biodiversity 3.3.0 → 3.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: df9dbe0a3ecca9197bcd2bfe2a029308be56a490
-  data.tar.gz: d5ed6ae35b3c1e6f3a6ee7656bdbf9a1c1c78bd9
+  metadata.gz: bd706ba900f6f136fae256b85f55223e955b4300
+  data.tar.gz: 00b5ee4e9b8b34d7fcb5084ae4b82e6db8635522
 SHA512:
-  metadata.gz: 7ebe5b7d92c94636c71b594396215c81e70dd12bd40e83bb4871b8997f4620937fd573b3dedb4b6c2b5e5c03775982e55bd8006387ff298312f0de93e19a8ee3
-  data.tar.gz: 3308bb5e6e49b4dd215f9d92a7d1a91ff14764c5ef04ff401f29b11295352b12f82aaefb2cf2c2a10feedf03ab5e38dd54c9a6dead379e59ee206bae94487f78
+  metadata.gz: f34fb2b4e2d7e97f748967f9cec04625416b33789287ac6de7d63de348b676d4504b4c8a631505d7074989aff2d0597b857b0843b3f45cb4c5c3f5d61e7336ca
+  data.tar.gz: 8886e2fccfb35456c326e461c7f48d42ef3c5f76d2044219b8e63af2c690851f91cc4c536b2727ffe0684fdf0c006425322d6d7664b7dc7fe4717e02ed17031d

data/CHANGELOG

@@ -1,3 +1,6 @@
+3.4.0 -- added ScientificNameParser.add_rank_to_canonical(parsed) method.
+         This method allows to add infraspecific rank to a canonical form
+         after the fact of parsing.
 
 3.3.0 -- parserver gets new option -- -H for host, default 127.0.0.1
 

data/README.md

@@ -128,6 +128,15 @@ parser.parse("      Plantago       major    ")[:scientificName][:normalized]
 parser.parse("Pseudocercospora dendrobii (H.C. Burnett) U. \
 Braun & Crous 2003")[:scientificName][:canonical]
 
+# to get canonical form with infraspecies ranks
+parsed = parser.parse("Seddera latifolia Hochst. & Steud. var. latifolia")
+ranked = ScientificNameParser.add_rank_to_canonical(parsed)
+ranked[:scientificName][:canonical]
+#or
+parser = ScientificNameParser.new(canonical_with_rank: true)
+ranked = parser.parse("Seddera latifolia Hochst. & Steud. var. latifolia")
+ranked[:scientificName][:canonical]
+
 # to get detailed information about elements of the name
 parser.parse("Pseudocercospora dendrobii (H.C. Burnett 1883) U. \
 Braun & Crous 2003")[:scientificName][:details]

data/lib/biodiversity/parser.rb

@@ -115,6 +115,20 @@ class ScientificNameParser
     }
   end
 
+  def self.add_rank_to_canonical(parsed)
+    return parsed if parsed[:scientificName][:hybrid]
+    name = parsed[:scientificName]
+    parts = name[:canonical].split(" ")
+    name_ary = parts[0..1]
+    name[:details][0][:infraspecies].each do |data|
+      infrasp = data[:string]
+      rank = data[:rank]
+      name_ary << (rank && rank != "n/a" ? "#{rank} #{infrasp}" : infrasp)
+    end
+    parsed[:scientificName][:canonical] = name_ary.join(" ")
+    parsed
+  end
+
   def self.version
     Biodiversity::VERSION
   end
@@ -230,13 +244,14 @@ class ScientificNameParser
       else
         res.merge!(self)
       end
+      res[:surrogate] = true if ScientificNameParser.surrogate?(res)
+      res = {:scientificName => res}
       if (canonical_with_rank &&
           canonical.count(" ") > 1 &&
-          res[:details][0][:infraspecies])
+          res[:scientificName][:details][0][:infraspecies])
         ScientificNameParser.add_rank_to_canonical(res)
       end
-      res[:surrogate] = true if ScientificNameParser.surrogate?(res)
-      res = {:scientificName => res}
+      res
     end
 
     def @parsed.pos_json
@@ -270,15 +285,4 @@ class ScientificNameParser
                      name.match(surrogate2))
     is_surrogate
   end
-
-  def self.add_rank_to_canonical(parsed)
-    parts = parsed[:canonical].split(" ")
-    name_ary = parts[0..1]
-    parsed[:details][0][:infraspecies].each do |data|
-      infrasp = data[:string]
-      rank = data[:rank]
-      name_ary << (rank && rank != "n/a" ? "#{rank} #{infrasp}" : infrasp)
-    end
-    parsed[:canonical] = name_ary.join(" ")
-  end
 end

data/lib/biodiversity/version.rb

@@ -1,3 +1,3 @@
 module Biodiversity
-  VERSION = "3.3.0"
+  VERSION = "3.4.0"
 end

data/spec/parser/scientific_name_spec.rb

@@ -7,23 +7,23 @@ describe ScientificNameParser do
     set_parser(ScientificNameParser.new)
   end
 
