Module: PreProcessor

Defined in:
lib/biodiversity/parser.rb

Constant Summary collapse

NOTES = 
/\s+(species\s+group|species\s+complex|group|author)\b.*$/i
TAXON_CONCEPTS1 = 
/\s+(sensu\.|sensu|auct\.|auct)\b.*$/i
TAXON_CONCEPTS2 = 
/\s+(\(?s\.\s?s\.|\(?s\.\s?l\.|\(?s\.\s?str\.|\(?s\.\s?lat\.|sec\.|sec|near)\b.*$/
TAXON_CONCEPTS3 = 
/(,\s*|\s+)(pro parte|p.\s?p.)\s*$/i
NOMEN_CONCEPTS = 
/(,\s*|\s+)(\(?nomen|\(?nom\.|\(?comb\.).*$/i
LAST_WORD_JUNK = 
/(,\s*|\s+)(von|van|sensu|new|non|nec|cf|ssp|subsp|subgen|hybrid|hort.|hort)\s*$/i

Class Method Summary collapse

Class Method Details

.clean(a_string) ⇒ Object 



17
18
19
20
21
22
23
 
# File 'lib/biodiversity/parser.rb', line 17

def self.clean(a_string)
  [NOTES, TAXON_CONCEPTS1, TAXON_CONCEPTS2, TAXON_CONCEPTS3, NOMEN_CONCEPTS, LAST_WORD_JUNK].each do |i|
    a_string = a_string.gsub(i, '')
  end
  a_string = a_string.tr('ſ','s') #old 's'
  a_string
end