component_jyut6ping3.schema.yaml

# Rime schema
# 咪就係粵拼輸入法攞
# encoding: utf-8

schema:
  schema_id: component_jyut6ping3
  name: 粵拼拆字
  version: "0.3.3"
  author:
    - sgalal <ayaka@mail.shn.hk>
    - LeiMaau <leimaau@qq.com>
    - chaaklau <chaakming@gmail.com>
    - laubonghaudoi <laubonghaudoi@icloud.com>
    - tanxpyox <tanxpyox@gmail.com>

  description: |-
    採用香港語言學學會粵語拼音方案
    https://www.lshk.org/jyutping
    支援聲調顯示同輸入
    Telegram 用户交流組 https://t.me/rime_cantonese

  dependencies:
    - luna_pinyin
    - stroke
    - cangjie5

switches:
  - name: ascii_mode
    reset: 0
    states: [ 中文, 西文 ]
  - name: full_shape
    states: [ 半角, 全角 ]
  - options: [ noop, trad_hk, trad_tw, simplification ]
    states: [ 傳統漢字, 香港傳統漢字, 臺灣傳統漢字, 大陆简化汉字 ]
  - name: ascii_punct
    states: [ 。，, ．， ]
  - name: emoji_suggestion
    reset: 1
    states: [ 無 Emoji, 有 Emoji ]

engine:
  processors:
    - ascii_composer
    - recognizer
    - key_binder
    - speller
    - punctuator
    - selector
    - navigator
    - express_editor
  segmentors:
    - ascii_segmentor
    - matcher
    - affix_segmentor@luna_pinyin
    - affix_segmentor@stroke
    - affix_segmentor@cangjie5
    - abc_segmentor
    - punct_segmentor
    - fallback_segmentor
  translators:
    - punct_translator
    - script_translator
    - script_translator@luna_pinyin
    - table_translator@stroke
    - table_translator@cangjie5
  filters:
    - simplifier@trad_tw
    - simplifier@trad_hk
    - simplifier
    - simplifier@emoji_suggestion
    - uniquifier
    - reverse_lookup_filter

speller:
  alphabet: zyxwvutsrqponmlkjihgfedcba
  delimiter: " '"
  algebra:
    # 取消下兩行註釋，支援疑影不分
    #- derive/^ng([aeiou])/$1/
    #- derive/^([aeiou])/ng$1/

    # 取消下行註釋，支援 n- 併入 l- ，如「你」讀若「理」
    #- derive/^n(?!g)/l/

    # 取消下行註釋，支援 gw-, kw- 併入 g-, k- ，如「國」讀若「各」、「廓」讀若「確」
    #- derive/^(g|k)w/$1/

    # 取消下行註釋，支援獨立鼻音韻 ng 併入 m，如「吳」讀若「唔」
    #- derive/^ng([123456])$/m$1/

    - derive/[123456]//           # 忽略声调
    - xform/1/v/                  # 陰平
    - xform/4/vv/                 # 陽平
    - xform/2/x/                  # 陰上
    - xform/5/xx/                 # 陽上
    - xform/3/q/                  # 陰去
    - xform/6/qq/                 # 陽去
    - derive/^([aeiou])/q$1/      # 增加 q 表示喉塞
    - derive/^jy?([aeiou])/y$1/   # 容錯 je -> ye, jyu -> yu
    - derive/^jyu/ju/             # 容錯 jyu -> ju
    - derive/yu/y/                # 容錯 jyu -> jy
    - derive/eoi$/oey/            # 容錯
    - derive/eoi$/eoy/            # 容錯
    - derive/eo/oe/               # 容錯 eo/oe 不分
    - derive/oe(ng|k)$/eo$1/      # 容錯
    - derive/aa$/a/               # 容錯
    - abbrev/^([a-z]).+$/$1/      # 首字母簡拼
    - abbrev/^([a-z]{2}).+$/$1/   # 首2字母簡拼

translator:
  dictionary: component_jyut6ping3
  spelling_hints: 8
  preedit_format:
    - xform/([aeioumngptk])vv/$1⁴/
    - xform/([aeioumngptk])xx/$1⁵/
    - xform/([aeioumngptk])qq/$1⁶/
    - xform/([aeioumngptk])v/$1¹/
    - xform/([aeioumngptk])x/$1²/
    - xform/([aeioumngptk])q/$1³/

luna_pinyin:
  tag: luna_pinyin
  dictionary: luna_pinyin
  prefix: "`"
  suffix: ";"
  tips: 〔拼音〕
  preedit_format:
    - xform/([nl])v/$1ü/
    - xform/([nl])ue/$1üe/
    - xform/([jqxy])v/$1u/

stroke:
  tag: stroke
  dictionary: stroke
  prefix: "x"
  suffix: ";"
  tips: 〔筆畫〕
  preedit_format:
    - xlit/hspnz/一丨丿丶乙/

cangjie5:
  tag: cangjie5
  dictionary: cangjie5
  enable_user_dict: false
  prefix: "v"
  suffix: ";"
  tips: 〔倉頡五代〕
  preedit_format:
    - 'xform/^([a-z]*)$/$1\t（\U$1\E）/'
    - "xlit|ABCDEFGHIJKLMNOPQRSTUVWXYZ|日月金木水火土竹戈十大中一弓人心手口尸廿山女田難卜符|"
  comment_format:
    - "xlit|abcdefghijklmnopqrstuvwxyz~|日月金木水火土竹戈十大中一弓人心手口尸廿山女田難卜符～|"
  disable_user_dict_for_patterns:
    - "^z.*$"
    - "^yyy.*$"

reverse_lookup:
  tags: [ luna_pinyin, stroke, cangjie5 ]
  overwrite_comment: false
  dictionary: jyut6ping3

trad_tw:
  option_name: trad_tw
  opencc_config: t2tw.json

trad_hk:
  option_name: trad_hk
  opencc_config: t2hk.json

emoji_suggestion:
  opencc_config: emoji.json
  option_name: emoji_suggestion
  tips: all

punctuator:
  import_preset: symbols_cantonese

key_binder:
  import_preset: default

recognizer:
  import_preset: default
  patterns:
    punct: "^/([0-9]0?|[a-z]+)$"
    luna_pinyin: "`[a-z']*;?$"
    stroke: "^x[hspnz]*;?$"
    cangjie5: "^v[a-z]*;?$"