datasketches 0.3.1 → 0.4.0

Sign up to get free protection for your applications and to get access to all the features.
Files changed (245) hide show
  1. checksums.yaml +4 -4
  2. data/CHANGELOG.md +10 -0
  3. data/NOTICE +1 -1
  4. data/README.md +0 -2
  5. data/ext/datasketches/cpc_wrapper.cpp +3 -3
  6. data/ext/datasketches/kll_wrapper.cpp +0 -10
  7. data/lib/datasketches/version.rb +1 -1
  8. data/lib/datasketches.rb +1 -1
  9. data/vendor/datasketches-cpp/CMakeLists.txt +23 -20
  10. data/vendor/datasketches-cpp/CODE_OF_CONDUCT.md +3 -0
  11. data/vendor/datasketches-cpp/CONTRIBUTING.md +50 -0
  12. data/vendor/datasketches-cpp/Doxyfile +2827 -0
  13. data/vendor/datasketches-cpp/LICENSE +0 -76
  14. data/vendor/datasketches-cpp/README.md +1 -3
  15. data/vendor/datasketches-cpp/common/CMakeLists.txt +12 -11
  16. data/vendor/datasketches-cpp/common/include/MurmurHash3.h +25 -27
  17. data/vendor/datasketches-cpp/common/include/common_defs.hpp +15 -10
  18. data/vendor/datasketches-cpp/common/include/count_zeros.hpp +11 -2
  19. data/vendor/datasketches-cpp/common/include/kolmogorov_smirnov.hpp +9 -6
  20. data/vendor/datasketches-cpp/common/include/memory_operations.hpp +5 -4
  21. data/vendor/datasketches-cpp/common/include/optional.hpp +148 -0
  22. data/vendor/datasketches-cpp/common/include/quantiles_sorted_view.hpp +95 -2
  23. data/vendor/datasketches-cpp/common/include/quantiles_sorted_view_impl.hpp +1 -1
  24. data/vendor/datasketches-cpp/common/include/serde.hpp +69 -20
  25. data/vendor/datasketches-cpp/common/test/CMakeLists.txt +2 -2
  26. data/vendor/datasketches-cpp/common/test/integration_test.cpp +6 -0
  27. data/vendor/datasketches-cpp/common/test/optional_test.cpp +85 -0
  28. data/vendor/datasketches-cpp/common/test/test_allocator.hpp +14 -14
  29. data/vendor/datasketches-cpp/{python/src/__init__.py → count/CMakeLists.txt} +25 -1
  30. data/vendor/datasketches-cpp/count/include/count_min.hpp +405 -0
  31. data/vendor/datasketches-cpp/count/include/count_min_impl.hpp +497 -0
  32. data/vendor/datasketches-cpp/{MANIFEST.in → count/test/CMakeLists.txt} +23 -20
  33. data/vendor/datasketches-cpp/count/test/count_min_allocation_test.cpp +155 -0
  34. data/vendor/datasketches-cpp/count/test/count_min_test.cpp +303 -0
  35. data/vendor/datasketches-cpp/cpc/include/cpc_common.hpp +14 -20
  36. data/vendor/datasketches-cpp/cpc/include/cpc_compressor.hpp +7 -4
  37. data/vendor/datasketches-cpp/cpc/include/cpc_compressor_impl.hpp +17 -17
  38. data/vendor/datasketches-cpp/cpc/include/cpc_confidence.hpp +3 -3
  39. data/vendor/datasketches-cpp/cpc/include/cpc_sketch.hpp +40 -40
  40. data/vendor/datasketches-cpp/cpc/include/cpc_sketch_impl.hpp +14 -11
  41. data/vendor/datasketches-cpp/cpc/include/cpc_union.hpp +35 -11
  42. data/vendor/datasketches-cpp/cpc/include/cpc_union_impl.hpp +8 -8
  43. data/vendor/datasketches-cpp/cpc/include/cpc_util.hpp +16 -8
  44. data/vendor/datasketches-cpp/cpc/include/u32_table.hpp +3 -2
  45. data/vendor/datasketches-cpp/cpc/include/u32_table_impl.hpp +5 -5
  46. data/vendor/datasketches-cpp/cpc/test/CMakeLists.txt +20 -7
