Method: Bio::FastaDefline#words

Defined in:
lib/bio/db/fasta/defline.rb

#words(case_sensitive = nil, kill_regexp = self.class::KillRegexpArray, kwhash = self.class::KillWordsHash) ⇒ Object

Shows words used in the defline. Returns an Array.



390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
 
# File 'lib/bio/db/fasta/defline.rb', line 390

def words(case_sensitive = nil, kill_regexp = self.class::KillRegexpArray,
          kwhash = self.class::KillWordsHash)
  a = descriptions.join(' ').split(/[\.\,\;\:\(\)\[\]\{\}\<\>\"\'\`\~\/\|\?\!\&\@\# \x00-\x1f\x7f]+/)
  a.collect! do |x|
    x.sub!(/\A[\$\*\-\+]+/, '')
    x.sub!(/[\$\*\-\=]+\z/, '')
    if x.size <= 1 then
      nil
    elsif kwhash[x.downcase] then
      nil
    else
      if kill_regexp.find { |expr| expr =~ x } then
        nil
      else
        x
      end
    end
  end
  a.compact!
  a.collect! { |x| x.downcase } unless case_sensitive
  a.sort!
  a.uniq!
  a
end