nucleotide_base.hpp

Go to the documentation of this file.

// -----------------------------------------------------------------------------------------------------
// Copyright (c) 2006-2021, Knut Reinert & Freie Universität Berlin
// Copyright (c) 2016-2021, Knut Reinert & MPI für molekulare Genetik
// This file may be used, modified and/or redistributed under the terms of the 3-clause BSD-License
// shipped with this file and also available at: https://github.com/seqan/seqan3/blob/master/LICENSE.md
// -----------------------------------------------------------------------------------------------------
 
#pragma once
 
#include <seqan3/alphabet/alphabet_base.hpp>
#include <seqan3/alphabet/detail/convert.hpp>
#include <seqan3/alphabet/nucleotide/concept.hpp>
#include <seqan3/utility/char_operations/transform.hpp>
 
#ifdef SEQAN3_DEPRECATED_310
namespace seqan3::detail
{
// helper concept to deprecate old char_to_rank lookup tables
template <typename alphabet_t>
SEQAN3_CONCEPT has_complement_table = requires()
{
    { alphabet_t::complement_table[0] };
};
} // namespace seqan3::detail
#endif // SEQAN3_DEPRECATED_310
 
namespace seqan3
{
 
template <typename derived_type, auto size>
class nucleotide_base : public alphabet_base<derived_type, size, char>
{
private:
    using base_t = alphabet_base<derived_type, size, char>;
 
    constexpr nucleotide_base()                                    noexcept = default; 
    constexpr nucleotide_base(nucleotide_base const &)             noexcept = default; 
    constexpr nucleotide_base(nucleotide_base &&)                  noexcept = default; 
    constexpr nucleotide_base & operator=(nucleotide_base const &) noexcept = default; 
    constexpr nucleotide_base & operator=(nucleotide_base &&)      noexcept = default; 
    ~nucleotide_base()                                             noexcept = default; 
 
 
    friend derived_type;
 
protected:
    // Import from base:
    using typename base_t::char_type;
    using typename base_t::rank_type;
 
public:
    using base_t::alphabet_size;
    using base_t::to_rank;
 
    // This constructor needs to be public, because constructor templates are not inherited otherwise
    template <typename other_nucl_type>
        requires (!std::same_as<nucleotide_base, other_nucl_type>) &&
                 (!std::same_as<derived_type, other_nucl_type>) &&
                 nucleotide_alphabet<other_nucl_type>
    explicit constexpr nucleotide_base(other_nucl_type const & other) noexcept
    {
        static_cast<derived_type &>(*this) =
            detail::convert_through_char_representation<derived_type, other_nucl_type>[seqan3::to_rank(other)];
    }
 
    constexpr derived_type complement() const noexcept
    {
#ifdef SEQAN3_DEPRECATED_310
        if constexpr (detail::has_complement_table<derived_type>)
            return complement_table_deprecated(to_rank());
        else
            return derived_type{}.assign_rank(derived_type{}.rank_complement(to_rank()));
#else // ^^^ before 3.1.0 release / after 3.1.0 release vvv
        return derived_type{}.assign_rank(derived_type{}.rank_complement(to_rank()));
#endif // SEQAN3_DEPRECATED_310
    }
 
#ifdef SEQAN3_DEPRECATED_310
private:
 
    SEQAN3_DEPRECATED_310 static constexpr derived_type complement_table_deprecated(rank_type const rank) noexcept
    {
        return derived_type::complement_table[rank];
    }
 
public:
#endif // SEQAN3_DEPRECATED_310
 
    static constexpr bool char_is_valid(char_type const c) noexcept
    {
        return valid_char_table[static_cast<uint8_t>(c)];
    }
 
private:
    static constexpr std::array<bool, 256> valid_char_table
    {
        [] () constexpr
        {
            // init with false
            std::array<bool, 256> ret{};
 
            // the original valid chars and their lower cases
            for (size_t rank = 0u; rank < derived_type::alphabet_size; ++rank)
            {
                uint8_t c{};
#ifdef SEQAN3_DEPRECATED_310
                if constexpr (detail::has_rank_to_char_table<derived_type>)
                    c = derived_type::rank_to_char[rank];
                else
                    c = derived_type::rank_to_char(rank);
#else // ^^^ before 3.1.0 release / after 3.1.0 release vvv
                c = derived_type::rank_to_char(rank);
#endif // SEQAN3_DEPRECATED_310
                ret[c] = true;
                ret[to_lower(c)] = true;
            }
 
            // U and T shall be accepted for all
            ret['U'] = true;
            ret['T'] = true;
            ret['u'] = true;
            ret['t'] = true;
 
            return ret;
        }()
    };
};
 
} // namespace seqan3