language: el
script: Greek

normalization:
  unicode_normal_form: NFC
  # Greek Unicode: precomposed forms.
  # Polytonic diacritics (breathing marks, accents, iota subscript)
  # are preserved. These carry morphological and semantic information
  # in Koine Greek that must not be stripped.
  strip_breathing_marks: false
  strip_accents_for_indexing: false
  # Accents in Greek affect meaning and morphological parsing.
  # They are never stripped from the canonical form.
  lowercase_for_indexing: true
  # Lemma lookup uses lowercased forms.
  # Surface forms preserve original casing.
  lunate_sigma: false
  # Use standard sigma/final-sigma (σ/ς) not lunate (ϲ).
  # Final sigma is morphologically relevant.

punctuation:
  strategy: preserve
  ano_teleia: preserve
  # The ano teleia (·) is the Greek semicolon and must be preserved.

spacing:
  strategy: whitespace-split

script_direction: ltr