Class: Kotoshu::Grammar::PatternMatchers::PhraseMatcher

Inherits:
BaseMatcher
  • Object
show all
Defined in:
lib/kotoshu/grammar/pattern_matchers/phrase_matcher.rb

Overview

Matcher for literal multi-word phrase confusions.

Catches phrases where every word is individually valid but the combination is a common error — typically phonetic confusions like "could of" (should be "could have"). The spelling checker passes these because each word is in the dictionary; only a phrase-level grammar check catches them.

The rule config declares the wrong phrase and its replacement:

conditions:
- type: phrase_check
  wrong_phrase: "could of"
  suggestion: "could have"

The matcher scans the token stream for consecutive tokens whose downcased text matches wrong_phrase (split on whitespace) and emits one error per match.

Instance Method Summary collapse

Methods inherited from BaseMatcher

#initialize

Constructor Details

This class inherits a constructor from Kotoshu::Grammar::PatternMatchers::BaseMatcher

Instance Method Details

#match(tokens, rule) ⇒ Array<Hash>

Match tokens against the phrase-confusion pattern.

Parameters:

  • tokens (Array<Hash>) —

    Array of token hashes

  • rule (Rule) —

    The rule being checked

Returns:

  • (Array<Hash>) —

    Array of error hashes



30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
 
# File 'lib/kotoshu/grammar/pattern_matchers/phrase_matcher.rb', line 30

def match(tokens, rule)
  wrong_phrase = phrase_condition&.dig('wrong_phrase')
  suggestion = phrase_condition&.dig('suggestion')
  return [] unless wrong_phrase && suggestion

  wrong_tokens = wrong_phrase.downcase.split
  return [] if wrong_tokens.empty?

  matches = []
  tokens.each_with_index do |start_token, start_idx|
    next unless start_token[:token]&.downcase == wrong_tokens.first

    match_idx = find_match(tokens, start_idx, wrong_tokens)
    next unless match_idx

    matches << build_error(tokens, match_idx, wrong_tokens.length,
                           wrong_phrase, suggestion, rule)
  end
  matches
end