wuss.hpp

Go to the documentation of this file.

// SPDX-FileCopyrightText: 2006-2025 Knut Reinert & Freie Universität Berlin
// SPDX-FileCopyrightText: 2016-2025 Knut Reinert & MPI für molekulare Genetik
// SPDX-License-Identifier: BSD-3-Clause

 
#pragma once
 
#include <cmath>
#include <limits>
#include <vector>
 
#include <seqan3/alphabet/alphabet_base.hpp>
#include <seqan3/alphabet/structure/concept.hpp>
#include <seqan3/utility/char_operations/transform.hpp>
 
// ------------------------------------------------------------------
// wuss
// ------------------------------------------------------------------
 
namespace seqan3
{

template <uint8_t SIZE = 51>
class wuss : public alphabet_base<wuss<SIZE>, SIZE>
{
    static_assert(SIZE >= 15 && SIZE <= 67 && SIZE % 2 == 1,
                  "The wuss<> alphabet size must be an odd number in range 15..67.");
 
private:
    using base_t = alphabet_base<wuss<SIZE>, SIZE>;

    friend base_t;
 
protected:
    using typename base_t::char_type;
    using typename base_t::rank_type;
 
public:
    using base_t::alphabet_size;
    using base_t::to_char;
    using base_t::to_rank;

    constexpr wuss() noexcept = default;                         
    constexpr wuss(wuss const &) noexcept = default;             
    constexpr wuss(wuss &&) noexcept = default;                  
    constexpr wuss & operator=(wuss const &) noexcept = default; 
    constexpr wuss & operator=(wuss &&) noexcept = default;      
    ~wuss() noexcept = default;                                  


    constexpr bool is_pair_open() const noexcept
    {
        return interaction_tab[to_rank()] < 0;
    }

    constexpr bool is_pair_close() const noexcept
    {
        return interaction_tab[to_rank()] > 0;
    }

    constexpr bool is_unpaired() const noexcept
    {
        return interaction_tab[to_rank()] == 0;
    }

    // formula: (alphabet size - 7 unpaired characters) / 2, as every bracket exists as opening/closing pair
    static constexpr uint8_t max_pseudoknot_depth{static_cast<uint8_t>((alphabet_size - 7) / 2)};

    constexpr std::optional<uint8_t> pseudoknot_id() const noexcept
    {
        if (interaction_tab[to_rank()] != 0)
            return std::abs(interaction_tab[to_rank()]) - 1;
        else
            return std::nullopt; // unpaired
    }

 
private:
    static constexpr char_type rank_to_char(rank_type const rank)
    {
        return rank_to_char_table[rank];
    }

    static constexpr rank_type char_to_rank(char_type const chr)
    {
        using index_t = std::make_unsigned_t<char_type>;
        return char_to_rank_table[static_cast<index_t>(chr)];
    }

    static constexpr std::array<char_type, alphabet_size> rank_to_char_table{
        []() constexpr
        {
            std::array<char_type, alphabet_size>
                chars{'.', ':', ',', '-', '_', '~', ';', '<', '(', '[', '{', '>', ')', ']', '}'};
 
            // pseudoknot letters
            for (rank_type rnk = 15u; rnk + 1u < alphabet_size; rnk += 2u)
            {
                char_type const off = static_cast<char_type>((rnk - 15u) / 2u);
                chars[rnk] = 'A' + off;
                chars[rnk + 1u] = 'a' + off;
            }
 
            return chars;
        }()};

    static constexpr std::array<rank_type, 256> char_to_rank_table{[]() constexpr
                                                                   {
                                                                       std::array<rank_type, 256> rank_table{};
 
                                                                       rank_table.fill(6u);
 
                                                                       // set alphabet values
                                                                       for (rank_type rnk = 0u; rnk < alphabet_size;
                                                                            ++rnk)
                                                                           rank_table[rank_to_char_table[rnk]] = rnk;
 
                                                                       return rank_table;
                                                                   }()};

    static constexpr std::array<int8_t, SIZE> interaction_tab{
        []() constexpr
        {
            static_assert(static_cast<int16_t>(std::numeric_limits<int8_t>::max()) >= SIZE);
            static_assert(-static_cast<int16_t>(std::numeric_limits<int8_t>::min()) >= SIZE);
 
            std::array<int8_t, alphabet_size> interaction_table{};
            int8_t cnt_open = 0;
            int8_t cnt_close = 0;
 
            for (rank_type rnk = 0u; rnk <= 6u; ++rnk)
                interaction_table[rnk] = 0;
 
            for (rank_type rnk = 7u; rnk <= 10u; ++rnk)
                interaction_table[rnk] = --cnt_open;
 
            for (rank_type rnk = 11u; rnk <= 14u; ++rnk)
                interaction_table[rnk] = ++cnt_close;
 
            for (rank_type rnk = 15u; rnk + 1u < alphabet_size; rnk += 2u)
            {
                interaction_table[rnk] = --cnt_open;
                interaction_table[rnk + 1u] = ++cnt_close;
            }
 
            return interaction_table;
        }()};
};

using wuss51 = wuss<51>;
 
inline namespace literals
{

constexpr wuss51 operator""_wuss51(char const ch) noexcept
{
    return wuss51{}.assign_char(ch);
}

SEQAN3_WORKAROUND_LITERAL std::vector<wuss51> operator""_wuss51(char const * str, std::size_t len)
{
    std::vector<wuss51> vec;
    vec.resize(len);
 
    for (size_t idx = 0ul; idx < len; ++idx)
        vec[idx].assign_char(str[idx]);
 
    return vec;
}

 
} // namespace literals
 
} // namespace seqan3