Class: Kotoshu::Language::Tokenizer::PortugueseTokenizer

Inherits:
Base
  • Object
show all
Defined in:
lib/kotoshu/language/tokenizer/portuguese_tokenizer.rb

Overview

Tokenizer for Portuguese text.

Ported from LanguageTool’s PortugueseWordTokenizer.

Handles:

  • Decimal comma between digits (3,14)

  • Dotted numbers (1.000.000)

  • Dates (01.01.2024, 2024-01-01)

  • Colons in time (12:25)

  • Hyphens with do-not-split list

  • Spaced decimals (2 000 000)

Constant Summary collapse

WORD_SEPARATORS =

Portuguese word separators - most punctuation and whitespace Note: We protect special patterns before splitting

/[\s"()\[\]{}<>@€£\\$%‰‱ºªᵃᵒˢ|`~#^·]/.freeze
DECIMAL_COMMA_SUBST =

Placeholder characters (using non-printing characters)

"\uE001"
NON_BREAKING_SPACE_SUBST = 
"\uE002"
NON_BREAKING_DOT_SUBST = 
"\uE003"
NON_BREAKING_COLON_SUBST = 
"\uE004"
DECIMAL_COMMA_PATTERN =

Decimal comma between digits: 3,14

/(\d),(\d)/
DOTTED_NUMBERS_PATTERN =

Dotted numbers: 1.000.000

/(\d)\.(\d)/
COLON_NUMBERS_PATTERN =

Colon in numbers (time): 12:25

/(\d):(\d)/
DATE_PATTERN =

Date patterns: 01.01.2024, 2024-01-01

/(\d{2})\.(\d{2})\.(\d{4})|(\d{4})\.(\d{2})\.(\d{2})|(\d{4})-(\d{2})-(\d{2})/
SPACED_DECIMAL_PATTERN =

Spaced decimals: 2 000 000

/(?<=^|[\s(])\d{1,3}( \d{3})+(?:[,#{DECIMAL_COMMA_SUBST}#{NON_BREAKING_DOT_SUBST}]\d+)?(?=\D|$)/
DO_NOT_SPLIT =

Do-not-split list (from LanguageTool)

%w[
  mers-cov mcgraw-hill sars-cov-2 sars-cov
  ph-metre ph-metres anti-ivg anti-uv anti-vih al-qaïda
].freeze

Instance Method Summary collapse

Methods inherited from Base

#normalize, #skip_token?, #tokenize_with_positions, #word_boundary_regex, #word_char?

Instance Method Details

#tokenize(text) ⇒ Object 



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
 
# File 'lib/kotoshu/language/tokenizer/portuguese_tokenizer.rb', line 49

def tokenize(text)
  return [] if text.nil? || text.strip.empty?

  # Handle decimal commas
  if text.include?(",")
    text = text.gsub(DECIMAL_COMMA_PATTERN, "\\1#{DECIMAL_COMMA_SUBST}\\2")
  end

  # Handle dots in numbers and dates
  if text.include?(".")
    # Handle dates first (before dotted numbers to avoid conflicts)
    text = text.gsub(DATE_PATTERN) do |match|
      # match[0] is the full match, match[1-9] are the capture groups
      if match[1] && match[2] && match[3] # DD.MM.YYYY
        "#{match[1]}#{NON_BREAKING_DOT_SUBST}#{match[2]}#{NON_BREAKING_DOT_SUBST}#{match[3]}"
      elsif match[4] && match[5] && match[6] # YYYY.MM.DD
        "#{match[4]}#{NON_BREAKING_DOT_SUBST}#{match[5]}#{NON_BREAKING_DOT_SUBST}#{match[6]}"
      elsif match[7] && match[8] && match[9] # YYYY-MM-DD (keep as-is)
        match[0]
      else
        match[0]
      end
    end
    text = text.gsub(DOTTED_NUMBERS_PATTERN, "\\1#{NON_BREAKING_DOT_SUBST}\\2")
  end

  # Handle spaced decimals: 2 000 000
  text = handle_spaced_decimals(text)

  # Handle colons in time: 12:25
  if text.include?(":")
    text = text.gsub(COLON_NUMBERS_PATTERN, "\\1#{NON_BREAKING_COLON_SUBST}\\2")
  end

  # Split on word boundaries
  raw_tokens = text.split(WORD_SEPARATORS)

  # Process each token
  tokens = []
  raw_tokens.each do |token|
    next if token.empty?

    # Restore placeholders
    token = restore_placeholders(token)

    # Handle hyphenated words
    parts = words_to_add(token)
    tokens.concat(parts)
  end

  # Filter and normalize
  tokens
    .map { |token| normalize(token) }
    .reject { |token| skip_token?(token) }
end