datasketches 0.3.2 → 0.4.1

data/vendor/datasketches-cpp/req/include/req_sketch_impl.hpp

@@ -35,8 +35,8 @@ max_nom_size_(0),
 num_retained_(0),
 n_(0),
 compactors_(allocator),
-min_item_(nullptr),
-max_item_(nullptr),
+min_item_(),
+max_item_(),
 sorted_view_(nullptr)
 {
   grow();
@@ -44,14 +44,6 @@ sorted_view_(nullptr)
 
 template<typename T, typename C, typename A>
 req_sketch<T, C, A>::~req_sketch() {
-  if (min_item_ != nullptr) {
-    min_item_->~T();
-    allocator_.deallocate(min_item_, 1);
-  }
-  if (max_item_ != nullptr) {
-    max_item_->~T();
-    allocator_.deallocate(max_item_, 1);
-  }
   reset_sorted_view();
 }
 
@@ -65,13 +57,10 @@ max_nom_size_(other.max_nom_size_),
 num_retained_(other.num_retained_),
 n_(other.n_),
 compactors_(other.compactors_),
-min_item_(nullptr),
-max_item_(nullptr),
+min_item_(other.min_item_),
+max_item_(other.max_item_),
 sorted_view_(nullptr)
-{
-  if (other.min_item_ != nullptr) min_item_ = new (allocator_.allocate(1)) T(*other.min_item_);
-  if (other.max_item_ != nullptr) max_item_ = new (allocator_.allocate(1)) T(*other.max_item_);
-}
+{}
 
 template<typename T, typename C, typename A>
 req_sketch<T, C, A>::req_sketch(req_sketch&& other) noexcept :
@@ -83,13 +72,10 @@ max_nom_size_(other.max_nom_size_),
 num_retained_(other.num_retained_),
 n_(other.n_),
 compactors_(std::move(other.compactors_)),
-min_item_(other.min_item_),
-max_item_(other.max_item_),
+min_item_(std::move(other.min_item_)),
+max_item_(std::move(other.max_item_)),
 sorted_view_(nullptr)
-{
-  other.min_item_ = nullptr;
-  other.max_item_ = nullptr;
-}
+{}
 
 template<typename T, typename C, typename A>
 req_sketch<T, C, A>& req_sketch<T, C, A>::operator=(const req_sketch& other) {
@@ -135,8 +121,8 @@ max_nom_size_(other.max_nom_size_),
 num_retained_(other.num_retained_),
 n_(other.n_),
 compactors_(allocator),
-min_item_(nullptr),
-max_item_(nullptr),
+min_item_(other.min_item_),
+max_item_(other.max_item_),
 sorted_view_(nullptr)
 {
   static_assert(
@@ -147,10 +133,6 @@ sorted_view_(nullptr)
   for (const auto& compactor: other.compactors_) {
     compactors_.push_back(req_compactor<T, C, A>(compactor, comparator_, allocator_));
   }
-  if (!other.is_empty()) {
-    min_item_ = new (allocator_.allocate(1)) T(other.get_min_item());
-    max_item_ = new (allocator_.allocate(1)) T(other.get_max_item());
-  }
 }
 
