Module: Export::Coldp::Files::Synonym

Defined in:
lib/export/coldp/files/synonym.rb

Overview

Bigger picture: understand how this maps to core name usage table in CoL

Class Method Summary collapse

Class Method Details

.generate(otus, project_members, reference_csv = nil, skip_name_ids = []) ⇒ Object

This is currently factored to use no ActiveRecord instances



29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
 
# File 'lib/export/coldp/files/synonym.rb', line 29

def self.generate(otus, project_members, reference_csv = nil, skip_name_ids = [])
  ::CSV.generate(col_sep: "\t") do |csv|

    csv << %w{taxonID nameID status remarks referenceID modified modifiedBy}

    # Only valid otus with taxon names, see lib/export/coldp.rb#otus
    #  ?! in groups of
    otus.select('otus.id id, taxon_names.cached cached, otus.taxon_name_id taxon_name_id')
      .pluck(:id, :cached, :taxon_name_id)
      .each do |o|

        # TODO: Confirm resolved: original combinations of invalid names are not being handled correclty in reified

        # Here we grab the hierarchy again, and filter it by
        #   1) allow only invalid names OR names with differing original combinations
        #   2) of 1) eliminate Combinations with identical names to current placement
        #
        a = TaxonName.that_is_invalid
          .where(cached_valid_taxon_name_id: o[2])
          .where.not("(taxon_names.type = 'Combination' AND taxon_names.cached = ?)", o[1]) # Hybrids allowed, intended?

        b = TaxonName.where(cached_valid_taxon_name_id: o[2])
          .where('(taxon_names.cached_original_combination != taxon_names.cached)')
          .where.not("(taxon_names.type = 'Combination' AND taxon_names.cached = ?)", o[1])

        c = ::Queries.union(TaxonName, [a,b])

        # Hernán notes:
        # TaxonName.where(cached_valid_taxon_name_id: 42).merge(TaxonName.where.not(type: 'Combination').or(TaxonName.where.not(cached: 'Forty two'))).to_sql
        # Mjy - "or" performance is bad? or?

        # Original concept
        # TaxonName
        #   .where(cached_valid_taxon_name_id: o[2]) # == .historical_taxon_names
        #   .where("( ((taxon_names.id != taxon_names.cached_valid_taxon_name_id) OR ((taxon_names.cached_original_combination != taxon_names.cached))) AND NOT (taxon_names.type = 'Combination' AND taxon_names.cached = ?))", o[1]) # see name.rb

        c.pluck(:id, :cached, :cached_original_combination, :type, :rank_class, :cached_secondary_homonym, :updated_at, :updated_by_id)
          .each do |t|
            reified_id = ::Export::Coldp.reified_id(t[0], t[1], t[2])
            next if skip_name_ids.include?               reified_id = ::Export::Coldp.reified_id(t[0], t[1], t[2])


            # skip duplicate protonyms created for family group relationships
            if t[4]&.include? 'FamilyGroup'
              tn = TaxonName.find(t[0])
              if tn.taxon_name_relationships.any? {|tnr| tnr.type == 'TaxonNameRelationship::Iczn::Invalidating::Usage::FamilyGroupNameForm'}
                if tn.name == o[1]  # only skip if it matches the accepted name, because there might be multiple protonyms added for family group relationships
                  next
                end
              end
            end

            # skips including parent binomial as a synonym of autonym trinomial
            # TODO: may need to handle cases in which the gender stems are not an exact match
            matches = t[1].match(/([A-Z][a-z]+) \(.+\) ([a-z]+)/)
            cached = t[1]
            if matches&.size == 3
              cached = "#{matches[1]} #{matches[2]}"
            end
            unless t[5].nil?
              if !t[1].nil? and cached.include? t[5] and (t[4].match(/::Subspecies$/) or t[4].match(/::Form$/) or t[4].match(/::Variety$/))
                next
              end

              if !t[2].nil? and t[2].include? t[5] and o[1].include? t[5] and t[4].match(/::Species$/)
                next
              end
            end

            # TODO: This code block is erroneously removing basionyms from the synonyms section but we may need an improved form of it to remove duplicate synonyms (https://github.com/SpeciesFileGroup/taxonworks/issues/3482)
            # matches = t[1].match(/([A-Z][a-z]+) \(.+\) ([a-z]+)/)
            #
            # if matches&.size == 3        # cached_original_combination != cached_secondary_homonym
            #   if t[5] == "#{matches[1]} #{matches[2]}" and t[2] != t[5]
            #     next
            #   end
            # end

            # skips combinations including parent binomial as a synonym of autonym trinomial
            if t[3] == 'Combination' and o[1].include? t[1]
              next
            end

            # skip making parent genus Aus a synonym of subgenus autonym (Aus) Aus
            autonym_test = t[1]&.gsub(/\(/, '')&.gsub(/\)/, '')&.split(' ')
            if t[4]&.include?('Subgenus') && autonym_test.size >= 2 && autonym_test[0] == autonym_test[1] && t[2] == autonym_test[0]
              next
            end

            csv << [
              o[0],                                             # taxonID attached to the current valid concept
              reified_id,                                       # nameID
              nil,                                              # status  TODO: def status(taxon_name_id)
              Export::Coldp.sanitize_remarks(remarks_field),    # remarks
              nil,                                              # referenceID   Unclear what this means in TW
              Export::Coldp.modified(t[6]),                     # modified
              Export::Coldp.modified_by(t[7], project_members)  # modifiedBy
            ]
          end
      end
  end
end

.reference_id_field(otu) ⇒ Object 



24
25
26
 
# File 'lib/export/coldp/files/synonym.rb', line 24

def self.reference_id_field(otu)
  nil
end

.remarks_field ⇒ Object 



20
21
22
 
# File 'lib/export/coldp/files/synonym.rb', line 20

def self.remarks_field
  nil
end

.status(o, t) ⇒ Object

Last 3 of api.checklistbank.org/vocab/taxonomicstatus

Returns:

  • String



11
12
13
14
15
16
17
18
 
# File 'lib/export/coldp/files/synonym.rb', line 11

def self.status(o, t)
  #'accepted'
  #'provisionally accepted'
  #'synonym'
  #'ambiguous synonym'
  #'missaplied'
  'synonym'
end