datasketches 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/CHANGELOG.md +3 -0
- data/LICENSE +310 -0
- data/NOTICE +11 -0
- data/README.md +126 -0
- data/ext/datasketches/cpc_wrapper.cpp +50 -0
- data/ext/datasketches/ext.cpp +12 -0
- data/ext/datasketches/extconf.rb +11 -0
- data/ext/datasketches/hll_wrapper.cpp +69 -0
- data/lib/datasketches.rb +9 -0
- data/lib/datasketches/version.rb +3 -0
- data/vendor/datasketches-cpp/CMakeLists.txt +126 -0
- data/vendor/datasketches-cpp/LICENSE +311 -0
- data/vendor/datasketches-cpp/MANIFEST.in +19 -0
- data/vendor/datasketches-cpp/NOTICE +11 -0
- data/vendor/datasketches-cpp/README.md +42 -0
- data/vendor/datasketches-cpp/common/CMakeLists.txt +45 -0
- data/vendor/datasketches-cpp/common/include/MurmurHash3.h +173 -0
- data/vendor/datasketches-cpp/common/include/binomial_bounds.hpp +458 -0
- data/vendor/datasketches-cpp/common/include/bounds_binomial_proportions.hpp +291 -0
- data/vendor/datasketches-cpp/common/include/ceiling_power_of_2.hpp +41 -0
- data/vendor/datasketches-cpp/common/include/common_defs.hpp +51 -0
- data/vendor/datasketches-cpp/common/include/conditional_back_inserter.hpp +68 -0
- data/vendor/datasketches-cpp/common/include/conditional_forward.hpp +70 -0
- data/vendor/datasketches-cpp/common/include/count_zeros.hpp +114 -0
- data/vendor/datasketches-cpp/common/include/inv_pow2_table.hpp +107 -0
- data/vendor/datasketches-cpp/common/include/memory_operations.hpp +57 -0
- data/vendor/datasketches-cpp/common/include/serde.hpp +196 -0
- data/vendor/datasketches-cpp/common/test/CMakeLists.txt +38 -0
- data/vendor/datasketches-cpp/common/test/catch.hpp +17618 -0
- data/vendor/datasketches-cpp/common/test/catch_runner.cpp +7 -0
- data/vendor/datasketches-cpp/common/test/test_allocator.cpp +31 -0
- data/vendor/datasketches-cpp/common/test/test_allocator.hpp +108 -0
- data/vendor/datasketches-cpp/common/test/test_runner.cpp +29 -0
- data/vendor/datasketches-cpp/common/test/test_type.hpp +137 -0
- data/vendor/datasketches-cpp/cpc/CMakeLists.txt +74 -0
- data/vendor/datasketches-cpp/cpc/include/compression_data.hpp +6022 -0
- data/vendor/datasketches-cpp/cpc/include/cpc_common.hpp +62 -0
- data/vendor/datasketches-cpp/cpc/include/cpc_compressor.hpp +147 -0
- data/vendor/datasketches-cpp/cpc/include/cpc_compressor_impl.hpp +742 -0
- data/vendor/datasketches-cpp/cpc/include/cpc_confidence.hpp +167 -0
- data/vendor/datasketches-cpp/cpc/include/cpc_sketch.hpp +311 -0
- data/vendor/datasketches-cpp/cpc/include/cpc_sketch_impl.hpp +810 -0
- data/vendor/datasketches-cpp/cpc/include/cpc_union.hpp +102 -0
- data/vendor/datasketches-cpp/cpc/include/cpc_union_impl.hpp +346 -0
- data/vendor/datasketches-cpp/cpc/include/cpc_util.hpp +137 -0
- data/vendor/datasketches-cpp/cpc/include/icon_estimator.hpp +274 -0
- data/vendor/datasketches-cpp/cpc/include/kxp_byte_lookup.hpp +81 -0
- data/vendor/datasketches-cpp/cpc/include/u32_table.hpp +84 -0
- data/vendor/datasketches-cpp/cpc/include/u32_table_impl.hpp +266 -0
- data/vendor/datasketches-cpp/cpc/test/CMakeLists.txt +44 -0
- data/vendor/datasketches-cpp/cpc/test/compression_test.cpp +67 -0
- data/vendor/datasketches-cpp/cpc/test/cpc_sketch_test.cpp +381 -0
- data/vendor/datasketches-cpp/cpc/test/cpc_union_test.cpp +149 -0
- data/vendor/datasketches-cpp/fi/CMakeLists.txt +54 -0
- data/vendor/datasketches-cpp/fi/include/frequent_items_sketch.hpp +319 -0
- data/vendor/datasketches-cpp/fi/include/frequent_items_sketch_impl.hpp +484 -0
- data/vendor/datasketches-cpp/fi/include/reverse_purge_hash_map.hpp +114 -0
- data/vendor/datasketches-cpp/fi/include/reverse_purge_hash_map_impl.hpp +345 -0
- data/vendor/datasketches-cpp/fi/test/CMakeLists.txt +44 -0
- data/vendor/datasketches-cpp/fi/test/frequent_items_sketch_custom_type_test.cpp +84 -0
- data/vendor/datasketches-cpp/fi/test/frequent_items_sketch_test.cpp +360 -0
- data/vendor/datasketches-cpp/fi/test/items_sketch_string_from_java.sk +0 -0
- data/vendor/datasketches-cpp/fi/test/items_sketch_string_utf8_from_java.sk +0 -0
- data/vendor/datasketches-cpp/fi/test/longs_sketch_from_java.sk +0 -0
- data/vendor/datasketches-cpp/fi/test/reverse_purge_hash_map_test.cpp +47 -0
- data/vendor/datasketches-cpp/hll/CMakeLists.txt +92 -0
- data/vendor/datasketches-cpp/hll/include/AuxHashMap-internal.hpp +303 -0
- data/vendor/datasketches-cpp/hll/include/AuxHashMap.hpp +83 -0
- data/vendor/datasketches-cpp/hll/include/CompositeInterpolationXTable-internal.hpp +811 -0
- data/vendor/datasketches-cpp/hll/include/CompositeInterpolationXTable.hpp +40 -0
- data/vendor/datasketches-cpp/hll/include/CouponHashSet-internal.hpp +291 -0
- data/vendor/datasketches-cpp/hll/include/CouponHashSet.hpp +59 -0
- data/vendor/datasketches-cpp/hll/include/CouponList-internal.hpp +417 -0
- data/vendor/datasketches-cpp/hll/include/CouponList.hpp +91 -0
- data/vendor/datasketches-cpp/hll/include/CubicInterpolation-internal.hpp +233 -0
- data/vendor/datasketches-cpp/hll/include/CubicInterpolation.hpp +43 -0
- data/vendor/datasketches-cpp/hll/include/HarmonicNumbers-internal.hpp +90 -0
- data/vendor/datasketches-cpp/hll/include/HarmonicNumbers.hpp +48 -0
- data/vendor/datasketches-cpp/hll/include/Hll4Array-internal.hpp +335 -0
- data/vendor/datasketches-cpp/hll/include/Hll4Array.hpp +69 -0
- data/vendor/datasketches-cpp/hll/include/Hll6Array-internal.hpp +124 -0
- data/vendor/datasketches-cpp/hll/include/Hll6Array.hpp +55 -0
- data/vendor/datasketches-cpp/hll/include/Hll8Array-internal.hpp +158 -0
- data/vendor/datasketches-cpp/hll/include/Hll8Array.hpp +56 -0
- data/vendor/datasketches-cpp/hll/include/HllArray-internal.hpp +706 -0
- data/vendor/datasketches-cpp/hll/include/HllArray.hpp +136 -0
- data/vendor/datasketches-cpp/hll/include/HllSketch-internal.hpp +462 -0
- data/vendor/datasketches-cpp/hll/include/HllSketchImpl-internal.hpp +149 -0
- data/vendor/datasketches-cpp/hll/include/HllSketchImpl.hpp +85 -0
- data/vendor/datasketches-cpp/hll/include/HllSketchImplFactory.hpp +170 -0
- data/vendor/datasketches-cpp/hll/include/HllUnion-internal.hpp +287 -0
- data/vendor/datasketches-cpp/hll/include/HllUtil.hpp +239 -0
- data/vendor/datasketches-cpp/hll/include/RelativeErrorTables-internal.hpp +112 -0
- data/vendor/datasketches-cpp/hll/include/RelativeErrorTables.hpp +46 -0
- data/vendor/datasketches-cpp/hll/include/coupon_iterator-internal.hpp +56 -0
- data/vendor/datasketches-cpp/hll/include/coupon_iterator.hpp +43 -0
- data/vendor/datasketches-cpp/hll/include/hll.hpp +669 -0
- data/vendor/datasketches-cpp/hll/include/hll.private.hpp +32 -0
- data/vendor/datasketches-cpp/hll/test/AuxHashMapTest.cpp +79 -0
- data/vendor/datasketches-cpp/hll/test/CMakeLists.txt +51 -0
- data/vendor/datasketches-cpp/hll/test/CouponHashSetTest.cpp +130 -0
- data/vendor/datasketches-cpp/hll/test/CouponListTest.cpp +181 -0
- data/vendor/datasketches-cpp/hll/test/CrossCountingTest.cpp +93 -0
- data/vendor/datasketches-cpp/hll/test/HllArrayTest.cpp +191 -0
- data/vendor/datasketches-cpp/hll/test/HllSketchTest.cpp +389 -0
- data/vendor/datasketches-cpp/hll/test/HllUnionTest.cpp +313 -0
- data/vendor/datasketches-cpp/hll/test/IsomorphicTest.cpp +141 -0
- data/vendor/datasketches-cpp/hll/test/TablesTest.cpp +44 -0
- data/vendor/datasketches-cpp/hll/test/ToFromByteArrayTest.cpp +168 -0
- data/vendor/datasketches-cpp/hll/test/array6_from_java.sk +0 -0
- data/vendor/datasketches-cpp/hll/test/compact_array4_from_java.sk +0 -0
- data/vendor/datasketches-cpp/hll/test/compact_set_from_java.sk +0 -0
- data/vendor/datasketches-cpp/hll/test/list_from_java.sk +0 -0
- data/vendor/datasketches-cpp/hll/test/updatable_array4_from_java.sk +0 -0
- data/vendor/datasketches-cpp/hll/test/updatable_set_from_java.sk +0 -0
- data/vendor/datasketches-cpp/kll/CMakeLists.txt +58 -0
- data/vendor/datasketches-cpp/kll/include/kll_helper.hpp +150 -0
- data/vendor/datasketches-cpp/kll/include/kll_helper_impl.hpp +319 -0
- data/vendor/datasketches-cpp/kll/include/kll_quantile_calculator.hpp +67 -0
- data/vendor/datasketches-cpp/kll/include/kll_quantile_calculator_impl.hpp +169 -0
- data/vendor/datasketches-cpp/kll/include/kll_sketch.hpp +559 -0
- data/vendor/datasketches-cpp/kll/include/kll_sketch_impl.hpp +1131 -0
- data/vendor/datasketches-cpp/kll/test/CMakeLists.txt +44 -0
- data/vendor/datasketches-cpp/kll/test/kll_sketch_custom_type_test.cpp +154 -0
- data/vendor/datasketches-cpp/kll/test/kll_sketch_float_one_item_v1.sk +0 -0
- data/vendor/datasketches-cpp/kll/test/kll_sketch_from_java.sk +0 -0
- data/vendor/datasketches-cpp/kll/test/kll_sketch_test.cpp +685 -0
- data/vendor/datasketches-cpp/kll/test/kll_sketch_validation.cpp +229 -0
- data/vendor/datasketches-cpp/pyproject.toml +17 -0
- data/vendor/datasketches-cpp/python/CMakeLists.txt +61 -0
- data/vendor/datasketches-cpp/python/README.md +78 -0
- data/vendor/datasketches-cpp/python/jupyter/CPCSketch.ipynb +345 -0
- data/vendor/datasketches-cpp/python/jupyter/FrequentItemsSketch.ipynb +354 -0
- data/vendor/datasketches-cpp/python/jupyter/HLLSketch.ipynb +346 -0
- data/vendor/datasketches-cpp/python/jupyter/KLLSketch.ipynb +463 -0
- data/vendor/datasketches-cpp/python/jupyter/ThetaSketchNotebook.ipynb +396 -0
- data/vendor/datasketches-cpp/python/src/__init__.py +2 -0
- data/vendor/datasketches-cpp/python/src/cpc_wrapper.cpp +90 -0
- data/vendor/datasketches-cpp/python/src/datasketches.cpp +40 -0
- data/vendor/datasketches-cpp/python/src/fi_wrapper.cpp +123 -0
- data/vendor/datasketches-cpp/python/src/hll_wrapper.cpp +136 -0
- data/vendor/datasketches-cpp/python/src/kll_wrapper.cpp +209 -0
- data/vendor/datasketches-cpp/python/src/theta_wrapper.cpp +162 -0
- data/vendor/datasketches-cpp/python/src/vector_of_kll.cpp +488 -0
- data/vendor/datasketches-cpp/python/src/vo_wrapper.cpp +140 -0
- data/vendor/datasketches-cpp/python/tests/__init__.py +0 -0
- data/vendor/datasketches-cpp/python/tests/cpc_test.py +64 -0
- data/vendor/datasketches-cpp/python/tests/fi_test.py +110 -0
- data/vendor/datasketches-cpp/python/tests/hll_test.py +131 -0
- data/vendor/datasketches-cpp/python/tests/kll_test.py +119 -0
- data/vendor/datasketches-cpp/python/tests/theta_test.py +121 -0
- data/vendor/datasketches-cpp/python/tests/vector_of_kll_test.py +148 -0
- data/vendor/datasketches-cpp/python/tests/vo_test.py +101 -0
- data/vendor/datasketches-cpp/sampling/CMakeLists.txt +48 -0
- data/vendor/datasketches-cpp/sampling/include/var_opt_sketch.hpp +392 -0
- data/vendor/datasketches-cpp/sampling/include/var_opt_sketch_impl.hpp +1752 -0
- data/vendor/datasketches-cpp/sampling/include/var_opt_union.hpp +239 -0
- data/vendor/datasketches-cpp/sampling/include/var_opt_union_impl.hpp +645 -0
- data/vendor/datasketches-cpp/sampling/test/CMakeLists.txt +43 -0
- data/vendor/datasketches-cpp/sampling/test/binaries_from_java.txt +67 -0
- data/vendor/datasketches-cpp/sampling/test/var_opt_sketch_test.cpp +509 -0
- data/vendor/datasketches-cpp/sampling/test/var_opt_union_test.cpp +358 -0
- data/vendor/datasketches-cpp/sampling/test/varopt_sketch_long_sampling.sk +0 -0
- data/vendor/datasketches-cpp/sampling/test/varopt_sketch_string_exact.sk +0 -0
- data/vendor/datasketches-cpp/sampling/test/varopt_union_double_sampling.sk +0 -0
- data/vendor/datasketches-cpp/setup.py +94 -0
- data/vendor/datasketches-cpp/theta/CMakeLists.txt +57 -0
- data/vendor/datasketches-cpp/theta/include/theta_a_not_b.hpp +73 -0
- data/vendor/datasketches-cpp/theta/include/theta_a_not_b_impl.hpp +83 -0
- data/vendor/datasketches-cpp/theta/include/theta_intersection.hpp +88 -0
- data/vendor/datasketches-cpp/theta/include/theta_intersection_impl.hpp +130 -0
- data/vendor/datasketches-cpp/theta/include/theta_sketch.hpp +533 -0
- data/vendor/datasketches-cpp/theta/include/theta_sketch_impl.hpp +939 -0
- data/vendor/datasketches-cpp/theta/include/theta_union.hpp +122 -0
- data/vendor/datasketches-cpp/theta/include/theta_union_impl.hpp +109 -0
- data/vendor/datasketches-cpp/theta/test/CMakeLists.txt +45 -0
- data/vendor/datasketches-cpp/theta/test/theta_a_not_b_test.cpp +244 -0
- data/vendor/datasketches-cpp/theta/test/theta_compact_empty_from_java.sk +0 -0
- data/vendor/datasketches-cpp/theta/test/theta_compact_estimation_from_java.sk +0 -0
- data/vendor/datasketches-cpp/theta/test/theta_compact_single_item_from_java.sk +0 -0
- data/vendor/datasketches-cpp/theta/test/theta_intersection_test.cpp +218 -0
- data/vendor/datasketches-cpp/theta/test/theta_sketch_test.cpp +438 -0
- data/vendor/datasketches-cpp/theta/test/theta_union_test.cpp +97 -0
- data/vendor/datasketches-cpp/theta/test/theta_update_empty_from_java.sk +0 -0
- data/vendor/datasketches-cpp/theta/test/theta_update_estimation_from_java.sk +0 -0
- data/vendor/datasketches-cpp/tuple/CMakeLists.txt +104 -0
- data/vendor/datasketches-cpp/tuple/include/array_of_doubles_a_not_b.hpp +52 -0
- data/vendor/datasketches-cpp/tuple/include/array_of_doubles_a_not_b_impl.hpp +32 -0
- data/vendor/datasketches-cpp/tuple/include/array_of_doubles_intersection.hpp +52 -0
- data/vendor/datasketches-cpp/tuple/include/array_of_doubles_intersection_impl.hpp +31 -0
- data/vendor/datasketches-cpp/tuple/include/array_of_doubles_sketch.hpp +179 -0
- data/vendor/datasketches-cpp/tuple/include/array_of_doubles_sketch_impl.hpp +238 -0
- data/vendor/datasketches-cpp/tuple/include/array_of_doubles_union.hpp +81 -0
- data/vendor/datasketches-cpp/tuple/include/array_of_doubles_union_impl.hpp +43 -0
- data/vendor/datasketches-cpp/tuple/include/bounds_on_ratios_in_sampled_sets.hpp +135 -0
- data/vendor/datasketches-cpp/tuple/include/bounds_on_ratios_in_theta_sketched_sets.hpp +135 -0
- data/vendor/datasketches-cpp/tuple/include/jaccard_similarity.hpp +172 -0
- data/vendor/datasketches-cpp/tuple/include/theta_a_not_b_experimental.hpp +53 -0
- data/vendor/datasketches-cpp/tuple/include/theta_a_not_b_experimental_impl.hpp +33 -0
- data/vendor/datasketches-cpp/tuple/include/theta_comparators.hpp +48 -0
- data/vendor/datasketches-cpp/tuple/include/theta_constants.hpp +34 -0
- data/vendor/datasketches-cpp/tuple/include/theta_helpers.hpp +54 -0
- data/vendor/datasketches-cpp/tuple/include/theta_intersection_base.hpp +59 -0
- data/vendor/datasketches-cpp/tuple/include/theta_intersection_base_impl.hpp +121 -0
- data/vendor/datasketches-cpp/tuple/include/theta_intersection_experimental.hpp +78 -0
- data/vendor/datasketches-cpp/tuple/include/theta_intersection_experimental_impl.hpp +43 -0
- data/vendor/datasketches-cpp/tuple/include/theta_set_difference_base.hpp +54 -0
- data/vendor/datasketches-cpp/tuple/include/theta_set_difference_base_impl.hpp +80 -0
- data/vendor/datasketches-cpp/tuple/include/theta_sketch_experimental.hpp +393 -0
- data/vendor/datasketches-cpp/tuple/include/theta_sketch_experimental_impl.hpp +481 -0
- data/vendor/datasketches-cpp/tuple/include/theta_union_base.hpp +60 -0
- data/vendor/datasketches-cpp/tuple/include/theta_union_base_impl.hpp +84 -0
- data/vendor/datasketches-cpp/tuple/include/theta_union_experimental.hpp +88 -0
- data/vendor/datasketches-cpp/tuple/include/theta_union_experimental_impl.hpp +47 -0
- data/vendor/datasketches-cpp/tuple/include/theta_update_sketch_base.hpp +259 -0
- data/vendor/datasketches-cpp/tuple/include/theta_update_sketch_base_impl.hpp +389 -0
- data/vendor/datasketches-cpp/tuple/include/tuple_a_not_b.hpp +57 -0
- data/vendor/datasketches-cpp/tuple/include/tuple_a_not_b_impl.hpp +33 -0
- data/vendor/datasketches-cpp/tuple/include/tuple_intersection.hpp +104 -0
- data/vendor/datasketches-cpp/tuple/include/tuple_intersection_impl.hpp +43 -0
- data/vendor/datasketches-cpp/tuple/include/tuple_sketch.hpp +496 -0
- data/vendor/datasketches-cpp/tuple/include/tuple_sketch_impl.hpp +587 -0
- data/vendor/datasketches-cpp/tuple/include/tuple_union.hpp +109 -0
- data/vendor/datasketches-cpp/tuple/include/tuple_union_impl.hpp +47 -0
- data/vendor/datasketches-cpp/tuple/test/CMakeLists.txt +53 -0
- data/vendor/datasketches-cpp/tuple/test/aod_1_compact_empty_from_java.sk +1 -0
- data/vendor/datasketches-cpp/tuple/test/aod_1_compact_estimation_from_java.sk +0 -0
- data/vendor/datasketches-cpp/tuple/test/aod_1_compact_non_empty_no_entries_from_java.sk +0 -0
- data/vendor/datasketches-cpp/tuple/test/aod_2_compact_exact_from_java.sk +0 -0
- data/vendor/datasketches-cpp/tuple/test/aod_3_compact_empty_from_java.sk +1 -0
- data/vendor/datasketches-cpp/tuple/test/array_of_doubles_sketch_test.cpp +298 -0
- data/vendor/datasketches-cpp/tuple/test/theta_a_not_b_experimental_test.cpp +250 -0
- data/vendor/datasketches-cpp/tuple/test/theta_compact_empty_from_java.sk +0 -0
- data/vendor/datasketches-cpp/tuple/test/theta_compact_estimation_from_java.sk +0 -0
- data/vendor/datasketches-cpp/tuple/test/theta_compact_single_item_from_java.sk +0 -0
- data/vendor/datasketches-cpp/tuple/test/theta_intersection_experimental_test.cpp +224 -0
- data/vendor/datasketches-cpp/tuple/test/theta_jaccard_similarity_test.cpp +144 -0
- data/vendor/datasketches-cpp/tuple/test/theta_sketch_experimental_test.cpp +247 -0
- data/vendor/datasketches-cpp/tuple/test/theta_union_experimental_test.cpp +44 -0
- data/vendor/datasketches-cpp/tuple/test/tuple_a_not_b_test.cpp +289 -0
- data/vendor/datasketches-cpp/tuple/test/tuple_intersection_test.cpp +235 -0
- data/vendor/datasketches-cpp/tuple/test/tuple_jaccard_similarity_test.cpp +98 -0
- data/vendor/datasketches-cpp/tuple/test/tuple_sketch_allocation_test.cpp +102 -0
- data/vendor/datasketches-cpp/tuple/test/tuple_sketch_test.cpp +249 -0
- data/vendor/datasketches-cpp/tuple/test/tuple_union_test.cpp +187 -0
- metadata +302 -0
@@ -0,0 +1,149 @@
|
|
1
|
+
/*
|
2
|
+
* Licensed to the Apache Software Foundation (ASF) under one
|
3
|
+
* or more contributor license agreements. See the NOTICE file
|
4
|
+
* distributed with this work for additional information
|
5
|
+
* regarding copyright ownership. The ASF licenses this file
|
6
|
+
* to you under the Apache License, Version 2.0 (the
|
7
|
+
* "License"); you may not use this file except in compliance
|
8
|
+
* with the License. You may obtain a copy of the License at
|
9
|
+
*
|
10
|
+
* http://www.apache.org/licenses/LICENSE-2.0
|
11
|
+
*
|
12
|
+
* Unless required by applicable law or agreed to in writing,
|
13
|
+
* software distributed under the License is distributed on an
|
14
|
+
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
15
|
+
* KIND, either express or implied. See the License for the
|
16
|
+
* specific language governing permissions and limitations
|
17
|
+
* under the License.
|
18
|
+
*/
|
19
|
+
|
20
|
+
#include <catch.hpp>
|
21
|
+
|
22
|
+
#include "cpc_union.hpp"
|
23
|
+
|
24
|
+
namespace datasketches {
|
25
|
+
|
26
|
+
static const double RELATIVE_ERROR_FOR_LG_K_11 = 0.02;
|
27
|
+
|
28
|
+
TEST_CASE("cpc union: lg k limits", "[cpc_union]") {
|
29
|
+
cpc_union u1(CPC_MIN_LG_K); // this should work
|
30
|
+
cpc_union u2(CPC_MAX_LG_K); // this should work
|
31
|
+
REQUIRE_THROWS_AS(cpc_union(CPC_MIN_LG_K - 1), std::invalid_argument);
|
32
|
+
REQUIRE_THROWS_AS(cpc_union(CPC_MAX_LG_K + 1), std::invalid_argument);
|
33
|
+
}
|
34
|
+
|
35
|
+
TEST_CASE("cpc union: empty", "[cpc_union]") {
|
36
|
+
cpc_union u(11);
|
37
|
+
auto s = u.get_result();
|
38
|
+
REQUIRE(s.is_empty());
|
39
|
+
REQUIRE(s.get_estimate() == 0.0);
|
40
|
+
}
|
41
|
+
|
42
|
+
TEST_CASE("cpc union: copy", "[cpc_union]") {
|
43
|
+
cpc_sketch s(11);
|
44
|
+
s.update(1);
|
45
|
+
cpc_union u1(11);
|
46
|
+
u1.update(s);
|
47
|
+
|
48
|
+
cpc_union u2 = u1; // copy constructor
|
49
|
+
auto s1 = u2.get_result();
|
50
|
+
REQUIRE_FALSE(s1.is_empty());
|
51
|
+
REQUIRE(s1.get_estimate() == Approx(1).margin(RELATIVE_ERROR_FOR_LG_K_11));
|
52
|
+
s.update(2);
|
53
|
+
u2.update(s);
|
54
|
+
u1 = u2; // operator=
|
55
|
+
auto s2 = u1.get_result();
|
56
|
+
REQUIRE_FALSE(s2.is_empty());
|
57
|
+
REQUIRE(s2.get_estimate() == Approx(2).margin(2 * RELATIVE_ERROR_FOR_LG_K_11));
|
58
|
+
}
|
59
|
+
|
60
|
+
TEST_CASE("cpc union: custom seed", "[cpc_union]") {
|
61
|
+
cpc_sketch s(11, 123);
|
62
|
+
|
63
|
+
s.update(1);
|
64
|
+
s.update(2);
|
65
|
+
s.update(3);
|
66
|
+
|
67
|
+
cpc_union u1(11, 123);
|
68
|
+
u1.update(s);
|
69
|
+
auto r = u1.get_result();
|
70
|
+
REQUIRE_FALSE(r.is_empty());
|
71
|
+
REQUIRE(r.get_estimate() == Approx(3).margin(3 * RELATIVE_ERROR_FOR_LG_K_11));
|
72
|
+
|
73
|
+
// incompatible seed
|
74
|
+
cpc_union u2(11, 234);
|
75
|
+
REQUIRE_THROWS_AS(u2.update(s), std::invalid_argument);
|
76
|
+
}
|
77
|
+
|
78
|
+
TEST_CASE("cpc union: large", "[cpc_union]") {
|
79
|
+
int key = 0;
|
80
|
+
cpc_sketch s(11);
|
81
|
+
cpc_union u(11);
|
82
|
+
for (int i = 0; i < 1000; i++) {
|
83
|
+
cpc_sketch tmp(11);
|
84
|
+
for (int i = 0; i < 10000; i++) {
|
85
|
+
s.update(key);
|
86
|
+
tmp.update(key);
|
87
|
+
key++;
|
88
|
+
}
|
89
|
+
u.update(tmp);
|
90
|
+
}
|
91
|
+
cpc_sketch r = u.get_result();
|
92
|
+
REQUIRE(r.get_num_coupons() == s.get_num_coupons());
|
93
|
+
REQUIRE(r.get_estimate() == Approx(s.get_estimate()).margin(s.get_estimate() * RELATIVE_ERROR_FOR_LG_K_11));
|
94
|
+
}
|
95
|
+
|
96
|
+
TEST_CASE("cpc union: reduce k empty", "[cpc_union]") {
|
97
|
+
cpc_sketch s(11);
|
98
|
+
for (int i = 0; i < 10000; i++) s.update(i);
|
99
|
+
cpc_union u(12);
|
100
|
+
u.update(s);
|
101
|
+
cpc_sketch r = u.get_result();
|
102
|
+
REQUIRE(r.get_lg_k() == 11);
|
103
|
+
REQUIRE(r.get_estimate() == Approx(10000).margin(10000 * RELATIVE_ERROR_FOR_LG_K_11));
|
104
|
+
}
|
105
|
+
|
106
|
+
TEST_CASE("cpc union: reduce k sparse", "[cpc_union]") {
|
107
|
+
cpc_union u(12);
|
108
|
+
|
109
|
+
cpc_sketch s12(12);
|
110
|
+
for (int i = 0; i < 100; i++) s12.update(i);
|
111
|
+
u.update(s12);
|
112
|
+
|
113
|
+
cpc_sketch s11(11);
|
114
|
+
for (int i = 0; i < 1000; i++) s11.update(i);
|
115
|
+
u.update(s11);
|
116
|
+
|
117
|
+
cpc_sketch r = u.get_result();
|
118
|
+
REQUIRE(r.get_lg_k() == 11);
|
119
|
+
REQUIRE(r.get_estimate() == Approx(1000).margin(1000 * RELATIVE_ERROR_FOR_LG_K_11));
|
120
|
+
}
|
121
|
+
|
122
|
+
TEST_CASE("cpc union: reduce k window", "[cpc_union]") {
|
123
|
+
cpc_union u(12);
|
124
|
+
|
125
|
+
cpc_sketch s12(12);
|
126
|
+
for (int i = 0; i < 500; i++) s12.update(i);
|
127
|
+
u.update(s12);
|
128
|
+
|
129
|
+
cpc_sketch s11(11);
|
130
|
+
for (int i = 0; i < 1000; i++) s11.update(i);
|
131
|
+
u.update(s11);
|
132
|
+
|
133
|
+
cpc_sketch r = u.get_result();
|
134
|
+
REQUIRE(r.get_lg_k() == 11);
|
135
|
+
REQUIRE(r.get_estimate() == Approx(1000).margin(1000 * RELATIVE_ERROR_FOR_LG_K_11));
|
136
|
+
}
|
137
|
+
|
138
|
+
TEST_CASE("cpc union: moving update", "[cpc_union]") {
|
139
|
+
cpc_union u(11);
|
140
|
+
cpc_sketch s(11);
|
141
|
+
for (int i = 0; i < 100; i++) s.update(i); // sparse
|
142
|
+
u.update(std::move(s));
|
143
|
+
cpc_sketch r = u.get_result();
|
144
|
+
REQUIRE(r.get_estimate() == Approx(100).margin(100 * RELATIVE_ERROR_FOR_LG_K_11));
|
145
|
+
}
|
146
|
+
|
147
|
+
|
148
|
+
|
149
|
+
} /* namespace datasketches */
|
@@ -0,0 +1,54 @@
|
|
1
|
+
# Licensed to the Apache Software Foundation (ASF) under one
|
2
|
+
# or more contributor license agreements. See the NOTICE file
|
3
|
+
# distributed with this work for additional information
|
4
|
+
# regarding copyright ownership. The ASF licenses this file
|
5
|
+
# to you under the Apache License, Version 2.0 (the
|
6
|
+
# "License"); you may not use this file except in compliance
|
7
|
+
# with the License. You may obtain a copy of the License at
|
8
|
+
#
|
9
|
+
# http://www.apache.org/licenses/LICENSE-2.0
|
10
|
+
#
|
11
|
+
# Unless required by applicable law or agreed to in writing,
|
12
|
+
# software distributed under the License is distributed on an
|
13
|
+
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
14
|
+
# KIND, either express or implied. See the License for the
|
15
|
+
# specific language governing permissions and limitations
|
16
|
+
# under the License.
|
17
|
+
|
18
|
+
add_library(fi INTERFACE)
|
19
|
+
|
20
|
+
add_library(${PROJECT_NAME}::FI ALIAS fi)
|
21
|
+
|
22
|
+
if (BUILD_TESTS)
|
23
|
+
add_subdirectory(test)
|
24
|
+
endif()
|
25
|
+
|
26
|
+
target_include_directories(fi
|
27
|
+
INTERFACE
|
28
|
+
$<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}/include>
|
29
|
+
$<INSTALL_INTERFACE:$<INSTALL_PREFIX>/include>
|
30
|
+
)
|
31
|
+
|
32
|
+
target_link_libraries(fi INTERFACE common)
|
33
|
+
target_compile_features(fi INTERFACE cxx_std_11)
|
34
|
+
|
35
|
+
set(fi_HEADERS "")
|
36
|
+
list(APPEND fi_HEADERS "include/frequent_items_sketch.hpp")
|
37
|
+
list(APPEND fi_HEADERS "include/frequent_items_sketch_impl.hpp")
|
38
|
+
list(APPEND fi_HEADERS "include/reverse_purge_hash_map.hpp")
|
39
|
+
list(APPEND fi_HEADERS "include/reverse_purge_hash_map_impl.hpp")
|
40
|
+
|
41
|
+
install(TARGETS fi
|
42
|
+
EXPORT ${PROJECT_NAME}
|
43
|
+
)
|
44
|
+
|
45
|
+
install(FILES ${fi_HEADERS}
|
46
|
+
DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}/DataSketches")
|
47
|
+
|
48
|
+
target_sources(fi
|
49
|
+
INTERFACE
|
50
|
+
${CMAKE_CURRENT_SOURCE_DIR}/include/frequent_items_sketch.hpp
|
51
|
+
${CMAKE_CURRENT_SOURCE_DIR}/include/frequent_items_sketch_impl.hpp
|
52
|
+
${CMAKE_CURRENT_SOURCE_DIR}/include/reverse_purge_hash_map.hpp
|
53
|
+
${CMAKE_CURRENT_SOURCE_DIR}/include/reverse_purge_hash_map_impl.hpp
|
54
|
+
)
|
@@ -0,0 +1,319 @@
|
|
1
|
+
/*
|
2
|
+
* Licensed to the Apache Software Foundation (ASF) under one
|
3
|
+
* or more contributor license agreements. See the NOTICE file
|
4
|
+
* distributed with this work for additional information
|
5
|
+
* regarding copyright ownership. The ASF licenses this file
|
6
|
+
* to you under the Apache License, Version 2.0 (the
|
7
|
+
* "License"); you may not use this file except in compliance
|
8
|
+
* with the License. You may obtain a copy of the License at
|
9
|
+
*
|
10
|
+
* http://www.apache.org/licenses/LICENSE-2.0
|
11
|
+
*
|
12
|
+
* Unless required by applicable law or agreed to in writing,
|
13
|
+
* software distributed under the License is distributed on an
|
14
|
+
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
15
|
+
* KIND, either express or implied. See the License for the
|
16
|
+
* specific language governing permissions and limitations
|
17
|
+
* under the License.
|
18
|
+
*/
|
19
|
+
|
20
|
+
#ifndef FREQUENT_ITEMS_SKETCH_HPP_
|
21
|
+
#define FREQUENT_ITEMS_SKETCH_HPP_
|
22
|
+
|
23
|
+
#include <memory>
|
24
|
+
#include <vector>
|
25
|
+
#include <iostream>
|
26
|
+
#include <functional>
|
27
|
+
#include <type_traits>
|
28
|
+
|
29
|
+
#include "reverse_purge_hash_map.hpp"
|
30
|
+
#include "common_defs.hpp"
|
31
|
+
#include "serde.hpp"
|
32
|
+
|
33
|
+
namespace datasketches {
|
34
|
+
|
35
|
+
/*
|
36
|
+
* Based on Java implementation here:
|
37
|
+
* https://github.com/DataSketches/sketches-core/blob/master/src/main/java/com/yahoo/sketches/frequencies/ItemsSketch.java
|
38
|
+
* author Alexander Saydakov
|
39
|
+
*/
|
40
|
+
|
41
|
+
enum frequent_items_error_type { NO_FALSE_POSITIVES, NO_FALSE_NEGATIVES };
|
42
|
+
|
43
|
+
// for serialization as raw bytes
|
44
|
+
template<typename A> using AllocU8 = typename std::allocator_traits<A>::template rebind_alloc<uint8_t>;
|
45
|
+
template<typename A> using vector_u8 = std::vector<uint8_t, AllocU8<A>>;
|
46
|
+
|
47
|
+
// type W for weight must be an arithmetic type (integral or floating point)
|
48
|
+
template<typename T, typename W = uint64_t, typename H = std::hash<T>, typename E = std::equal_to<T>, typename S = serde<T>, typename A = std::allocator<T>>
|
49
|
+
class frequent_items_sketch {
|
50
|
+
public:
|
51
|
+
|
52
|
+
/**
|
53
|
+
* Construct this sketch with parameters lg_max_map_size and lg_start_map_size.
|
54
|
+
*
|
55
|
+
* @param lg_max_map_size Log2 of the physical size of the internal hash map managed by this
|
56
|
+
* sketch. The maximum capacity of this internal hash map is 0.75 times 2^lg_max_map_size.
|
57
|
+
* Both the ultimate accuracy and size of this sketch are functions of lg_max_map_size.
|
58
|
+
*
|
59
|
+
* @param lg_start_map_size Log2 of the starting physical size of the internal hash
|
60
|
+
* map managed by this sketch.
|
61
|
+
*/
|
62
|
+
explicit frequent_items_sketch(uint8_t lg_max_map_size, uint8_t lg_start_map_size = LG_MIN_MAP_SIZE);
|
63
|
+
|
64
|
+
/**
|
65
|
+
* Update this sketch with an item and a positive weight (frequency count).
|
66
|
+
* @param item for which the weight should be increased (lvalue)
|
67
|
+
* @param weight the amount by which the weight of the item should be increased
|
68
|
+
* A count of zero is a no-op, and a negative count will throw an exception.
|
69
|
+
*/
|
70
|
+
void update(const T& item, W weight = 1);
|
71
|
+
|
72
|
+
/**
|
73
|
+
* Update this sketch with an item and a positive weight (frequency count).
|
74
|
+
* @param item for which the weight should be increased (rvalue)
|
75
|
+
* @param weight the amount by which the weight of the item should be increased
|
76
|
+
* A count of zero is a no-op, and a negative count will throw an exception.
|
77
|
+
*/
|
78
|
+
void update(T&& item, W weight = 1);
|
79
|
+
|
80
|
+
/**
|
81
|
+
* This function merges the other sketch into this one.
|
82
|
+
* The other sketch may be of a different size.
|
83
|
+
* @param other sketch to be merged into this (lvalue)
|
84
|
+
*/
|
85
|
+
void merge(const frequent_items_sketch& other);
|
86
|
+
|
87
|
+
/**
|
88
|
+
* This function merges the other sketch into this one.
|
89
|
+
* The other sketch may be of a different size.
|
90
|
+
* @param other sketch to be merged into this (rvalue)
|
91
|
+
*/
|
92
|
+
void merge(frequent_items_sketch&& other);
|
93
|
+
|
94
|
+
/**
|
95
|
+
* @return true if this sketch is empty
|
96
|
+
*/
|
97
|
+
bool is_empty() const;
|
98
|
+
|
99
|
+
/**
|
100
|
+
* @return the number of active items in the sketch
|
101
|
+
*/
|
102
|
+
uint32_t get_num_active_items() const;
|
103
|
+
|
104
|
+
/**
|
105
|
+
* Returns the sum of the weights (frequencies) in the stream seen so far by the sketch
|
106
|
+
*
|
107
|
+
* @return the total weight of all items in the stream seen so far by the sketch
|
108
|
+
*/
|
109
|
+
W get_total_weight() const;
|
110
|
+
|
111
|
+
/**
|
112
|
+
* Returns the estimate of the weight (frequency) of the given item.
|
113
|
+
* Note: The true frequency of a item would be the sum of the counts as a result of the
|
114
|
+
* two update functions.
|
115
|
+
*
|
116
|
+
* @param item the given item
|
117
|
+
* @return the estimate of the weight (frequency) of the given item
|
118
|
+
*/
|
119
|
+
W get_estimate(const T& item) const;
|
120
|
+
|
121
|
+
/**
|
122
|
+
* Returns the guaranteed lower bound weight (frequency) of the given item.
|
123
|
+
*
|
124
|
+
* @param item the given item.
|
125
|
+
* @return the guaranteed lower bound weight of the given item. That is, a number which
|
126
|
+
* is guaranteed to be no larger than the real weight.
|
127
|
+
*/
|
128
|
+
W get_lower_bound(const T& item) const;
|
129
|
+
|
130
|
+
/**
|
131
|
+
* Returns the guaranteed upper bound weight (frequency) of the given item.
|
132
|
+
*
|
133
|
+
* @param item the given item
|
134
|
+
* @return the guaranteed upper bound weight of the given item. That is, a number which
|
135
|
+
* is guaranteed to be no smaller than the real frequency.
|
136
|
+
*/
|
137
|
+
W get_upper_bound(const T& item) const;
|
138
|
+
|
139
|
+
/**
|
140
|
+
* @return An upper bound on the maximum error of get_estimate(item) for any item.
|
141
|
+
* This is equivalent to the maximum distance between the upper bound and the lower bound
|
142
|
+
* for any item.
|
143
|
+
*/
|
144
|
+
W get_maximum_error() const;
|
145
|
+
|
146
|
+
/**
|
147
|
+
* Returns epsilon value of this sketch.
|
148
|
+
* This is just the value <i>3.5 / max_map_size</i>.
|
149
|
+
* @return epsilon used by the sketch to compute error.
|
150
|
+
*/
|
151
|
+
double get_epsilon() const;
|
152
|
+
|
153
|
+
/**
|
154
|
+
* Returns epsilon used to compute <i>a priori</i> error.
|
155
|
+
* This is just the value <i>3.5 / maxMapSize</i>.
|
156
|
+
* @param maxMapSize the planned map size to be used when constructing this sketch.
|
157
|
+
* @return epsilon used to compute <i>a priori</i> error.
|
158
|
+
*/
|
159
|
+
static double get_epsilon(uint8_t lg_max_map_size);
|
160
|
+
|
161
|
+
/**
|
162
|
+
* Returns the estimated <i>a priori</i> error given the max_map_size for the sketch and the
|
163
|
+
* estimated_total_stream_weight.
|
164
|
+
* @param lg_max_map_size the planned map size to be used when constructing this sketch.
|
165
|
+
* @param estimated_total_stream_weight the estimated total stream weight.
|
166
|
+
* @return the estimated <i>a priori</i> error.
|
167
|
+
*/
|
168
|
+
static double get_apriori_error(uint8_t lg_max_map_size, W estimated_total_weight);
|
169
|
+
|
170
|
+
class row;
|
171
|
+
typedef typename std::vector<row, typename std::allocator_traits<A>::template rebind_alloc<row>> vector_row; // alias for users
|
172
|
+
|
173
|
+
/**
|
174
|
+
* Returns an array of rows that include frequent items, estimates, upper and lower bounds
|
175
|
+
* given an error_type and using get_maximum_error() as a threshold.
|
176
|
+
*
|
177
|
+
* <p>The method first examines all active items in the sketch (items that have a counter).
|
178
|
+
*
|
179
|
+
* <p>If <i>error_type = NO_FALSE_NEGATIVES</i>, this will include an item in the result
|
180
|
+
* list if get_upper_bound(item) > threshold.
|
181
|
+
* There will be no false negatives, i.e., no Type II error.
|
182
|
+
* There may be items in the set with true frequencies less than the threshold
|
183
|
+
* (false positives).</p>
|
184
|
+
*
|
185
|
+
* <p>If <i>error_type = NO_FALSE_POSITIVES</i>, this will include an item in the result
|
186
|
+
* list if get_lower_bound(item) > threshold.
|
187
|
+
* There will be no false positives, i.e., no Type I error.
|
188
|
+
* There may be items omitted from the set with true frequencies greater than the
|
189
|
+
* threshold (false negatives).</p>
|
190
|
+
*
|
191
|
+
* @param error_type determines whether no false positives or no false negatives are desired.
|
192
|
+
* @return an array of frequent items
|
193
|
+
*/
|
194
|
+
vector_row get_frequent_items(frequent_items_error_type err_type) const;
|
195
|
+
|
196
|
+
/**
|
197
|
+
* Returns an array of rows that include frequent items, estimates, upper and lower bounds
|
198
|
+
* given an error_type and a threshold.
|
199
|
+
*
|
200
|
+
* <p>The method first examines all active items in the sketch (items that have a counter).
|
201
|
+
*
|
202
|
+
* <p>If <i>error_type = NO_FALSE_NEGATIVES</i>, this will include an item in the result
|
203
|
+
* list if get_upper_bound(item) > threshold.
|
204
|
+
* There will be no false negatives, i.e., no Type II error.
|
205
|
+
* There may be items in the set with true frequencies less than the threshold
|
206
|
+
* (false positives).</p>
|
207
|
+
*
|
208
|
+
* <p>If <i>error_type = NO_FALSE_POSITIVES</i>, this will include an item in the result
|
209
|
+
* list if get_lower_bound(item) > threshold.
|
210
|
+
* There will be no false positives, i.e., no Type I error.
|
211
|
+
* There may be items omitted from the set with true frequencies greater than the
|
212
|
+
* threshold (false negatives).</p>
|
213
|
+
*
|
214
|
+
* @param error_type determines whether no false positives or no false negatives are desired.
|
215
|
+
* @param threshold to include items in the result list
|
216
|
+
* @return an array of frequent items
|
217
|
+
*/
|
218
|
+
vector_row get_frequent_items(frequent_items_error_type err_type, W threshold) const;
|
219
|
+
|
220
|
+
/**
|
221
|
+
* Computes size needed to serialize the current state of the sketch.
|
222
|
+
* This can be expensive since every item needs to be looked at.
|
223
|
+
* @return size in bytes needed to serialize this sketch
|
224
|
+
*/
|
225
|
+
size_t get_serialized_size_bytes() const;
|
226
|
+
|
227
|
+
/**
|
228
|
+
* This method serializes the sketch into a given stream in a binary form
|
229
|
+
* @param os output stream
|
230
|
+
*/
|
231
|
+
void serialize(std::ostream& os) const;
|
232
|
+
|
233
|
+
// This is a convenience alias for users
|
234
|
+
// The type returned by the following serialize method
|
235
|
+
typedef vector_u8<A> vector_bytes;
|
236
|
+
|
237
|
+
/**
|
238
|
+
* This method serializes the sketch as a vector of bytes.
|
239
|
+
* An optional header can be reserved in front of the sketch.
|
240
|
+
* It is a blank space of a given size.
|
241
|
+
* This header is used in Datasketches PostgreSQL extension.
|
242
|
+
* @param header_size_bytes space to reserve in front of the sketch
|
243
|
+
* @return serialized sketch as a vector of bytes
|
244
|
+
*/
|
245
|
+
vector_bytes serialize(unsigned header_size_bytes = 0) const;
|
246
|
+
|
247
|
+
/**
|
248
|
+
* This method deserializes a sketch from a given stream.
|
249
|
+
* @param is input stream
|
250
|
+
* @return an instance of the sketch
|
251
|
+
*/
|
252
|
+
static frequent_items_sketch deserialize(std::istream& is);
|
253
|
+
|
254
|
+
/**
|
255
|
+
* This method deserializes a sketch from a given array of bytes.
|
256
|
+
* @param bytes pointer to the array of bytes
|
257
|
+
* @param size the size of the array
|
258
|
+
* @return an instance of the sketch
|
259
|
+
*/
|
260
|
+
static frequent_items_sketch deserialize(const void* bytes, size_t size);
|
261
|
+
|
262
|
+
/**
|
263
|
+
* Returns a human readable summary of this sketch
|
264
|
+
* @param print_items if true include the list of items retained by the sketch
|
265
|
+
*/
|
266
|
+
string<A> to_string(bool print_items = false) const;
|
267
|
+
|
268
|
+
private:
|
269
|
+
static const uint8_t LG_MIN_MAP_SIZE = 3;
|
270
|
+
static const uint8_t SERIAL_VERSION = 1;
|
271
|
+
static const uint8_t FAMILY_ID = 10;
|
272
|
+
static const uint8_t PREAMBLE_LONGS_EMPTY = 1;
|
273
|
+
static const uint8_t PREAMBLE_LONGS_NONEMPTY = 4;
|
274
|
+
static constexpr double EPSILON_FACTOR = 3.5;
|
275
|
+
enum flags { IS_EMPTY };
|
276
|
+
W total_weight;
|
277
|
+
W offset;
|
278
|
+
reverse_purge_hash_map<T, W, H, E, A> map;
|
279
|
+
static void check_preamble_longs(uint8_t preamble_longs, bool is_empty);
|
280
|
+
static void check_serial_version(uint8_t serial_version);
|
281
|
+
static void check_family_id(uint8_t family_id);
|
282
|
+
static void check_size(uint8_t lg_cur_size, uint8_t lg_max_size);
|
283
|
+
|
284
|
+
// version for integral signed type
|
285
|
+
template<typename WW = W, typename std::enable_if<std::is_integral<WW>::value && std::is_signed<WW>::value, int>::type = 0>
|
286
|
+
static inline void check_weight(WW weight);
|
287
|
+
|
288
|
+
// version for integral unsigned type
|
289
|
+
template<typename WW = W, typename std::enable_if<std::is_integral<WW>::value && std::is_unsigned<WW>::value, int>::type = 0>
|
290
|
+
static inline void check_weight(WW weight);
|
291
|
+
|
292
|
+
// version for floating point type
|
293
|
+
template<typename WW = W, typename std::enable_if<std::is_floating_point<WW>::value, int>::type = 0>
|
294
|
+
static inline void check_weight(WW weight);
|
295
|
+
|
296
|
+
// for deserialize
|
297
|
+
class items_deleter;
|
298
|
+
};
|
299
|
+
|
300
|
+
template<typename T, typename W, typename H, typename E, typename S, typename A>
|
301
|
+
class frequent_items_sketch<T, W, H, E, S, A>::row {
|
302
|
+
public:
|
303
|
+
row(const T* item, W weight, W offset):
|
304
|
+
item(item), weight(weight), offset(offset) {}
|
305
|
+
const T& get_item() const { return *item; }
|
306
|
+
W get_estimate() const { return weight + offset; }
|
307
|
+
W get_lower_bound() const { return weight; }
|
308
|
+
W get_upper_bound() const { return weight + offset; }
|
309
|
+
private:
|
310
|
+
const T* item;
|
311
|
+
W weight;
|
312
|
+
W offset;
|
313
|
+
};
|
314
|
+
|
315
|
+
}
|
316
|
+
|
317
|
+
#include "frequent_items_sketch_impl.hpp"
|
318
|
+
|
319
|
+
# endif
|