 template<typename T, typename C, typename A>
@@ -188,8 +170,8 @@ template<typename FwdT>
 void req_sketch<T, C, A>::update(FwdT&& item) {
   if (!check_update_item(item)) { return; }
   if (is_empty()) {
-    min_item_ = new (allocator_.allocate(1)) T(item);
-    max_item_ = new (allocator_.allocate(1)) T(item);
+    min_item_.emplace(item);
+    max_item_.emplace(item);
   } else {
     if (comparator_(item, *min_item_)) *min_item_ = item;
     if (comparator_(*max_item_, item)) *max_item_ = item;
@@ -207,8 +189,8 @@ void req_sketch<T, C, A>::merge(FwdSk&& other) {
   if (is_HRA() != other.is_HRA()) throw std::invalid_argument("merging HRA and LRA is not valid");
   if (other.is_empty()) return;
   if (is_empty()) {
-    min_item_ = new (allocator_.allocate(1)) T(conditional_forward<FwdSk>(*other.min_item_));
-    max_item_ = new (allocator_.allocate(1)) T(conditional_forward<FwdSk>(*other.max_item_));
+    min_item_.emplace(conditional_forward<FwdSk>(*other.min_item_));
+    max_item_.emplace(conditional_forward<FwdSk>(*other.max_item_));
   } else {
     if (comparator_(*other.min_item_, *min_item_)) *min_item_ = conditional_forward<FwdSk>(*other.min_item_);
     if (comparator_(*max_item_, *other.max_item_)) *max_item_ = conditional_forward<FwdSk>(*other.max_item_);
@@ -283,25 +265,6 @@ auto req_sketch<T, C, A>::get_quantile(double rank, bool inclusive) const -> qua
   return sorted_view_->get_quantile(rank, inclusive);
 }
 
-template<typename T, typename C, typename A>
-std::vector<T, A> req_sketch<T, C, A>::get_quantiles(const double* ranks, uint32_t size, bool inclusive) const {
-  if (is_empty()) throw std::runtime_error("operation is undefined for an empty sketch");
-  std::vector<T, A> quantiles(allocator_);
-  quantiles.reserve(size);
-
-  // possible side-effect of sorting level zero
-  setup_sorted_view();
-
-  for (uint32_t i = 0; i < size; ++i) {
-    const double rank = ranks[i];
-    if ((rank < 0.0) || (rank > 1.0)) {
-      throw std::invalid_argument("Normalized rank cannot be less than 0 or greater than 1");
-    }
-    quantiles.push_back(sorted_view_->get_quantile(rank, inclusive));
-  }
-  return quantiles;
-}
-
 template<typename T, typename C, typename A>
 quantiles_sorted_view<T, C, A> req_sketch<T, C, A>::get_sorted_view() const {
   if (!compactors_[0].is_sorted()) {
@@ -310,7 +273,7 @@ quantiles_sorted_view<T, C, A> req_sketch<T, C, A>::get_sorted_view() const {
   quantiles_sorted_view<T, C, A> view(get_num_retained(), comparator_, allocator_);
 
   for (auto& compactor: compactors_) {
-    view.add(compactor.begin(), compactor.end(), 1 << compactor.get_lg_weight());
+    view.add(compactor.begin(), compactor.end(), 1ULL << compactor.get_lg_weight());
   }
 
   view.convert_to_cummulative();
@@ -426,8 +389,8 @@ void req_sketch<T, C, A>::serialize(std::ostream& os, const SerDe& sd) const {
   if (is_empty()) return;
   if (is_estimation_mode()) {
     write(os, n_);
-    sd.serialize(os, min_item_, 1);
-    sd.serialize(os, max_item_, 1);
+    sd.serialize(os, &*min_item_, 1);
+    sd.serialize(os, &*max_item_, 1);
   }
   if (raw_items) {
     sd.serialize(os, compactors_[0].begin(), num_raw_items);
@@ -466,8 +429,8 @@ auto req_sketch<T, C, A>::serialize(unsigned header_size_bytes, const SerDe& sd)
   if (!is_empty()) {
     if (is_estimation_mode()) {
       ptr += copy_to_mem(n_, ptr);
-      ptr += sd.serialize(ptr, end_ptr - ptr, min_item_, 1);
-      ptr += sd.serialize(ptr, end_ptr - ptr, max_item_, 1);
+      ptr += sd.serialize(ptr, end_ptr - ptr, &*min_item_, 1);
+      ptr += sd.serialize(ptr, end_ptr - ptr, &*max_item_, 1);
     }
     if (raw_items) {
       ptr += sd.serialize(ptr, end_ptr - ptr, compactors_[0].begin(), num_raw_items);
@@ -498,12 +461,9 @@ req_sketch<T, C, A> req_sketch<T, C, A>::deserialize(std::istream& is, const Ser
   const bool hra = flags_byte & (1 << flags::IS_HIGH_RANK);
   if (is_empty) return req_sketch(k, hra, comparator, allocator);
 