-  it 'returns version number' do
+  it "returns version number" do
     expect(ScientificNameParser.version).to match /^\d+\.\d+\.\d+/
   end
 
-  it 'fixes cases' do
+  it "fixes cases" do
     names = [
-      ['QUERCUS ALBA', 'Quercus alba'],
-      ['QUERCUS (QUERCUS) ALBA', 'Quercus (Quercus) alba'],
-      ['QÜERCUS', 'Qüercus'],
-      ['PARDOSA MOéSTA', 'Pardosa moésta'],
+      ["QUERCUS ALBA", "Quercus alba"],
+      ["QUERCUS (QUERCUS) ALBA", "Quercus (Quercus) alba"],
+      ["QÜERCUS", "Qüercus"],
+      ["PARDOSA MOéSTA", "Pardosa moésta"],
     ]
     names.each do |name, capitalization|
       expect(ScientificNameParser::fix_case(name)).to eq capitalization
     end
   end
 
-  it 'generates standardized json' do
+  it "generates standardized json" do
     read_test_file do |y|
       expect(JSON.load(json(y[:name]))).to eq JSON.
         load(y[:jsn]) unless y[:comment]
@@ -31,9 +31,9 @@ describe ScientificNameParser do
   end
 
 
-  # it 'generates new test_file' do
+  # it "generates new test_file" do
   #   new_test = open(File.expand_path(dir +
-  #                    '../../spec/parser/test_data_new.txt'),'w')
+  #                    "../../spec/parser/test_data_new.txt"),"w")
   #   read_test_file do |y|
   #     if y[:comment]
   #       new_test.write y[:comment]
@@ -45,39 +45,39 @@ describe ScientificNameParser do
   #   end
   # end
 
-  it 'generates reasonable output if parser failed' do
-    sn = 'ddd sljlkj 3223452432'
+  it "generates reasonable output if parser failed" do
+    sn = "ddd sljlkj 3223452432"
     expect(json(sn)).to eq "{\"scientificName\":" \
       "{\"id\":\"3ebf93d9-b62a-5198-8715-4c8302f0a5d7\",\"parsed\":false," \
       "\"parser_version\":\"test_version\"," \
       "\"verbatim\":\"ddd sljlkj 3223452432\"}}"
   end
 
-  it 'shows version when the flag :show_version set to true' do
-    expect(parse('Homo sapiens')[:scientificName][:parser_version]).
+  it "shows version when the flag :show_version set to true" do
+    expect(parse("Homo sapiens")[:scientificName][:parser_version]).
       to_not be_nil
   end
 
-  it 'shows version for not spelled names' do
-    expect(parse('not_a_name')[:scientificName][:parser_version]).to_not be_nil
+  it "shows version for not spelled names" do
+    expect(parse("not_a_name")[:scientificName][:parser_version]).to_not be_nil
   end
 
-  it 'generates version for viruses' do
-    expect(parse('Nile virus')[:scientificName][:parser_version]).to_not be_nil
+  it "generates version for viruses" do
+    expect(parse("Nile virus")[:scientificName][:parser_version]).to_not be_nil
   end
 end
 
-describe 'ScientificNameParser with ranked canonicals' do
+describe "ScientificNameParser with ranked canonicals" do
   before(:all) do
     @parser = ScientificNameParser.new(canonical_with_rank: true)
   end
 
-  it 'does not influence output for uninomials and binomials' do
+  it "does not influence output for uninomials and binomials" do
     data = [
-      ['Ekbainacanthus Yakowlew 1902','Ekbainacanthus'],
-      ['Ekboarmia sagnesi herrerai Exposito 2007',
-       'Ekboarmia sagnesi herrerai'],
-      ['Ekboarmia holli Oberthür', 'Ekboarmia holli']]
+      ["Ekbainacanthus Yakowlew 1902","Ekbainacanthus"],
+      ["Ekboarmia sagnesi herrerai Exposito 2007",
+       "Ekboarmia sagnesi herrerai"],
+      ["Ekboarmia holli Oberthür", "Ekboarmia holli"]]
 
     data.each do |d|
       parsed = @parser.parse(d[0])[:scientificName][:canonical]
@@ -85,14 +85,14 @@ describe 'ScientificNameParser with ranked canonicals' do
     end
   end
 
