edit_distance_trace_matrix_full.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 <bitset>
 
#include <seqan3/alignment/matrix/detail/trace_directions.hpp>
#include <seqan3/alignment/pairwise/edit_distance_fwd.hpp>
#include <seqan3/utility/detail/bits_of.hpp>
 
namespace seqan3::detail
{
 
template <typename word_t, bool is_semi_global, bool use_max_errors>
class edit_distance_trace_matrix_full
{
public:
    template <std::ranges::viewable_range database_t,
              std::ranges::viewable_range query_t,
              typename align_config_t,
              typename edit_traits>
    friend class edit_distance_unbanded;
 
    edit_distance_trace_matrix_full() = default;                                                    
    edit_distance_trace_matrix_full(edit_distance_trace_matrix_full const &) = default;             
    edit_distance_trace_matrix_full(edit_distance_trace_matrix_full &&) = default;                  
    edit_distance_trace_matrix_full & operator=(edit_distance_trace_matrix_full const &) = default; 
    edit_distance_trace_matrix_full & operator=(edit_distance_trace_matrix_full &&) = default;      
    ~edit_distance_trace_matrix_full() = default;                                                   
 
protected:
    template <typename derived_t, typename edit_traits>
    friend class edit_distance_unbanded_trace_matrix_policy;
 
    edit_distance_trace_matrix_full(size_t const rows_size) : rows_size{rows_size}, columns{}
    {}
 
private:
    struct trace_path_iterator;
 
public:
    using word_type = word_t;
 
    static constexpr auto word_size = bits_of<word_type>;
 
    using value_type = detail::trace_directions;
 
    using reference = value_type;
 
    using size_type = size_t;
 
    void reserve(size_t const new_capacity)
    {
        columns.reserve(new_capacity);
    }
 
public:
    reference at(matrix_coordinate const & coordinate) const noexcept
    {
        size_t row = coordinate.row;
        size_t col = coordinate.col;
 
        assert(row < rows());
        assert(col < cols());
 
        column_type const & column = columns[col];
 
        if constexpr (use_max_errors)
            if (!(row < column.max_rows))
                return detail::trace_directions::none;
 
        if (row == 0u)
        {
            if constexpr (is_semi_global)
                return detail::trace_directions::none;
 
            if (col == 0u)
                return detail::trace_directions::none;
 
            return detail::trace_directions::left;
        }
 
        size_t const idx = (row - 1u) / word_size;
        size_t const offset = (row - 1u) % word_size;
 
        bool const left = std::bitset<word_size>(column.left[idx])[offset];
        bool const diagonal = std::bitset<word_size>(column.diagonal[idx])[offset];
        bool const up = std::bitset<word_size>(column.up[idx])[offset];
 
        auto const dir = (left ? detail::trace_directions::left : detail::trace_directions::none)
                       | (diagonal ? detail::trace_directions::diagonal : detail::trace_directions::none)
                       | (up ? detail::trace_directions::up : detail::trace_directions::none);
 
        return dir;
    }
 
    size_t rows() const noexcept
    {
        return rows_size;
    }
 
    size_t cols() const noexcept
    {
        return columns.size();
    }
 
    auto trace_path(matrix_coordinate const & trace_begin) const
    {
        if (trace_begin.row >= rows() || trace_begin.col >= cols())
            throw std::invalid_argument{"The given coordinate exceeds the matrix in vertical or horizontal direction."};
 
        using path_t = std::ranges::subrange<trace_path_iterator, std::default_sentinel_t>;
        return path_t{trace_path_iterator{this, trace_begin}, std::default_sentinel};
    }
 
protected:
    struct max_errors_state
    {
        size_t max_rows{};
    };
 
    struct trace_matrix_state
    {
        std::vector<word_type> left{};
        std::vector<word_type> diagonal{};
        std::vector<word_type> up{};
    };
 
    struct column_type : enable_state_t<true, trace_matrix_state>, enable_state_t<use_max_errors, max_errors_state>
    {};
 
    void add_column(std::vector<word_type> left, std::vector<word_type> diagonal, std::vector<word_type> up)
        requires (!use_max_errors)
    {
        column_type column{};
        column.left = std::move(left);
        column.diagonal = std::move(diagonal);
        column.up = std::move(up);
 
        columns.push_back(std::move(column));
    }
 
    void add_column(std::vector<word_type> left,
                    std::vector<word_type> diagonal,
                    std::vector<word_type> up,
                    size_t const max_rows)
        requires use_max_errors
    {
        column_type column{};
        column.left = std::move(left);
        column.diagonal = std::move(diagonal);
        column.up = std::move(up);
        column.max_rows = max_rows;
 
        columns.push_back(std::move(column));
    }
 
private:
    size_t rows_size{};
    std::vector<column_type> columns{};
};
 
template <typename word_t, bool is_semi_global, bool use_max_errors>
struct edit_distance_trace_matrix_full<word_t, is_semi_global, use_max_errors>::trace_path_iterator
{
    using iterator_category = std::input_iterator_tag;
    using value_type = detail::trace_directions;
    using difference_type = std::ptrdiff_t;
 
    static constexpr value_type D = value_type::diagonal;
    static constexpr value_type L = value_type::left;
    static constexpr value_type U = value_type::up;
    static constexpr value_type N = value_type::none;
 
    constexpr value_type operator*() const
    {
        value_type dir = parent->at(coordinate());
 
        if (dir == N)
            return N;
 
        if ((dir & L) == L)
            return L;
        else if ((dir & U) == U)
            return U;
        else
            return D;
    }
 
    [[nodiscard]] constexpr matrix_coordinate const & coordinate() const
    {
        return coordinate_;
    }
 
    constexpr trace_path_iterator & operator++()
    {
        value_type dir = *(*this);
 
        if ((dir & L) == L)
        {
            coordinate_.col = std::max<size_t>(coordinate_.col, 1) - 1;
        }
        else if ((dir & U) == U)
        {
            coordinate_.row = std::max<size_t>(coordinate_.row, 1) - 1;
        }
        else if ((dir & D) == D)
        {
            coordinate_.row = std::max<size_t>(coordinate_.row, 1) - 1;
            coordinate_.col = std::max<size_t>(coordinate_.col, 1) - 1;
        }
 
        // seqan3::trace_direction::none in an inner cell of the trace matrix is not possible in
        // non-local alignments, e.g. global, semi-global, max-error, banded. On the other hand,
        // this can happen in a cell of the first row or first colomn.
        assert(dir != N || coordinate_.row == 0 || coordinate_.col == 0);
 
        return *this;
    }
 
    constexpr void operator++(int)
    {
        ++(*this);
    }
 
    friend bool operator==(trace_path_iterator const & it, std::default_sentinel_t)
    {
        return *it == value_type::none;
    }
 
    friend bool operator==(std::default_sentinel_t, trace_path_iterator const & it)
    {
        return it == std::default_sentinel;
    }
 
    friend bool operator!=(trace_path_iterator const & it, std::default_sentinel_t)
    {
        return !(it == std::default_sentinel);
    }
 
    friend bool operator!=(std::default_sentinel_t, trace_path_iterator const & it)
    {
        return it != std::default_sentinel;
    }
 
    edit_distance_trace_matrix_full const * parent{nullptr};
    matrix_coordinate coordinate_{};
};
 
} // namespace seqan3::detail