-  A alloc(allocator);
-  auto item_buffer_deleter = [&alloc](T* ptr) { alloc.deallocate(ptr, 1); };
-  std::unique_ptr<T, decltype(item_buffer_deleter)> min_item_buffer(alloc.allocate(1), item_buffer_deleter);
-  std::unique_ptr<T, decltype(item_buffer_deleter)> max_item_buffer(alloc.allocate(1), item_buffer_deleter);
-  std::unique_ptr<T, item_deleter> min_item(nullptr, item_deleter(allocator));
-  std::unique_ptr<T, item_deleter> max_item(nullptr, item_deleter(allocator));
+  optional<T> tmp; // space to deserialize min and max
+  optional<T> min_item;
+  optional<T> max_item;
 
   const bool raw_items = flags_byte & (1 << flags::RAW_ITEMS);
   const bool is_level_0_sorted = flags_byte & (1 << flags::IS_LEVEL_ZERO_SORTED);
@@ -512,12 +472,14 @@ req_sketch<T, C, A> req_sketch<T, C, A>::deserialize(std::istream& is, const Ser
   uint64_t n = 1;
   if (num_levels > 1) {
     n = read<uint64_t>(is);
-    sd.deserialize(is, min_item_buffer.get(), 1);
-    // serde call did not throw, repackage with destrtuctor
-    min_item = std::unique_ptr<T, item_deleter>(min_item_buffer.release(), item_deleter(allocator));
-    sd.deserialize(is, max_item_buffer.get(), 1);
-    // serde call did not throw, repackage with destrtuctor
-    max_item = std::unique_ptr<T, item_deleter>(max_item_buffer.release(), item_deleter(allocator));
+    sd.deserialize(is, &*tmp, 1);
+    // serde call did not throw, repackage and cleanup
+    min_item.emplace(*tmp);
+    (*tmp).~T();
+    sd.deserialize(is, &*tmp, 1);
+    // serde call did not throw, repackage and cleanup
+    max_item.emplace(*tmp);
+    (*tmp).~T();
   }
 
   if (raw_items) {
@@ -537,12 +499,8 @@ req_sketch<T, C, A> req_sketch<T, C, A>::deserialize(std::istream& is, const Ser
       if (comparator(*it, *min_it)) min_it = it;
       if (comparator(*max_it, *it)) max_it = it;
     }
-    new (min_item_buffer.get()) T(*min_it);
-    // copy did not throw, repackage with destrtuctor
-    min_item = std::unique_ptr<T, item_deleter>(min_item_buffer.release(), item_deleter(allocator));
-    new (max_item_buffer.get()) T(*max_it);
-    // copy did not throw, repackage with destrtuctor
-    max_item = std::unique_ptr<T, item_deleter>(max_item_buffer.release(), item_deleter(allocator));
+    min_item.emplace(*min_it);
+    max_item.emplace(*max_it);
   }
 
   if (!is.good()) throw std::runtime_error("error reading from std::istream");
@@ -579,12 +537,9 @@ req_sketch<T, C, A> req_sketch<T, C, A>::deserialize(const void* bytes, size_t s
   const bool hra = flags_byte & (1 << flags::IS_HIGH_RANK);
   if (is_empty) return req_sketch(k, hra, comparator, allocator);
 
-  A alloc(allocator);
-  auto item_buffer_deleter = [&alloc](T* ptr) { alloc.deallocate(ptr, 1); };
-  std::unique_ptr<T, decltype(item_buffer_deleter)> min_item_buffer(alloc.allocate(1), item_buffer_deleter);
-  std::unique_ptr<T, decltype(item_buffer_deleter)> max_item_buffer(alloc.allocate(1), item_buffer_deleter);
-  std::unique_ptr<T, item_deleter> min_item(nullptr, item_deleter(allocator));
-  std::unique_ptr<T, item_deleter> max_item(nullptr, item_deleter(allocator));
+  optional<T> tmp; // space to deserialize min and max
+  optional<T> min_item;
+  optional<T> max_item;
 
   const bool raw_items = flags_byte & (1 << flags::RAW_ITEMS);
   const bool is_level_0_sorted = flags_byte & (1 << flags::IS_LEVEL_ZERO_SORTED);
@@ -594,12 +549,14 @@ req_sketch<T, C, A> req_sketch<T, C, A>::deserialize(const void* bytes, size_t s
   if (num_levels > 1) {
     ensure_minimum_memory(end_ptr - ptr, sizeof(n));
     ptr += copy_from_mem(ptr, n);
-    ptr += sd.deserialize(ptr, end_ptr - ptr, min_item_buffer.get(), 1);
-    // serde call did not throw, repackage with destrtuctor
-    min_item = std::unique_ptr<T, item_deleter>(min_item_buffer.release(), item_deleter(allocator));
-    ptr += sd.deserialize(ptr, end_ptr - ptr, max_item_buffer.get(), 1);
-    // serde call did not throw, repackage with destrtuctor
-    max_item = std::unique_ptr<T, item_deleter>(max_item_buffer.release(), item_deleter(allocator));
+    ptr += sd.deserialize(ptr, end_ptr - ptr, &*tmp, 1);
+    // serde call did not throw, repackage and cleanup
+    min_item.emplace(*tmp);
+    (*tmp).~T();
+    ptr += sd.deserialize(ptr, end_ptr - ptr, &*tmp, 1);
+    // serde call did not throw, repackage and cleanup
+    max_item.emplace(*tmp);
+    (*tmp).~T();
   }
 
   if (raw_items) {
@@ -623,12 +580,8 @@ req_sketch<T, C, A> req_sketch<T, C, A>::deserialize(const void* bytes, size_t s
       if (comparator(*it, *min_it)) min_it = it;
       if (comparator(*max_it, *it)) max_it = it;
     }
-    new (min_item_buffer.get()) T(*min_it);
-    // copy did not throw, repackage with destrtuctor
-    min_item = std::unique_ptr<T, item_deleter>(min_item_buffer.release(), item_deleter(allocator));
-    new (max_item_buffer.get()) T(*max_it);
-    // copy did not throw, repackage with destrtuctor
-    max_item = std::unique_ptr<T, item_deleter>(max_item_buffer.release(), item_deleter(allocator));
+    min_item.emplace(*min_it);
+    max_item.emplace(*max_it);
   }
 
   return req_sketch(k, hra, n, std::move(min_item), std::move(max_item), std::move(compactors), comparator);
@@ -721,23 +674,9 @@ string<A> req_sketch<T, C, A>::to_string(bool print_levels, bool print_items) co
   return string<A>(os.str().c_str(), allocator_);
 }
 
-template<typename T, typename C, typename A>
-class req_sketch<T, C, A>::item_deleter {
-  public:
-  item_deleter(const A& allocator): allocator_(allocator) {}
-  void operator() (T* ptr) {
-    if (ptr != nullptr) {
-      ptr->~T();
-      allocator_.deallocate(ptr, 1);
-    }
-  }
-  private:
-  A allocator_;
-};
-
 template<typename T, typename C, typename A>
 req_sketch<T, C, A>::req_sketch(uint16_t k, bool hra, uint64_t n,
-    std::unique_ptr<T, item_deleter> min_item, std::unique_ptr<T, item_deleter> max_item,
+    optional<T>&& min_item, optional<T>&& max_item,
     std::vector<Compactor, AllocCompactor>&& compactors, const C& comparator):
 comparator_(comparator),
 allocator_(compactors.get_allocator()),
@@ -747,8 +686,8 @@ max_nom_size_(0),
 num_retained_(0),
 n_(n),
 compactors_(std::move(compactors)),
-min_item_(min_item.release()),
-max_item_(max_item.release()),
+min_item_(std::move(min_item)),
+max_item_(std::move(max_item)),
 sorted_view_(nullptr)
 {
   update_max_nom_size();

data/vendor/datasketches-cpp/req/test/CMakeLists.txt

@@ -20,7 +20,6 @@ add_executable(req_test)
 target_link_libraries(req_test req common_test_lib)
 
 set_target_properties(req_test PROPERTIES
-  CXX_STANDARD 11
   CXX_STANDARD_REQUIRED YES
 )
 
@@ -41,3 +40,17 @@ target_sources(req_test
     req_sketch_test.cpp
     req_sketch_custom_type_test.cpp
 )
+
+if (SERDE_COMPAT)
+target_sources(req_test
+  PRIVATE
+    req_sketch_deserialize_from_java_test.cpp
+)
+endif()
+
+if (GENERATE)
+target_sources(req_test
+  PRIVATE
+    req_sketch_serialize_for_java.cpp
+)
+endif()

data/vendor/datasketches-cpp/req/test/req_sketch_deserialize_from_java_test.cpp

@@ -0,0 +1,55 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied.  See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+
+#include <catch2/catch.hpp>
+#include <fstream>
+#include <req_sketch.hpp>
+
+namespace datasketches {
+
+// assume the binary sketches for this test have been generated by datasketches-java code
+// in the subdirectory called "java" in the root directory of this project
+static std::string testBinaryInputPath = std::string(TEST_BINARY_INPUT_PATH) + "../../java/";
+
+TEST_CASE("req float", "[serde_compat]") {
+  const unsigned n_arr[] = {0, 1, 10, 100, 1000, 10000, 100000, 1000000};
+  for (const unsigned n: n_arr) {
+    std::ifstream is;
+    is.exceptions(std::ios::failbit | std::ios::badbit);
+    is.open(testBinaryInputPath + "req_float_n" + std::to_string(n) + "_java.sk", std::ios::binary);
+    const auto sketch = req_sketch<float>::deserialize(is);
+    REQUIRE(sketch.is_HRA());
+    REQUIRE(sketch.is_empty() == (n == 0));
+    REQUIRE(sketch.is_estimation_mode() == (n > 10));
+    REQUIRE(sketch.get_n() == n);
+    if (n > 0) {
+      REQUIRE(sketch.get_min_item() == 1.0f);
+      REQUIRE(sketch.get_max_item() == static_cast<float>(n));
+      uint64_t weight = 0;
+      for (const auto pair: sketch) {
+        REQUIRE(pair.first >= sketch.get_min_item());
+        REQUIRE(pair.first <= sketch.get_max_item());
+        weight += pair.second;
+      }
+      REQUIRE(weight == sketch.get_n());
+    }
+  }
+}
+
+} /* namespace datasketches */

data/vendor/datasketches-cpp/{tuple/include/array_of_doubles_intersection_impl.hpp → req/test/req_sketch_serialize_for_java.cpp}

@@ -17,15 +17,20 @@
  * under the License.
  */
 
-namespace datasketches {
+#include <catch2/catch.hpp>
+#include <fstream>
+#include <req_sketch.hpp>
 
-template<typename P, typename A>
-array_of_doubles_intersection<P, A>::array_of_doubles_intersection(uint64_t seed, const P& policy, const A& allocator):
-Base(seed, policy, allocator) {}
+namespace datasketches {
 
-template<typename P, typename A>
-auto array_of_doubles_intersection<P, A>::get_result(bool ordered) const -> CompactSketch {
-  return compact_array_of_doubles_sketch_alloc<A>(this->state_.get_policy().get_policy().get_num_values(), Base::get_result(ordered));
+TEST_CASE("req sketch float generate", "[serialize_for_java]") {
+  const unsigned n_arr[] = {0, 1, 10, 100, 1000, 10000, 100000, 1000000};
+  for (const unsigned n: n_arr) {
+    req_sketch<float> sketch(12);
+    for (unsigned i = 1; i <= n; ++i) sketch.update(i);
+    std::ofstream os("req_float_n" + std::to_string(n) + "_cpp.sk", std::ios::binary);
+    sketch.serialize(os);
+  }
 }
 
