Method: Ariel::Learner#lengthen_landmark

Defined in:

lib/ariel/learner.rb

#lengthen_landmark(landmark, index) ⇒ Object

Implements landmark refinements. Landmarks are lengthened to make them more specific.

• Takes a landmark and its index in the current rule.

• Applies the rule consisting of all previous landmarks in the current rule, so the landmark can be considered in the context of the point from which it shall be applied.

• Every point at which the landmark matches after the cur_loc is considered.

• Two extended landmarks are generated - a landmark that includes the token before the match, and a landmark that includes that token after the match.

• Rules are generated incorporating these extended landmarks, including alternative landmark extensions that use relevant wildcards.

# File 'lib/ariel/learner.rb', line 187

def lengthen_landmark(landmark, index)
  current_seed.rewind #In case apply_rule isn't called as index=0
  result = @current_rule.partial(0..(index-1)).closest_match current_seed if index > 0 #Don't care about already matched tokens
  return 0 unless result # Rule doesn't match, no point refining
  refined_rules=[]
  width = landmark.size
  while current_seed.skip_to(*landmark) #Probably should stop when cur_pos > label_index
    break if current_seed.cur_pos > current_seed.label_index
    match_start = (current_seed.cur_pos - 1) - width #pos of first matched token
    match_end = current_seed.cur_pos - 1 #pos of last matched token
    preceding_token = current_seed.tokens[match_start-1]
    trailing_token = current_seed.tokens[match_end+1]
    front_extended_landmark = landmark.clone.insert(0, preceding_token.text) if preceding_token
    back_extended_landmark = landmark.clone.insert(-1, trailing_token.text) if trailing_token
    f = current_rule.deep_clone
    b = current_rule.deep_clone
    f.landmarks[index] = front_extended_landmark if front_extended_landmark
    b.landmarks[index] = back_extended_landmark if back_extended_landmark
    refined_rules << f
    refined_rules.concat f.generalise_feature(index, 0)
    refined_rules << b
    refined_rules.concat b.generalise_feature(index, -1)
  end
  @candidates.concat refined_rules
  Log.debug "#{refined_rules.size} landmark refinements generated"
  return refined_rules.size
end