biodiversity 0.7.0 → 0.7.1

data/VERSION

@@ -1 +1 @@
-0.7.0
+0.7.1

data/lib/biodiversity/parser.rb

@@ -50,6 +50,10 @@ class ScientificNameParser
     !!(a_string.match(/\sICTV\s*$/) || a_string.match(/\s(virus|phage|viroid|satellite|prion)\b/i))
   end 
 
+  def unknown_placement?(a_string)
+    !!(a_string.match(/incertae\s+sedis/i) || a_string.match(/inc\.\s*sed\./i))
+  end
+
   def parsed
     @parsed
   end
@@ -60,6 +64,8 @@ class ScientificNameParser
     
     if virus?(a_string)
       @parsed = { :verbatim => a_string, :virus => true }
+    elsif unknown_placement?(a_string)
+      @parsed = { :verbatim => a_string }
     else
       @parsed = @clean.parse(a_string) || @dirty.parse(a_string) || @canonical.parse(a_string) || { :verbatim => a_string }
     end

data/spec/parser/test_data.txt

@@ -245,6 +245,15 @@ Achillea bonarota nom. in herb.|{"scientificName":{"parsed":true, "verbatim":"Ac
 Aconitum napellus var. formosum (Rchb.) W. D. J. Koch (nom. ambig.)|{"scientificName":{"parsed":true, "verbatim":"Aconitum napellus var. formosum (Rchb.) W. D. J. Koch (nom. ambig.)", "normalized":"Aconitum napellus var. formosum (Rchb.) W. D. J. Koch", "canonical":"Aconitum napellus formosum", "hybrid":false, "details":[{"genus":{"string":"Aconitum"}, "species":{"string":"napellus"}, "infraspecies":[{"string":"formosum", "rank":"var.", "authorship":"(Rchb.) W. D. J. Koch", "combinationAuthorTeam":{"authorTeam":"W. D. J. Koch", "author":["W. D. J. Koch"]}, "basionymAuthorTeam":{"authorTeam":"Rchb.", "author":["Rchb."]}}]}],  "parser_version":"test_version", "parser_run":1, "positions":{"0":["genus", 8], "9":["species", 17], "23":["infraspecies", 31], "33":["author_word", 38], "40":["author_word", 42], "43":["author_word", 45], "46":["author_word", 48], "49":["author_word", 53]}}}
 × Dialaeliopsis Hort.|{"scientificName":{"parsed":true, "verbatim":"× Dialaeliopsis Hort.", "normalized":"× Dialaeliopsis", "canonical":"Dialaeliopsis", "hybrid":true, "details":[{"uninomial":{"string":"Dialaeliopsis"}}],  "parser_version":"test_version", "parser_run":1, "positions":{"2":["uninomial", 15]}}}
 
+#should not parse incertae sedis
+
+Incertae sedis|{"scientificName":{"parsed":false, "parser_version":"test_version", "verbatim":"Incertae sedis"}}
+incertae sedis|{"scientificName":{"parsed":false, "parser_version":"test_version", "verbatim":"incertae sedis"}}
+Inc.   sed.|{"scientificName":{"parsed":false, "parser_version":"test_version", "verbatim":"Inc.   sed."}}
+inc.sed.|{"scientificName":{"parsed":false, "parser_version":"test_version", "verbatim":"inc.sed."}}
+inc.   sed.|{"scientificName":{"parsed":false, "parser_version":"test_version", "verbatim":"inc.   sed."}}
+
+
 #should not parse viruses
 Abutilon mosaic virus [X15983] [X15984] Abutilon mosaic virus ICTV|{"scientificName":{"parsed":false, "parser_version":"test_version", "verbatim":"Abutilon mosaic virus [X15983] [X15984] Abutilon mosaic virus ICTV", "virus":true}}
 Acute bee paralysis virus [AF150629] Acute bee paralysis virus|{"scientificName":{"parsed":false, "parser_version":"test_version", "verbatim":"Acute bee paralysis virus [AF150629] Acute bee paralysis virus", "virus":true}}

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: biodiversity
 version: !ruby/object:Gem::Version 
-  hash: 3
+  hash: 1
   prerelease: false
   segments: 
   - 0
   - 7
-  - 0
-  version: 0.7.0
+  - 1
+  version: 0.7.1
 platform: ruby
 authors: 
 - Dmitry Mozzherin
@@ -15,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2010-09-11 00:00:00 -04:00
+date: 2010-09-17 00:00:00 -04:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency