Merge pull request #131 from eseiler/misc/insert_iterator

[MISC] Allow insert_iterator on std::vector

include/hibf/config.hpp

@@ -7,14 +7,12 @@
 
 #pragma once
 
-#include <cinttypes>  // for uint16_t, uint32_t, uint64_t, uint8_t
+#include <cinttypes>  // for uint32_t, uint8_t
 #include <cstddef>    // for size_t
 #include <functional> // for function
 #include <iosfwd>     // for istream, ostream
-#include <iterator>   // for insert_iterator
 
-#include <hibf/cereal/path.hpp>        // IWYU pragma: keep
-#include <hibf/contrib/robin_hood.hpp> // for unordered_flat_set
+#include <hibf/misc/insert_iterator.hpp> // for insert_iterator
 #include <hibf/platform.hpp>
 
 #include <cereal/access.hpp> // for access
@@ -23,8 +21,6 @@
 namespace seqan::hibf
 {
 
-using insert_iterator = std::insert_iterator<robin_hood::unordered_flat_set<uint64_t>>;
-
 /*!\brief The configuration used to build an (H)IBF
  * \ingroup hibf
  *

include/hibf/misc/insert_iterator.hpp

@@ -0,0 +1,86 @@
+// --------------------------------------------------------------------------------------------------
+// Copyright (c) 2006-2022, Knut Reinert & Freie Universität Berlin
+// Copyright (c) 2016-2022, 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/raptor/blob/main/LICENSE.md
+// --------------------------------------------------------------------------------------------------
+
+#pragma once
+
+#include <algorithm> // for max
+#include <cassert>   // for assert
+#include <cinttypes> // for uint64_t
+#include <cstddef>   // for ptrdiff_t
+#include <iterator>  // for output_iterator_tag
+#include <utility>   // for addressof, move
+#include <vector>    // for vector
+
+#include <hibf/contrib/robin_hood.hpp> // for unordered_flat_set
+#include <hibf/platform.hpp>
+
+// IWYU pragma: private, include <hibf/config.hpp>
+
+namespace seqan::hibf
+{
+
+class insert_iterator
+{
+public:
+    using iterator_category = std::output_iterator_tag;
+    using value_type = void;
+    using difference_type = ptrdiff_t;
+    using pointer = void;
+    using reference = void;
+
+    insert_iterator() = delete;
+    insert_iterator(insert_iterator const &) = default;
+    insert_iterator(insert_iterator &&) = default;
+    insert_iterator & operator=(insert_iterator const &) = default;
+    insert_iterator & operator=(insert_iterator &&) = default;
+    ~insert_iterator() = default;
+
+    explicit constexpr insert_iterator(robin_hood::unordered_flat_set<uint64_t> & set) :
+        set{std::addressof(set)},
+        is_set{true}
+    {}
+
+    explicit constexpr insert_iterator(std::vector<uint64_t> & vec) : vec{std::addressof(vec)}, is_set{false}
+    {}
+
+    insert_iterator & operator=(uint64_t const value) noexcept
+    {
+        if (is_set)
+        {
+            assert(set != nullptr);
+            set->emplace(std::move(value));
+        }
+        else
+        {
+            assert(vec != nullptr);
+            vec->emplace_back(std::move(value));
+        }
+        return *this;
+    }
+
+    [[nodiscard]] constexpr insert_iterator & operator*() noexcept
+    {
+        return *this;
+    }
+
+    constexpr insert_iterator & operator++() noexcept
+    {
+        return *this;
+    }
+
+    constexpr insert_iterator operator++(int) noexcept
+    {
+        return *this;
+    }
+
+private:
+    robin_hood::unordered_flat_set<uint64_t> * set{nullptr};
+    std::vector<uint64_t> * vec{nullptr};
+    bool is_set{false};
+};
+
+} // namespace seqan::hibf

src/build/compute_kmers.cpp

@@ -30,7 +30,7 @@ void compute_kmers(robin_hood::unordered_flat_set<uint64_t> & kmers,
 {
     timer<concurrent::no> local_user_bin_io_timer{};
     local_user_bin_io_timer.start();
-    data.config.input_fn(record.idx, std::inserter(kmers, kmers.begin()));
+    data.config.input_fn(record.idx, insert_iterator{kmers});
     local_user_bin_io_timer.stop();
     data.user_bin_io_timer += local_user_bin_io_timer;
 }

src/build/insert_into_ibf.cpp

@@ -54,11 +54,11 @@ void insert_into_ibf(build_data const & data,
                      seqan::hibf::interleaved_bloom_filter & ibf)
 {
     auto const bin_index = seqan::hibf::bin_index{static_cast<size_t>(record.storage_TB_id)};
-    robin_hood::unordered_flat_set<uint64_t> values;
+    std::vector<uint64_t> values;
 
     timer<concurrent::no> local_user_bin_io_timer{};
     local_user_bin_io_timer.start();
-    data.config.input_fn(record.idx, std::inserter(values, values.begin()));
+    data.config.input_fn(record.idx, insert_iterator{values});
     local_user_bin_io_timer.stop();
     data.user_bin_io_timer += local_user_bin_io_timer;
 

src/interleaved_bloom_filter.cpp

@@ -57,7 +57,7 @@ size_t max_bin_size(config & configuration)
     for (size_t i = 0u; i < configuration.number_of_user_bins; ++i)
     {
         kmers.clear();
-        configuration.input_fn(i, std::inserter(kmers, kmers.begin()));
+        configuration.input_fn(i, insert_iterator{kmers});
 
 #pragma omp critical
         max_size = std::max(max_size, kmers.size());
@@ -81,7 +81,7 @@ interleaved_bloom_filter::interleaved_bloom_filter(config & configuration) :
     for (size_t i = 0u; i < configuration.number_of_user_bins; ++i)
     {
         kmers.clear();
-        configuration.input_fn(i, std::inserter(kmers, kmers.begin()));
+        configuration.input_fn(i, insert_iterator{kmers});
 
         for (uint64_t const hash : kmers)
             emplace(hash, seqan::hibf::bin_index{i});

src/sketch/compute_sketches.cpp

@@ -29,7 +29,7 @@ void compute_sketches(config const & config,
         seqan::hibf::sketch::hyperloglog sketch(config.sketch_bits);
 
         kmers.clear();
-        config.input_fn(i, std::inserter(kmers, kmers.begin()));
+        config.input_fn(i, insert_iterator{kmers});
 
         for (auto k_hash : kmers)
             sketch.add(k_hash);