UnicodeData.hpp

// Copyright 2019 Pokitec
// All rights reserved.
 
#pragma once
 
#include "TTauri/Text/Grapheme.hpp"
#include "TTauri/Text/UnicodeBidi.hpp"
#include "TTauri/Foundation/ResourceView.hpp"
#include "TTauri/Foundation/math.hpp"
#include "TTauri/Foundation/URL.hpp"
#include "TTauri/Foundation/required.hpp"
#include <nonstd/span>
 
namespace tt {
struct UnicodeData_Description;
 
enum class GraphemeUnitType : uint8_t {
    Other = 0,
    CR = 1,
    LF = 2,
    Control = 3,
    Extend = 4,
    ZWJ = 5,
    Regional_Indicator = 6,
    Prepend = 7,
    SpacingMark = 8,
    L = 9,
    V = 10,
    T = 11,
    LV = 12,
    LVT = 13,
    Extended_Pictographic = 14
};
 
 
 
struct GraphemeBreakState {
    GraphemeUnitType previous = GraphemeUnitType::Other;
    int RICount = 0;
    bool firstCharacter = true;
    bool inExtendedPictographic = false;
 
    void reset() noexcept {
        previous = GraphemeUnitType::Other;
        RICount = 0;
        firstCharacter = true;
        inExtendedPictographic = false;
    }
};
 
 
enum class BidiClass : uint8_t {
    Unknown = 0,
    L = 1, 
    R = 2, 
    AL = 3, 
    EN = 4, 
    ES = 5, 
    ET = 6, 
    AN = 7, 
    CS = 8, 
    NSM = 9, 
    BN = 10, 
    B = 11, 
    S = 12, 
    WS = 13, 
    ON = 14, 
    // Explicit values.
    LRE, 
    LRO, 
    RLE, 
    RLO, 
    PDF, 
    LRI, 
    RLI, 
    FSI, 
    PDI 
};
 
enum GeneralCharacterClass {
    Unknown,
    Digit,
    Letter,
    WhiteSpace,
    ParagraphSeparator
};
 
[[nodiscard]] constexpr GeneralCharacterClass to_GeneralCharacterClass(BidiClass bidiClass) noexcept {
    switch (bidiClass) {
    case BidiClass::Unknown: return GeneralCharacterClass::Unknown;
    case BidiClass::L: return GeneralCharacterClass::Letter;
    case BidiClass::R: return GeneralCharacterClass::Letter;
    case BidiClass::AL: return GeneralCharacterClass::Letter;
    case BidiClass::EN: return GeneralCharacterClass::Digit;
    case BidiClass::ES: return GeneralCharacterClass::Unknown;
    case BidiClass::ET: return GeneralCharacterClass::Unknown;
    case BidiClass::AN: return GeneralCharacterClass::Digit;
    case BidiClass::CS: return GeneralCharacterClass::Unknown;
    case BidiClass::NSM: return GeneralCharacterClass::Unknown;
    case BidiClass::BN: return GeneralCharacterClass::Unknown;
    case BidiClass::B: return GeneralCharacterClass::ParagraphSeparator;
    case BidiClass::S: return GeneralCharacterClass::Unknown;
    case BidiClass::WS: return GeneralCharacterClass::WhiteSpace;
    case BidiClass::ON: return GeneralCharacterClass::Unknown;
    case BidiClass::LRE: return GeneralCharacterClass::Unknown;
    case BidiClass::LRO: return GeneralCharacterClass::Unknown;
    case BidiClass::RLE: return GeneralCharacterClass::Unknown;
    case BidiClass::RLO: return GeneralCharacterClass::Unknown;
    case BidiClass::PDF: return GeneralCharacterClass::Unknown;
    case BidiClass::LRI: return GeneralCharacterClass::Unknown;
    case BidiClass::RLI: return GeneralCharacterClass::Unknown;
    case BidiClass::FSI: return GeneralCharacterClass::Unknown;
    case BidiClass::PDI: return GeneralCharacterClass::Unknown;
    default: tt_no_default;
    }
}
 
class UnicodeData {
private:
    nonstd::span<std::byte const> bytes;
 
    std::unique_ptr<ResourceView> view;
 
    size_t descriptions_offset;
    size_t descriptions_count;
 
    size_t compositions_offset;
    size_t compositions_count;
public:
    UnicodeData(nonstd::span<std::byte const> bytes);
 
    UnicodeData(std::unique_ptr<ResourceView> view);
 
    UnicodeData() = delete;
    UnicodeData(UnicodeData const &other) = delete;
    UnicodeData &operator=(UnicodeData const &other) = delete;
    UnicodeData(UnicodeData &&other) = delete;
    UnicodeData &operator=(UnicodeData &&other) = delete;
    ~UnicodeData() = default;
 
    std::u32string toNFD(std::u32string_view text, bool decomposeLigatures=false) const noexcept;
 
    std::u32string toNFC(std::u32string_view text, bool decomposeLigatures=false, bool composeCRLF=false) const noexcept;
 
    std::u32string toNFKD(std::u32string_view text) const noexcept;
 
    std::u32string toNFKC(std::u32string_view text, bool composeCRLF=false) const noexcept;
 
    bool checkGraphemeBreak(char32_t codeUnit, GraphemeBreakState &state) const noexcept;
 
    BidiClass getBidiClass(char32_t codePoint) const noexcept;
 
private:
    void initialize();
 
    UnicodeData_Description const *getDescription(char32_t codePoint) const noexcept;
    GraphemeUnitType getGraphemeUnitType(char32_t codePoint) const noexcept;
    uint8_t getDecompositionOrder(char32_t codePoint) const noexcept;
 
    char32_t compose(char32_t startCharacter, char32_t composingCharacter, bool composeCRLF) const noexcept;
    void decomposeCodePoint(std::u32string &result, char32_t codePoint, bool decomposeCompatible, bool decomposeLigatures) const noexcept;
    std::u32string decompose(std::u32string_view text, bool decomposeCompatible, bool decomposeLigatures=false) const noexcept;
 
 
    static void reorder(std::u32string &text) noexcept;
 
    static void clean(std::u32string &text) noexcept;
 
 
    void compose(std::u32string &text, bool composeCRLF=false) const noexcept;
};
 
}
 
namespace tt {
 
template<>
std::unique_ptr<tt::UnicodeData> parseResource(URL const &location);
 
}