 } /* namespace datasketches */

data/vendor/datasketches-cpp/req/test/req_sketch_test.cpp

@@ -47,8 +47,6 @@ TEST_CASE("req sketch: empty", "[req_sketch]") {
   REQUIRE_THROWS_AS(sketch.get_max_item(), std::runtime_error);
   REQUIRE_THROWS_AS(sketch.get_rank(0), std::runtime_error);
   REQUIRE_THROWS_AS(sketch.get_quantile(0), std::runtime_error);
-  const double ranks[3] {0, 0.5, 1};
-  REQUIRE_THROWS_AS(sketch.get_quantiles(ranks, 3), std::runtime_error);
 
   const float split_points[1] {0};
   REQUIRE_THROWS_AS(sketch.get_CDF(split_points, 1), std::runtime_error);
@@ -71,13 +69,6 @@ TEST_CASE("req sketch: single value, lra", "[req_sketch]") {
   REQUIRE(sketch.get_quantile(0.5, false) == 1);
   REQUIRE(sketch.get_quantile(1, false) == 1);
 
-  const double ranks[3] {0, 0.5, 1};
-  auto quantiles = sketch.get_quantiles(ranks, 3);
-  REQUIRE(quantiles.size() == 3);
-  REQUIRE(quantiles[0] == 1);
-  REQUIRE(quantiles[1] == 1);
-  REQUIRE(quantiles[2] == 1);
-
   unsigned count = 0;
   for (auto pair: sketch) {
     REQUIRE(pair.second == 1);
@@ -145,13 +136,6 @@ TEST_CASE("req sketch: exact mode", "[req_sketch]") {
   REQUIRE(sketch.get_quantile(0.9) == 9);
   REQUIRE(sketch.get_quantile(1) == 10);
 
-  const double ranks[3] {0, 0.5, 1};
-  auto quantiles = sketch.get_quantiles(ranks, 3);
-  REQUIRE(quantiles.size() == 3);
-  REQUIRE(quantiles[0] == 1);
-  REQUIRE(quantiles[1] == 5);
-  REQUIRE(quantiles[2] == 10);
-
   const float splits[3] {2, 6, 9};
   auto cdf = sketch.get_CDF(splits, 3, false);
   REQUIRE(cdf[0] == 0.1);
@@ -356,76 +340,6 @@ TEST_CASE("req sketch: serialize deserialize stream and bytes equivalence", "[re
   REQUIRE(sketch2.get_max_item() == sketch.get_max_item());
 }
 
