj|capt|cmdr|lt|sgt|cpl|pvt|prof|dr|md|ph\.?d|rev|mme|abbé|ptre|bro|esq)\.?|docteur|father|cantor|vicar|père|pastor|rabbi|reverend|pere|soeur|sister|professor)(\s+|$)/i
 
 end

data/lib/dwc_agent/version.rb

@@ -3,7 +3,7 @@ module DwcAgent
 
     MAJOR = 1
     MINOR = 4
-    PATCH = 6
+    PATCH = 11
     BUILD = nil
 
     def self.version

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: dwc_agent
 version: !ruby/object:Gem::Version
-  version: 1.4.6
+  version: 1.4.11
 platform: ruby
 authors:
 - David P. Shorthouse
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-02-11 00:00:00.000000000 Z
+date: 2020-07-27 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: namae
@@ -98,7 +98,7 @@ files:
 - lib/dwc_agent/similarity.rb
 - lib/dwc_agent/utility.rb
 - lib/dwc_agent/version.rb
-homepage: https://github.com/dshorthouse/dwc_agent
+homepage: https://github.com/bionomia/dwc_agent
 licenses:
 - MIT
 metadata: {}