search.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/std/algorithm>
#include <seqan3/std/ranges>
 
#include <seqan3/core/algorithm/algorithm_result_generator_range.hpp>
#include <seqan3/core/algorithm/detail/algorithm_executor_blocking.hpp>
#include <seqan3/core/configuration/configuration.hpp>
#include <seqan3/core/detail/persist_view.hpp>
#include <seqan3/search/configuration/default_configuration.hpp>
#include <seqan3/search/configuration/on_result.hpp>
#include <seqan3/search/configuration/parallel.hpp>
#include <seqan3/search/detail/search_configurator.hpp>
#include <seqan3/search/detail/search_traits.hpp>
#include <seqan3/utility/views/convert.hpp>
#include <seqan3/utility/views/deep.hpp>
#include <seqan3/utility/views/zip.hpp>
 
namespace seqan3::detail
{
struct search_configuration_validator
{
    template <typename query_t>
    static void validate_query_type()
    {
        using pure_query_t = std::remove_cvref_t<query_t>;
        if constexpr(range_dimension_v<pure_query_t> == 1u)
        {
            static_assert(std::ranges::random_access_range<pure_query_t>,
                          "The query sequence must model random_access_range.");
            static_assert(std::ranges::sized_range<pure_query_t>, "The query sequence must model sized_range.");
        }
        else
        {
            static_assert(std::ranges::forward_range<pure_query_t>, "The query collection must model forward_range.");
            static_assert(std::ranges::sized_range<pure_query_t>, "The query collection must model sized_range.");
            static_assert(std::ranges::random_access_range<std::ranges::range_value_t<pure_query_t>>,
                          "Elements of the query collection must model random_access_range.");
            static_assert(std::ranges::sized_range<std::ranges::range_value_t<pure_query_t>>,
                          "Elements of the query collection must model sized_range.");
        }
    }
};
} // namespace seqan3::detail
 
namespace seqan3
{
 
template <typename index_t,
          std::ranges::forward_range queries_t,
          typename configuration_t = decltype(search_cfg::default_configuration)>
    requires std::ranges::forward_range<std::ranges::range_reference_t<queries_t>> &&
             std::same_as<range_innermost_value_t<queries_t>, typename index_t::alphabet_type>
inline auto search(queries_t && queries,
                   index_t const & index,
                   configuration_t const & cfg = search_cfg::default_configuration)
{
    auto updated_cfg = detail::search_configurator::add_defaults(cfg);
 
    detail::search_configuration_validator::validate_query_type<queries_t>();
 
    size_t queries_size = std::ranges::distance(queries);
    auto indexed_queries = views::zip(std::views::iota(size_t{0}, queries_size), queries);
 
    using indexed_queries_t = decltype(indexed_queries);
 
    using query_t = std::ranges::range_reference_t<indexed_queries_t>;
    auto [algorithm, complete_config] = detail::search_configurator::configure_algorithm<query_t>(updated_cfg, index);
 
    using complete_configuration_t = decltype(complete_config);
    using traits_t = detail::search_traits<complete_configuration_t>;
    using algorithm_result_t = typename traits_t::search_result_type;
    using execution_handler_t = std::conditional_t<
                                    complete_configuration_t::template exists<search_cfg::parallel>(),
                                    detail::execution_handler_parallel,
                                    detail::execution_handler_sequential>;
 
    // Select the execution handler for the search configuration.
    auto select_execution_handler = [&] ()
    {
        if constexpr (std::same_as<execution_handler_t, detail::execution_handler_parallel>)
        {
            auto thread_count = get<search_cfg::parallel>(complete_config).thread_count;
            if (!thread_count)
                throw std::runtime_error{"You must configure the number of threads in seqan3::search_cfg::parallel."};
 
            return execution_handler_t{*thread_count};
        }
        else
        {
            return execution_handler_t{};
        }
    };
 
    // Finally, choose between two way execution returning an algorithm range or calling a user callback on every hit.
    if constexpr (traits_t::has_user_callback)
    {
        select_execution_handler().bulk_execute(algorithm,
                                                indexed_queries,
                                                get<search_cfg::on_result>(complete_config).callback);
    }
    else
    {
        using executor_t = detail::algorithm_executor_blocking<indexed_queries_t,
                                                               decltype(algorithm),
                                                               algorithm_result_t,
                                                               execution_handler_t>;
 
        return algorithm_result_generator_range{executor_t{std::move(indexed_queries),
                                                           std::move(algorithm),
                                                           algorithm_result_t{},
                                                           select_execution_handler()}};
    }
}
 
// Convert query sequence if it does not match the alphabet type of the index.
template <typename index_t,
          std::ranges::forward_range queries_t,
          typename configuration_t = decltype(search_cfg::default_configuration)>
    requires std::ranges::forward_range<std::ranges::range_reference_t<queries_t>> &&
             (!std::same_as<range_innermost_value_t<queries_t>, typename index_t::alphabet_type>)
inline auto search(queries_t && queries,
                   index_t const & index,
                   configuration_t const & cfg = search_cfg::default_configuration)
{
    static_assert(std::convertible_to<range_innermost_value_t<queries_t>, typename index_t::alphabet_type>,
                  "The alphabet of the text collection must be convertible to the alphabet of the index.");
 
    if constexpr (range_dimension_v<queries_t> == 2u)
        return search(queries | views::deep{views::convert<typename index_t::alphabet_type>}, index, cfg);
    else
        return search(queries | views::convert<typename index_t::alphabet_type>, index, cfg);
}
 
// Overload for a single query (not a collection of queries)
template <typename index_t,
          std::ranges::forward_range query_t,
          typename configuration_t = decltype(search_cfg::default_configuration)>
inline auto search(query_t && query,
                   index_t const & index,
                   configuration_t const & cfg = search_cfg::default_configuration)
{
    return search(std::views::single(std::forward<query_t>(query)), index, cfg);
}
 
template <typename index_t, typename configuration_t = decltype(search_cfg::default_configuration)>
inline auto search(char const * const queries,
                   index_t const & index,
                   configuration_t const & cfg = search_cfg::default_configuration)
{
    return search(std::string_view{queries}, index, cfg);
}
 
template <typename index_t, typename configuration_t = decltype(search_cfg::default_configuration)>
inline auto search(std::initializer_list<char const * const> const & queries,
                   index_t const & index,
                   configuration_t const & cfg = search_cfg::default_configuration)
{
    std::vector<std::string_view> query;
    query.reserve(std::ranges::size(queries));
    std::ranges::for_each(queries, [&query] (char const * const q) { query.push_back(std::string_view{q}); });
    return search(std::move(query) | detail::persist, index, cfg);
}
 
 
} // namespace seqan3