  47. data/vendor/datasketches-cpp/cpc/test/cpc_sketch_deserialize_from_java_test.cpp +60 -0
  48. data/vendor/datasketches-cpp/cpc/test/cpc_sketch_serialize_for_java.cpp +38 -0
  49. data/vendor/datasketches-cpp/cpc/test/cpc_sketch_test.cpp +4 -29
  50. data/vendor/datasketches-cpp/cpc/test/cpc_union_test.cpp +4 -4
  51. data/vendor/datasketches-cpp/{tox.ini → density/CMakeLists.txt} +24 -8
  52. data/vendor/datasketches-cpp/density/include/density_sketch.hpp +256 -0
  53. data/vendor/datasketches-cpp/density/include/density_sketch_impl.hpp +543 -0
  54. data/vendor/datasketches-cpp/{python/datasketches/__init__.py → density/test/CMakeLists.txt} +15 -3
  55. data/vendor/datasketches-cpp/density/test/density_sketch_test.cpp +244 -0
  56. data/vendor/datasketches-cpp/fi/include/frequent_items_sketch.hpp +21 -9
  57. data/vendor/datasketches-cpp/fi/include/frequent_items_sketch_impl.hpp +6 -4
  58. data/vendor/datasketches-cpp/fi/include/reverse_purge_hash_map.hpp +9 -3
  59. data/vendor/datasketches-cpp/fi/test/CMakeLists.txt +14 -1
  60. data/vendor/datasketches-cpp/fi/test/frequent_items_sketch_deserialize_from_java_test.cpp +95 -0
  61. data/vendor/datasketches-cpp/fi/test/frequent_items_sketch_serialize_for_java.cpp +83 -0
  62. data/vendor/datasketches-cpp/fi/test/frequent_items_sketch_test.cpp +3 -42
  63. data/vendor/datasketches-cpp/hll/include/CouponList-internal.hpp +2 -2
  64. data/vendor/datasketches-cpp/hll/include/CouponList.hpp +3 -1
  65. data/vendor/datasketches-cpp/hll/include/Hll4Array-internal.hpp +19 -11
  66. data/vendor/datasketches-cpp/hll/include/Hll4Array.hpp +2 -5
  67. data/vendor/datasketches-cpp/hll/include/Hll6Array-internal.hpp +19 -7
  68. data/vendor/datasketches-cpp/hll/include/Hll6Array.hpp +1 -1
  69. data/vendor/datasketches-cpp/hll/include/Hll8Array-internal.hpp +98 -42
  70. data/vendor/datasketches-cpp/hll/include/Hll8Array.hpp +2 -0
  71. data/vendor/datasketches-cpp/hll/include/HllArray-internal.hpp +94 -61
  72. data/vendor/datasketches-cpp/hll/include/HllArray.hpp +20 -8
  73. data/vendor/datasketches-cpp/hll/include/HllSketch-internal.hpp +4 -4
  74. data/vendor/datasketches-cpp/hll/include/HllSketchImpl.hpp +3 -1
  75. data/vendor/datasketches-cpp/hll/include/HllSketchImplFactory.hpp +3 -21
  76. data/vendor/datasketches-cpp/hll/include/HllUnion-internal.hpp +8 -0
  77. data/vendor/datasketches-cpp/hll/include/HllUtil.hpp +14 -18
  78. data/vendor/datasketches-cpp/hll/include/coupon_iterator-internal.hpp +1 -1
  79. data/vendor/datasketches-cpp/hll/include/coupon_iterator.hpp +8 -2
  80. data/vendor/datasketches-cpp/hll/include/hll.hpp +79 -65
  81. data/vendor/datasketches-cpp/hll/test/CMakeLists.txt +14 -1
  82. data/vendor/datasketches-cpp/hll/test/HllUnionTest.cpp +7 -1
  83. data/vendor/datasketches-cpp/hll/test/ToFromByteArrayTest.cpp +0 -68
  84. data/vendor/datasketches-cpp/hll/test/hll_sketch_deserialize_from_java_test.cpp +69 -0
  85. data/vendor/datasketches-cpp/hll/test/hll_sketch_serialize_for_java.cpp +52 -0
  86. data/vendor/datasketches-cpp/kll/include/kll_helper.hpp +0 -1