-TEST_CASE("req sketch: stream deserialize from Java - empty", "[req_sketch]") {
-  std::ifstream is;
-  is.exceptions(std::ios::failbit | std::ios::badbit);
-  is.open(input_path + "req_float_empty_from_java.sk", std::ios::binary);
-  auto sketch = req_sketch<float>::deserialize(is);
-  REQUIRE(sketch.is_empty());
-  REQUIRE_FALSE(sketch.is_estimation_mode());
-  REQUIRE(sketch.get_n() == 0);
-  REQUIRE(sketch.get_num_retained() == 0);
-  REQUIRE_THROWS_AS(sketch.get_min_item(), std::runtime_error);
-  REQUIRE_THROWS_AS(sketch.get_max_item(), std::runtime_error);
-}
-
-TEST_CASE("req sketch: stream deserialize from Java - single item", "[req_sketch]") {
-  std::ifstream is;
-  is.exceptions(std::ios::failbit | std::ios::badbit);
-  is.open(input_path + "req_float_single_item_from_java.sk", std::ios::binary);
-  auto sketch = req_sketch<float>::deserialize(is);
-  REQUIRE_FALSE(sketch.is_empty());
-  REQUIRE_FALSE(sketch.is_estimation_mode());
-  REQUIRE(sketch.get_n() == 1);
-  REQUIRE(sketch.get_num_retained() == 1);
-  REQUIRE(sketch.get_min_item() == 1);
-  REQUIRE(sketch.get_max_item() == 1);
-  REQUIRE(sketch.get_rank(1.0f, false) == 0);
-  REQUIRE(sketch.get_rank(1.0f) == 1);
-}
-
-TEST_CASE("req sketch: stream deserialize from Java - raw items", "[req_sketch]") {
-  std::ifstream is;
-  is.exceptions(std::ios::failbit | std::ios::badbit);
-  is.open(input_path + "req_float_raw_items_from_java.sk", std::ios::binary);
-  auto sketch = req_sketch<float>::deserialize(is);
-  REQUIRE_FALSE(sketch.is_empty());
-  REQUIRE_FALSE(sketch.is_estimation_mode());
-  REQUIRE(sketch.get_n() == 4);
-  REQUIRE(sketch.get_num_retained() == 4);
-  REQUIRE(sketch.get_min_item() == 0);
-  REQUIRE(sketch.get_max_item() == 3);
-  REQUIRE(sketch.get_rank(2.0f, false) == 0.5);
-}
-
-TEST_CASE("req sketch: stream deserialize from Java - exact mode", "[req_sketch]") {
-  std::ifstream is;
-  is.exceptions(std::ios::failbit | std::ios::badbit);
-  is.open(input_path + "req_float_exact_from_java.sk", std::ios::binary);
-  auto sketch = req_sketch<float>::deserialize(is);
-  REQUIRE_FALSE(sketch.is_empty());
-  REQUIRE_FALSE(sketch.is_estimation_mode());
-  REQUIRE(sketch.get_n() == 100);
-  REQUIRE(sketch.get_num_retained() == 100);
-  REQUIRE(sketch.get_min_item() == 0);
-  REQUIRE(sketch.get_max_item() == 99);
-  REQUIRE(sketch.get_rank(50.0f, false) == 0.5);
-}
-
-TEST_CASE("req sketch: stream deserialize from Java - estimation mode", "[req_sketch]") {
-  std::ifstream is;
-  is.exceptions(std::ios::failbit | std::ios::badbit);
-  is.open(input_path + "req_float_estimation_from_java.sk", std::ios::binary);
-  auto sketch = req_sketch<float>::deserialize(is);
-  REQUIRE_FALSE(sketch.is_empty());
-  REQUIRE(sketch.is_estimation_mode());
-  REQUIRE(sketch.get_n() == 10000);
-  REQUIRE(sketch.get_num_retained() == 2942);
-  REQUIRE(sketch.get_min_item() == 0);
-  REQUIRE(sketch.get_max_item() == 9999);
-  REQUIRE(sketch.get_rank(5000.0f, false) == 0.5);
-}
-
 TEST_CASE("req sketch: merge into empty", "[req_sketch]") {
   req_sketch<float> sketch1(40);
 
@@ -555,11 +469,11 @@ TEST_CASE("req sketch: type conversion - custom types") {
 
 TEST_CASE("get_rank equivalence") {
   req_sketch<int> sketch(12);
-  const size_t n = 1000;
-  for (size_t i = 0; i < n; ++i) sketch.update(i);
+  const int n = 1000;
+  for (int i = 0; i < n; ++i) sketch.update(i);
   REQUIRE(sketch.get_n() == n);
   auto view = sketch.get_sorted_view();
-  for (size_t i = 0; i < n; ++i) {
+  for (int i = 0; i < n; ++i) {
     REQUIRE(sketch.get_rank(i) == view.get_rank(i));
   }
 }

data/vendor/datasketches-cpp/sampling/CMakeLists.txt

@@ -41,4 +41,8 @@ install(FILES
 		include/var_opt_sketch_impl.hpp
 		include/var_opt_union.hpp
 		include/var_opt_union_impl.hpp
+    include/ebpps_sample.hpp
+    include/ebpps_sample_impl.hpp
+    include/ebpps_sketch.hpp
+    include/ebpps_sketch_impl.hpp
   DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}/DataSketches")