-  it 'preserves rank for ranked multinomials' do
+  it "preserves rank for ranked multinomials" do
     data = [
-      ['Cola cordifolia var. puberula A. Chev.',
-       'Cola cordifolia var. puberula'],
-      ['Abies homolepis forma umbilicata (Mayr) Schelle',
-       'Abies homolepis forma umbilicata'],
-      ['Quercus ilex ssp. ballota (Desf.) Samp',
-       'Quercus ilex ssp. ballota'],
+      ["Cola cordifolia var. puberula A. Chev.",
+       "Cola cordifolia var. puberula"],
+      ["Abies homolepis forma umbilicata (Mayr) Schelle",
+       "Abies homolepis forma umbilicata"],
+      ["Quercus ilex ssp. ballota (Desf.) Samp",
+       "Quercus ilex ssp. ballota"],
       ["Physarum globuliferum forma. flavum Leontyev & Dudka",
        "Physarum globuliferum forma. flavum"]
     ]
@@ -101,16 +101,57 @@ describe 'ScientificNameParser with ranked canonicals' do
       expect(parsed).to eq d[1]
     end
   end
+end
+
+describe ".add_rank_to_canonical" do
+  subject(:parser) { ScientificNameParser.new }
 
+  it "adds rank to infraspecies with rank" do
+    data = [
+      ["Cola cordifolia var. puberula A. Chev.",
+       "Cola cordifolia puberula",
+       "Cola cordifolia var. puberula"],
+      ["Abies homolepis forma umbilicata (Mayr) Schelle",
+       "Abies homolepis umbilicata",
+       "Abies homolepis forma umbilicata"],
+      ["Quercus ilex ssp. ballota (Desf.) Samp",
+       "Quercus ilex ballota",
+       "Quercus ilex ssp. ballota"],
+      ["Physarum globuliferum forma. flavum Leontyev & Dudka",
+       "Physarum globuliferum flavum",
+       "Physarum globuliferum forma. flavum"]
+    ]
+    data.each do |d|
+      parsed = parser.parse(d[0])
+      canonical1 = parsed[:scientificName][:canonical]
+      expect(canonical1).to eq d[1]
+      ScientificNameParser.add_rank_to_canonical(parsed)
+      canonical2 = parsed[:scientificName][:canonical]
+      expect(canonical2).to eq d[2]
+    end
+  end
+
+  it "does not work for hybrids yet" do
+    data = [["Corda X cordiflora var. puberula",
+            "Corda cordiflora puberula"]]
+    data.each do |d|
+      parsed = parser.parse(d[0])
+      canonical1 = parsed[:scientificName][:canonical]
+      expect(canonical1).to eq d[1]
+      ScientificNameParser.add_rank_to_canonical(parsed)
+      canonical2 = parsed[:scientificName][:canonical]
+      expect(canonical2).to eq d[1]
+    end
+  end
 end
 
 describe ParallelParser do
-  it 'finds number of cpus' do
+  it "finds number of cpus" do
     pparser = ParallelParser.new
     expect(pparser.cpu_num).to be > 0
   end
 
-  it 'parses several names in parallel' do
+  it "parses several names in parallel" do
     names = []
     read_test_file { |n| names << (n[:name]) if n[:name] }
     names.uniq!
@@ -120,7 +161,7 @@ describe ParallelParser do
     expect(res.keys.size).to eq names.size
   end
 
-  it 'parses several names in parallel with given num of processes' do
+  it "parses several names in parallel with given num of processes" do
     names = []
     read_test_file { |n| names << (n[:name]) if n[:name] }
     names.uniq!
@@ -130,8 +171,8 @@ describe ParallelParser do
     expect(res.keys.size).to eq names.size
   end
 
-  it 'has parsed name in native ruby format and in returned as \
-      a hash with name as a key and parsed data as value' do
+  it "has parsed name in native ruby format and in returned as \
+      a hash with name as a key and parsed data as value" do
     names = []
     read_test_file { |n| names << (n[:name]) if n[:name] }
     names.uniq!

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: biodiversity
 version: !ruby/object:Gem::Version
-  version: 3.3.0
+  version: 3.4.0
 platform: ruby
 authors:
 - Dmitry Mozzherin
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-08-15 00:00:00.000000000 Z
+date: 2015-08-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: treetop