dwc_agent 1.4.5 → 1.4.10

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7b8890828a04d30f7989f3545df032be216d1bab086f1ec6576cbe93569231bd
-  data.tar.gz: d94e120615540e65f29a13132d46b34639e3380e2265439fd05cf5febc580ff6
+  metadata.gz: c9776505a303866cf6fb7ffed89c26f0d654c09bc89c0546de763084c55d483d
+  data.tar.gz: 1552777d4f2b4b8f6a72052a817dd7396e384f7b3f42b1dab1c4db7802dbfee7
 SHA512:
-  metadata.gz: f2e30d42a8c2d743791c2322a51d3fe022e5ef458f68ebd30763cfc74e5ec9a1914faca1ff90cfb3364bead66cde22bccbc0ab47aef6a7d17a78c6954c8cbeff
-  data.tar.gz: d2de0c26c64f77fd273db67465d8772aef5480eb0c4dd1b48446dbb6a6495164bc9e2fbd133280c18ef7e975f82bf9f46dcd65e87727531fa7c514ebc09e0be7
+  metadata.gz: ed8284bf695e107e8aa19af434376d7db3ad96235e1b5fc0f394a321edf527dd911ac9daa768630d4719e3ddcef0168077aa3c0a0d98180b29fc89cf116b4184
+  data.tar.gz: 67498c5b597e6ef3619a90e1b336a30c2840c33d71bb0c7a38b06a9db99d58e3caadcde0c292d65b88146dcd2a17be6ac4e8f5b77916ac578083e0ea841fb498

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|
@@ -67,11 +68,11 @@ module DwcAgent
     ORCID|
     MRI(\s|-)PAS|
     urn\:qm\.qld\.gov\.au\:collector|
-    (?i:University\s+of\s+California)\,?\s+?(?i:Berkeley)?|
+    (?i:University\s+of\s+(Southern\s+)?California(,\s+Berkeley)?)|
     (?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 +80,7 @@ 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|
+    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 +108,7 @@ 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 +135,9 @@ module DwcAgent
 
   CHAR_SUBS = {
     '|' => ' | ',
+    'ǀ' => ' | ',
+    '∣' => ' | ',
+    '│' => ' | ',
     '(' => ' ',
     ')' => ' ',
     '?' => '',
@@ -174,6 +178,7 @@ 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)|
@@ -196,8 +201,9 @@ 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)?)|
     (?i:not?\s+(entered|stated))|
     (?i:nomenclatur(e|al)\s+adjustment)|
@@ -222,7 +228,7 @@ 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)|
@@ -242,6 +248,7 @@ module DwcAgent
     "of",
     "adjustment",
     "available",
+    "arachnology",
     "catalogue",
     "curators",
     "data",

data/lib/dwc_agent/version.rb

@@ -3,7 +3,7 @@ module DwcAgent
 
     MAJOR = 1
     MINOR = 4
-    PATCH = 5
+    PATCH = 10
     BUILD = nil
 
     def self.version

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: dwc_agent
 version: !ruby/object:Gem::Version
-  version: 1.4.5
+  version: 1.4.10
 platform: ruby
 authors:
 - David P. Shorthouse
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-01-22 00:00:00.000000000 Z
+date: 2020-06-15 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: {}
@@ -119,7 +119,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.3
+rubygems_version: 3.0.6
 signing_key: 
 specification_version: 4
 summary: Parse Darwin Core agent terms such as recordedBy and identifiedBy