🥮 mooncakes.io
    Sign in

    emoji

    Unicode Emoji Segmentation — MoonBit implementation of UTS #51 v17.0.0

    emoji
    unicode
    uts51
    segmentation
    Download zip
    Author
    Version
    0.3.0
    License
    MIT
    Repository
    Last updated
    5 months ago
    Downloads
    29

    #moonwell-emoji — Unicode Emoji Segmentation

    A MoonBit library for Unicode emoji segmentation implementing UTS #51 v17.0.0.

    MoonBit port of c-emoji.

    The core is a DFA that classifies and segments emoji sequences in a single O(n) pass with no heap allocation beyond result storage. Two scanning modes are provided — strict and greedy — that differ in how they handle incomplete sequences at boundaries.

    #Usage

    let text : Array[Int] = [0x1F468, 0x1F3FB, 0x200D, 0x1F4BB] // 👨🏻‍💻
    let seqs = @emoji.scan_strict(text)

    for i = 0; i < seqs.length(); i = i + 1 {
      println("emoji at [\{seqs[i].start}, \{seqs[i].end}]")
    }
    // emoji at [0, 3]

    start and end are inclusive indices into the codepoint array.

    #Strict vs greedy

    Inputstrictgreedy
    👨 ZWJ (trailing ZWJ)∅{0,0}
    👨 ZWJ ZWJ 👩 (double ZWJ){3,3}{0,0} {3,3}
    👨 ZWJ A (invalid ZWJ target)∅{0,0}
    🏴 gb eng (tag without cancel)∅{0,0}

    #API

    #Scanning

    • scan_strict(codepoints : Array[Int]) -> Array[ScanRange] — conservative scan
    • scan_greedy(codepoints : Array[Int]) -> Array[ScanRange] — permissive scan

    #Sequence classification

    • classify_type(bitmask : Int) -> EmojiSequenceType — from recorded bitmask
    • classify_style(bitmask : Int) -> EmojiPresentationStyle — from recorded bitmask

    #Presentation resolution

    • resolve(type, style, leading_cp) -> EmojiPresentationStyle — full resolution
    • resolve_default(type, leading_cp) -> EmojiPresentationStyle — when no VS present

    #UCD predicates

    • is_emoji(cp : Int) -> Bool
    • is_presentation(cp : Int) -> Bool
    • is_modifier_base(cp : Int) -> Bool
    • is_modifier(cp : Int) -> Bool
    • is_keycap_base(cp : Int) -> Bool
    • is_regional_indicator(cp : Int) -> Bool

    #Static memory

    ComponentSize
    Emoji trie952 bytes
    Modifier base trie520 bytes
    Presentation trie744 bytes
    DFA transition table169 bytes
    Total static data2,385 bytes

    #Unicode version

    Unicode 17.0.0.

    #License

    MIT License.

    Copyright (c) 2026 Christian Hansen chansen@cpan.org.

    Copyright (c) 2026 fundon fundon@pindash.io.

    Class

    source
    pub(all) enum Class {
      OTHER
      EMOJI
      MODIFIER_BASE
      MODIFIER
      VS15
      VS16
      RI
      KEYCAP_BASE
      KEYCAP_TERM
      TAG_BASE
      TAG_SPEC
      TAG_TERM
      ZWJ
    } derive(Eq, 
    Debug    )

    EmojiPresentationStyle

    source
    pub(all) enum EmojiPresentationStyle {
      DEFAULT
      TEXT
      EMOJI
    } derive(Eq, 
    Debug    )

    EmojiSequenceType

    source
    pub(all) enum EmojiSequenceType {
      BASIC
      KEYCAP
      FLAG
      TAG
      MODIFIER
      ZWJ
    } derive(Eq, 
    Debug    )

    ScanRange

    source
    pub struct ScanRange {
      start : Int
      end : Int
    }

    State

    source
    pub(all) enum State {
      REJECT
      START
      TERMINAL
      EMOJI
      MODIFIER_BASE
      OPTIONAL_ZWJ
      KEYCAP_VS
      TAG_BASE
      RI
      TAG_SPEC
      TAG_EMPTY
      KEYCAP_BASE
      ZWJ
    } derive(Compare, Eq, 
    Debug    )

    RECORD_STATE_SHIFT

    source
    let RECORD_STATE_SHIFT : Int

    class_to_int

    source
    fn class_to_int(c : Class) -> Int

    Convert Class enum to its integer discriminant.

    classify_style

    source
    fn classify_style(bitmask : Int) -> EmojiPresentationStyle

    classify_type

    source
    fn classify_type(bitmask : Int) -> EmojiSequenceType

    dfa_is_accepting

    source
    fn dfa_is_accepting(state : State) -> Bool

    dfa_is_boundary

    source
    fn dfa_is_boundary(state : State) -> Bool

    dfa_is_start

    source
    fn dfa_is_start(state : State) -> Bool

    dfa_step

    source
    fn dfa_step(state : State, klass : Class) -> State

    dfa_step_record

    source
    fn dfa_step_record(state : State, klass : Class, bitmask : Int) -> (State, Int)

    Step the DFA and record class/state bits into the bitmask accumulator.

    int_to_state

    source
    fn int_to_state(n : Int) -> State

    Convert an integer back to State enum.

    is_emoji

    source
    fn is_emoji(cp : Int) -> Bool

    is_emoji_presentation_selector

    source
    fn is_emoji_presentation_selector(cp : Int) -> Bool

    is_enclosing_keycap

    source
    fn is_enclosing_keycap(cp : Int) -> Bool

    is_keycap_base

    source
    fn is_keycap_base(cp : Int) -> Bool

    is_modifier

    source
    fn is_modifier(cp : Int) -> Bool

    is_modifier_base

    source
    fn is_modifier_base(cp : Int) -> Bool

    is_presentation

    source
    fn is_presentation(cp : Int) -> Bool

    is_regional_indicator

    source
    fn is_regional_indicator(cp : Int) -> Bool

    is_tag_base

    source
    fn is_tag_base(cp : Int) -> Bool

    is_tag_spec

    source
    fn is_tag_spec(cp : Int) -> Bool

    is_tag_term

    source
    fn is_tag_term(cp : Int) -> Bool

    is_text_presentation_selector

    source
    fn is_text_presentation_selector(cp : Int) -> Bool

    resolve

    source
    fn resolve(type_ : EmojiSequenceType, style : EmojiPresentationStyle, leading_cp : Int) -> EmojiPresentationStyle

    resolve_default

    source
    fn resolve_default(type_ : EmojiSequenceType, leading_cp : Int) -> EmojiPresentationStyle

    scan_greedy

    source
    fn scan_greedy(codepoints : Array[Int]) -> Array[ScanRange]

    scan_strict

    source
    fn scan_strict(codepoints : Array[Int]) -> Array[ScanRange]

    state_to_int

    source
    fn state_to_int(s : State) -> Int

    Convert State enum to its integer discriminant.

    ucd_classify

    source
    fn ucd_classify(cp : Int) -> Class

    Source Files

    Powered by MoonBit

    Site sourceReport issuePackagesBuild queueSkillsStatistics

    © 2026 mooncakes.io