  87. data/vendor/datasketches-cpp/kll/include/kll_helper_impl.hpp +2 -2
  88. data/vendor/datasketches-cpp/kll/include/kll_sketch.hpp +79 -53
  89. data/vendor/datasketches-cpp/kll/include/kll_sketch_impl.hpp +61 -132
  90. data/vendor/datasketches-cpp/kll/test/CMakeLists.txt +14 -1
  91. data/vendor/datasketches-cpp/kll/test/kll_sketch_deserialize_from_java_test.cpp +103 -0
  92. data/vendor/datasketches-cpp/kll/test/kll_sketch_serialize_for_java.cpp +62 -0
  93. data/vendor/datasketches-cpp/kll/test/kll_sketch_test.cpp +5 -40
  94. data/vendor/datasketches-cpp/quantiles/include/quantiles_sketch.hpp +76 -54
  95. data/vendor/datasketches-cpp/quantiles/include/quantiles_sketch_impl.hpp +66 -136
  96. data/vendor/datasketches-cpp/quantiles/test/CMakeLists.txt +14 -1
  97. data/vendor/datasketches-cpp/quantiles/test/quantiles_sketch_deserialize_from_java_test.cpp +84 -0
  98. data/vendor/datasketches-cpp/quantiles/test/quantiles_sketch_serialize_for_java.cpp +52 -0
  99. data/vendor/datasketches-cpp/quantiles/test/quantiles_sketch_test.cpp +15 -39
  100. data/vendor/datasketches-cpp/req/include/req_common.hpp +7 -3
  101. data/vendor/datasketches-cpp/req/include/req_compactor_impl.hpp +2 -4
  102. data/vendor/datasketches-cpp/req/include/req_sketch.hpp +105 -26
  103. data/vendor/datasketches-cpp/req/include/req_sketch_impl.hpp +50 -111
  104. data/vendor/datasketches-cpp/req/test/CMakeLists.txt +14 -1
  105. data/vendor/datasketches-cpp/req/test/req_sketch_deserialize_from_java_test.cpp +55 -0
  106. data/vendor/datasketches-cpp/{tuple/include/array_of_doubles_intersection_impl.hpp → req/test/req_sketch_serialize_for_java.cpp} +12 -7
  107. data/vendor/datasketches-cpp/req/test/req_sketch_test.cpp +3 -89
  108. data/vendor/datasketches-cpp/sampling/CMakeLists.txt +4 -0
  109. data/vendor/datasketches-cpp/sampling/include/ebpps_sample.hpp +210 -0
  110. data/vendor/datasketches-cpp/sampling/include/ebpps_sample_impl.hpp +535 -0
  111. data/vendor/datasketches-cpp/sampling/include/ebpps_sketch.hpp +281 -0
  112. data/vendor/datasketches-cpp/sampling/include/ebpps_sketch_impl.hpp +531 -0
  113. data/vendor/datasketches-cpp/sampling/include/var_opt_sketch.hpp +89 -32
  114. data/vendor/datasketches-cpp/sampling/include/var_opt_sketch_impl.hpp +33 -19
  115. data/vendor/datasketches-cpp/sampling/include/var_opt_union.hpp +13 -10
  116. data/vendor/datasketches-cpp/sampling/include/var_opt_union_impl.hpp +23 -19
  117. data/vendor/datasketches-cpp/sampling/test/CMakeLists.txt +55 -8
  118. data/vendor/datasketches-cpp/sampling/test/ebpps_allocation_test.cpp +96 -0
  119. data/vendor/datasketches-cpp/sampling/test/ebpps_sample_test.cpp +137 -0
  120. data/vendor/datasketches-cpp/sampling/test/ebpps_sketch_test.cpp +266 -0
  121. data/vendor/datasketches-cpp/sampling/test/var_opt_sketch_deserialize_from_java_test.cpp +81 -0
  122. data/vendor/datasketches-cpp/sampling/test/var_opt_sketch_serialize_for_java.cpp +54 -0
  123. data/vendor/datasketches-cpp/sampling/test/var_opt_sketch_test.cpp +33 -51
  124. data/vendor/datasketches-cpp/sampling/test/var_opt_union_deserialize_from_java_test.cpp +50 -0
  125. data/vendor/datasketches-cpp/sampling/test/var_opt_union_serialize_for_java.cpp +56 -0
  126. data/vendor/datasketches-cpp/sampling/test/var_opt_union_test.cpp +0 -20
  127. data/vendor/datasketches-cpp/theta/CMakeLists.txt +1 -0
  128. data/vendor/datasketches-cpp/theta/include/bit_packing.hpp +6279 -0
  129. data/vendor/datasketches-cpp/theta/include/bounds_on_ratios_in_sampled_sets.hpp +1 -0
  130. data/vendor/datasketches-cpp/theta/include/bounds_on_ratios_in_theta_sketched_sets.hpp +7 -6
  131. data/vendor/datasketches-cpp/theta/include/compact_theta_sketch_parser.hpp +14 -8
  132. data/vendor/datasketches-cpp/theta/include/compact_theta_sketch_parser_impl.hpp +60 -46
  133. data/vendor/datasketches-cpp/theta/include/theta_a_not_b.hpp +20 -5
  134. data/vendor/datasketches-cpp/theta/include/theta_constants.hpp +10 -4
  135. data/vendor/datasketches-cpp/theta/include/theta_helpers.hpp +3 -1
  136. data/vendor/datasketches-cpp/theta/include/theta_intersection.hpp +13 -5
  137. data/vendor/datasketches-cpp/theta/include/theta_intersection_base_impl.hpp +5 -5
  138. data/vendor/datasketches-cpp/theta/include/theta_intersection_impl.hpp +3 -3
  139. data/vendor/datasketches-cpp/theta/include/theta_jaccard_similarity.hpp +2 -1
  140. data/vendor/datasketches-cpp/theta/include/theta_jaccard_similarity_base.hpp +1 -0
  141. data/vendor/datasketches-cpp/theta/include/theta_set_difference_base_impl.hpp +1 -1
  142. data/vendor/datasketches-cpp/theta/include/theta_sketch.hpp +180 -33
  143. data/vendor/datasketches-cpp/theta/include/theta_sketch_impl.hpp +430 -130
  144. data/vendor/datasketches-cpp/theta/include/theta_union.hpp +17 -10
  145. data/vendor/datasketches-cpp/theta/include/theta_union_base_impl.hpp +10 -10
  146. data/vendor/datasketches-cpp/theta/include/theta_union_impl.hpp +3 -3
  147. data/vendor/datasketches-cpp/theta/include/theta_update_sketch_base.hpp +21 -6
  148. data/vendor/datasketches-cpp/theta/include/theta_update_sketch_base_impl.hpp +13 -3
  149. data/vendor/datasketches-cpp/theta/test/CMakeLists.txt +15 -1
  150. data/vendor/datasketches-cpp/theta/test/bit_packing_test.cpp +80 -0
  151. data/vendor/datasketches-cpp/theta/test/theta_sketch_deserialize_from_java_test.cpp +57 -0
  152. data/vendor/datasketches-cpp/theta/test/theta_sketch_serialize_for_java.cpp +61 -0
  153. data/vendor/datasketches-cpp/theta/test/theta_sketch_test.cpp +39 -188
  154. data/vendor/datasketches-cpp/theta/test/theta_union_test.cpp +25 -0
  155. data/vendor/datasketches-cpp/tuple/CMakeLists.txt +8 -7
  156. data/vendor/datasketches-cpp/tuple/include/array_of_doubles_sketch.hpp +19 -144
  157. data/vendor/datasketches-cpp/tuple/include/{array_of_doubles_a_not_b.hpp → array_tuple_a_not_b.hpp} +24 -16
  158. data/vendor/datasketches-cpp/tuple/include/{array_of_doubles_a_not_b_impl.hpp → array_tuple_a_not_b_impl.hpp} +4 -4
  159. data/vendor/datasketches-cpp/tuple/include/array_tuple_intersection.hpp +65 -0
  160. data/vendor/datasketches-cpp/tuple/include/array_tuple_intersection_impl.hpp +31 -0
  161. data/vendor/datasketches-cpp/tuple/include/array_tuple_sketch.hpp +237 -0
  162. data/vendor/datasketches-cpp/tuple/include/{array_of_doubles_sketch_impl.hpp → array_tuple_sketch_impl.hpp} +40 -41
  163. data/vendor/datasketches-cpp/tuple/include/array_tuple_union.hpp +81 -0
  164. data/vendor/datasketches-cpp/tuple/include/array_tuple_union_impl.hpp +43 -0
  165. data/vendor/datasketches-cpp/tuple/include/tuple_a_not_b.hpp +11 -2
  166. data/vendor/datasketches-cpp/tuple/include/tuple_intersection.hpp +17 -10
  167. data/vendor/datasketches-cpp/tuple/include/tuple_jaccard_similarity.hpp +2 -1
  168. data/vendor/datasketches-cpp/tuple/include/tuple_sketch.hpp +95 -32
  169. data/vendor/datasketches-cpp/tuple/include/tuple_sketch_impl.hpp +2 -1
  170. data/vendor/datasketches-cpp/tuple/include/tuple_union.hpp +19 -11
  171. data/vendor/datasketches-cpp/tuple/test/CMakeLists.txt +16 -1
  172. data/vendor/datasketches-cpp/tuple/test/aod_sketch_deserialize_from_java_test.cpp +76 -0
  173. data/vendor/datasketches-cpp/tuple/test/aod_sketch_serialize_for_java.cpp +62 -0
  174. data/vendor/datasketches-cpp/tuple/test/array_of_doubles_sketch_test.cpp +5 -129
  175. data/vendor/datasketches-cpp/tuple/test/engagement_test.cpp +85 -89
  176. data/vendor/datasketches-cpp/tuple/test/tuple_jaccard_similarity_test.cpp +3 -1
  177. data/vendor/datasketches-cpp/tuple/test/tuple_sketch_deserialize_from_java_test.cpp +47 -0
  178. data/vendor/datasketches-cpp/{python/src/datasketches.cpp → tuple/test/tuple_sketch_serialize_for_java.cpp} +16 -30
  179. data/vendor/datasketches-cpp/tuple/test/tuple_sketch_test.cpp +1 -1
  180. data/vendor/datasketches-cpp/version.cfg.in +1 -1
  181. metadata +61 -79
  182. data/vendor/datasketches-cpp/fi/test/items_sketch_string_from_java.sk +0 -0
  183. data/vendor/datasketches-cpp/fi/test/items_sketch_string_utf8_from_java.sk +0 -0
  184. data/vendor/datasketches-cpp/fi/test/longs_sketch_from_java.sk +0 -0
  185. data/vendor/datasketches-cpp/hll/test/array6_from_java.sk +0 -0
  186. data/vendor/datasketches-cpp/hll/test/compact_array4_from_java.sk +0 -0
  187. data/vendor/datasketches-cpp/hll/test/compact_set_from_java.sk +0 -0
  188. data/vendor/datasketches-cpp/hll/test/list_from_java.sk +0 -0
  189. data/vendor/datasketches-cpp/hll/test/updatable_array4_from_java.sk +0 -0
  190. data/vendor/datasketches-cpp/hll/test/updatable_set_from_java.sk +0 -0
  191. data/vendor/datasketches-cpp/kll/test/kll_sketch_from_java.sk +0 -0
  192. data/vendor/datasketches-cpp/pyproject.toml +0 -23
  193. data/vendor/datasketches-cpp/python/CMakeLists.txt +0 -81
  194. data/vendor/datasketches-cpp/python/README.md +0 -85
  195. data/vendor/datasketches-cpp/python/datasketches/PySerDe.py +0 -104
  196. data/vendor/datasketches-cpp/python/include/py_serde.hpp +0 -113
  197. data/vendor/datasketches-cpp/python/jupyter/CPCSketch.ipynb +0 -345
  198. data/vendor/datasketches-cpp/python/jupyter/FrequentItemsSketch.ipynb +0 -354
  199. data/vendor/datasketches-cpp/python/jupyter/HLLSketch.ipynb +0 -346
  200. data/vendor/datasketches-cpp/python/jupyter/KLLSketch.ipynb +0 -463
  201. data/vendor/datasketches-cpp/python/jupyter/ThetaSketchNotebook.ipynb +0 -403
  202. data/vendor/datasketches-cpp/python/pybind11Path.cmd +0 -21
  203. data/vendor/datasketches-cpp/python/src/cpc_wrapper.cpp +0 -90
  204. data/vendor/datasketches-cpp/python/src/fi_wrapper.cpp +0 -128
  205. data/vendor/datasketches-cpp/python/src/hll_wrapper.cpp +0 -134
  206. data/vendor/datasketches-cpp/python/src/kll_wrapper.cpp +0 -210
  207. data/vendor/datasketches-cpp/python/src/ks_wrapper.cpp +0 -68
  208. data/vendor/datasketches-cpp/python/src/py_serde.cpp +0 -111
  209. data/vendor/datasketches-cpp/python/src/quantiles_wrapper.cpp +0 -204
  210. data/vendor/datasketches-cpp/python/src/req_wrapper.cpp +0 -215
  211. data/vendor/datasketches-cpp/python/src/theta_wrapper.cpp +0 -172
  212. data/vendor/datasketches-cpp/python/src/vector_of_kll.cpp +0 -490
  213. data/vendor/datasketches-cpp/python/src/vo_wrapper.cpp +0 -173
  214. data/vendor/datasketches-cpp/python/tests/__init__.py +0 -16
  215. data/vendor/datasketches-cpp/python/tests/cpc_test.py +0 -64
  216. data/vendor/datasketches-cpp/python/tests/fi_test.py +0 -110
  217. data/vendor/datasketches-cpp/python/tests/hll_test.py +0 -130
  218. data/vendor/datasketches-cpp/python/tests/kll_test.py +0 -125
  219. data/vendor/datasketches-cpp/python/tests/quantiles_test.py +0 -126
  220. data/vendor/datasketches-cpp/python/tests/req_test.py +0 -126
  221. data/vendor/datasketches-cpp/python/tests/theta_test.py +0 -146
  222. data/vendor/datasketches-cpp/python/tests/vector_of_kll_test.py +0 -148
  223. data/vendor/datasketches-cpp/python/tests/vo_test.py +0 -125
  224. data/vendor/datasketches-cpp/req/test/req_float_empty_from_java.sk +0 -0
  225. data/vendor/datasketches-cpp/req/test/req_float_estimation_from_java.sk +0 -0
  226. data/vendor/datasketches-cpp/req/test/req_float_exact_from_java.sk +0 -0
  227. data/vendor/datasketches-cpp/req/test/req_float_raw_items_from_java.sk +0 -0
  228. data/vendor/datasketches-cpp/req/test/req_float_single_item_from_java.sk +0 -0
  229. data/vendor/datasketches-cpp/sampling/test/binaries_from_java.txt +0 -67
  230. data/vendor/datasketches-cpp/sampling/test/varopt_sketch_long_sampling.sk +0 -0
  231. data/vendor/datasketches-cpp/sampling/test/varopt_sketch_string_exact.sk +0 -0
  232. data/vendor/datasketches-cpp/sampling/test/varopt_union_double_sampling.sk +0 -0
  233. data/vendor/datasketches-cpp/setup.py +0 -110
  234. data/vendor/datasketches-cpp/theta/test/theta_compact_empty_from_java.sk +0 -0
  235. data/vendor/datasketches-cpp/theta/test/theta_compact_estimation_from_java.sk +0 -0
  236. data/vendor/datasketches-cpp/theta/test/theta_compact_exact_from_java.sk +0 -0
  237. data/vendor/datasketches-cpp/theta/test/theta_compact_single_item_from_java.sk +0 -0
  238. data/vendor/datasketches-cpp/tuple/include/array_of_doubles_intersection.hpp +0 -52
  239. data/vendor/datasketches-cpp/tuple/include/array_of_doubles_union.hpp +0 -81
  240. data/vendor/datasketches-cpp/tuple/include/array_of_doubles_union_impl.hpp +0 -43
  241. data/vendor/datasketches-cpp/tuple/test/aod_1_compact_empty_from_java.sk +0 -1
  242. data/vendor/datasketches-cpp/tuple/test/aod_1_compact_estimation_from_java.sk +0 -0
  243. data/vendor/datasketches-cpp/tuple/test/aod_1_compact_non_empty_no_entries_from_java.sk +0 -0
  244. data/vendor/datasketches-cpp/tuple/test/aod_2_compact_exact_from_java.sk +0 -0
  245. data/vendor/datasketches-cpp/tuple/test/aod_3_compact_empty_from_java.sk +0 -1
@@ -32,10 +32,10 @@ namespace datasketches {
32
32
  static const double RELATIVE_ERROR_FOR_LG_K_11 = 0.02;
33
33
 
34
34
  TEST_CASE("cpc sketch: lg k limits", "[cpc_sketch]") {
35
- cpc_sketch s1(CPC_MIN_LG_K); // this should work
36
- cpc_sketch s2(CPC_MAX_LG_K); // this should work
37
- REQUIRE_THROWS_AS(cpc_sketch(CPC_MIN_LG_K - 1), std::invalid_argument);
38
- REQUIRE_THROWS_AS(cpc_sketch(CPC_MAX_LG_K + 1), std::invalid_argument);
35
+ cpc_sketch s1(cpc_constants::MIN_LG_K); // this should work
36
+ cpc_sketch s2(cpc_constants::MAX_LG_K); // this should work
37
+ REQUIRE_THROWS_AS(cpc_sketch(cpc_constants::MIN_LG_K - 1), std::invalid_argument);
38
+ REQUIRE_THROWS_AS(cpc_sketch(cpc_constants::MAX_LG_K + 1), std::invalid_argument);
39
39
  }
40
40
 
41
41
  TEST_CASE("cpc sketch: empty", "[cpc_sketch]") {
@@ -88,9 +88,6 @@ TEST_CASE("cpc sketch: serialize deserialize empty", "[cpc_sketch]") {
88
88
  REQUIRE(deserialized.is_empty() == sketch.is_empty());
89
89
  REQUIRE(deserialized.get_estimate() == sketch.get_estimate());
90
90
  REQUIRE(deserialized.validate());
91
-
92
- std::ofstream os("cpc-empty.bin");
93
- sketch.serialize(os);
94
91
  }
95
92
 
96
93
  TEST_CASE("cpc sketch: serialize deserialize sparse", "[cpc_sketch]") {
@@ -108,9 +105,6 @@ TEST_CASE("cpc sketch: serialize deserialize sparse", "[cpc_sketch]") {
108
105
  for (int i = 0; i < n; i++) deserialized.update(i);
109
106
  REQUIRE(deserialized.get_estimate() == sketch.get_estimate());
110
107
  REQUIRE(deserialized.validate());
111
-
112
- std::ofstream os("cpc-sparse.bin");
113
- sketch.serialize(os);
114
108
  }
115
109
 
116
110
  TEST_CASE("cpc sketch: serialize deserialize hybrid", "[cpc_sketch]") {
@@ -128,9 +122,6 @@ TEST_CASE("cpc sketch: serialize deserialize hybrid", "[cpc_sketch]") {
128
122
  for (int i = 0; i < n; i++) deserialized.update(i);
129
123
  REQUIRE(deserialized.get_estimate() == sketch.get_estimate());
130
124
  REQUIRE(deserialized.validate());
131
-
132
- std::ofstream os("cpc-hybrid.bin");
133
- sketch.serialize(os);
134
125
  }
135
126
 
136
127
  TEST_CASE("cpc sketch: serialize deserialize pinned", "[cpc_sketch]") {
@@ -148,9 +139,6 @@ TEST_CASE("cpc sketch: serialize deserialize pinned", "[cpc_sketch]") {
148
139
  for (int i = 0; i < n; i++) deserialized.update(i);
149
140
  REQUIRE(deserialized.get_estimate() == sketch.get_estimate());
150
141
  REQUIRE(deserialized.validate());
151
-
152
- std::ofstream os("cpc-pinned.bin");
153
- sketch.serialize(os);
154
142
  }
155
143
 
156
144
  TEST_CASE("cpc sketch: serialize deserialize sliding", "[cpc_sketch]") {
@@ -168,9 +156,6 @@ TEST_CASE("cpc sketch: serialize deserialize sliding", "[cpc_sketch]") {
168
156
  for (int i = 0; i < n; i++) deserialized.update(i);
169
157
  REQUIRE(deserialized.get_estimate() == sketch.get_estimate());
170
158
  REQUIRE(deserialized.validate());
171
-
172
- std::ofstream os("cpc-sliding.bin");
173
- sketch.serialize(os);
174
159
  }
175
160
 
176
161
  TEST_CASE("cpc sketch: serializing deserialize sliding large", "[cpc_sketch]") {
@@ -188,9 +173,6 @@ TEST_CASE("cpc sketch: serializing deserialize sliding large", "[cpc_sketch]") {
188
173
  for (int i = 0; i < n; i++) deserialized.update(i);
189
174
  REQUIRE(deserialized.get_estimate() == sketch.get_estimate());
190
175
  REQUIRE(deserialized.validate());
191
-
192
- std::ofstream os("cpc-sliding-large.bin");
193
- sketch.serialize(os);
194
176
  }
195
177
 
196
178
  TEST_CASE("cpc sketch: serialize deserialize empty, bytes", "[cpc_sketch]") {
@@ -201,9 +183,6 @@ TEST_CASE("cpc sketch: serialize deserialize empty, bytes", "[cpc_sketch]") {
201
183
  REQUIRE(deserialized.get_estimate() == sketch.get_estimate());
202
184
  REQUIRE(deserialized.validate());
203
185
  REQUIRE_THROWS_AS(cpc_sketch::deserialize(bytes.data(), bytes.size() - 1), std::out_of_range);
204
-
205
- std::ofstream os("cpc-empty.bin");
206
- sketch.serialize(os);
207
186
  }
208
187
 
209
188
  TEST_CASE("cpc sketch: serialize deserialize sparse, bytes", "[cpc_sketch]") {
@@ -261,8 +240,6 @@ TEST_CASE("cpc sketch: serialize deserialize pinned, bytes", "[cpc_sketch]") {
261
240
  for (int i = 0; i < n; i++) deserialized.update(i);
262
241
  REQUIRE(deserialized.get_estimate() == sketch.get_estimate());
263
242
  REQUIRE(deserialized.validate());
264
-
265
- std::cout << sketch.to_string();
266
243
  }
267
244
 
268
245
  TEST_CASE("cpc sketch: serialize deserialize sliding, bytes", "[cpc_sketch]") {
@@ -380,8 +357,6 @@ TEST_CASE("cpc sketch: update int equivalence", "[cpc_sketch]") {
380
357
  sketch.update((uint8_t) -1);
381
358
  sketch.update((int8_t) -1);
382
359
  REQUIRE(sketch.get_estimate() == Approx(1).margin(RELATIVE_ERROR_FOR_LG_K_11));
383
- std::ofstream os("cpc-negative-one.bin"); // to compare with Java
384
- sketch.serialize(os);
385
360
  }
386
361
 
387
362
  TEST_CASE("cpc sketch: update float equivalence", "[cpc_sketch]") {
@@ -28,10 +28,10 @@ namespace datasketches {
28
28
  static const double RELATIVE_ERROR_FOR_LG_K_11 = 0.02;
29
29
 
30
30
  TEST_CASE("cpc union: lg k limits", "[cpc_union]") {
31
- cpc_union u1(CPC_MIN_LG_K); // this should work
32
- cpc_union u2(CPC_MAX_LG_K); // this should work
33
- REQUIRE_THROWS_AS(cpc_union(CPC_MIN_LG_K - 1), std::invalid_argument);
34
- REQUIRE_THROWS_AS(cpc_union(CPC_MAX_LG_K + 1), std::invalid_argument);
31
+ cpc_union u1(cpc_constants::MIN_LG_K); // this should work
32
+ cpc_union u2(cpc_constants::MAX_LG_K); // this should work
33
+ REQUIRE_THROWS_AS(cpc_union(cpc_constants::MIN_LG_K - 1), std::invalid_argument);
34
+ REQUIRE_THROWS_AS(cpc_union(cpc_constants::MAX_LG_K + 1), std::invalid_argument);
35
35
  }
36
36
 
37
37
  TEST_CASE("cpc union: empty", "[cpc_union]") {
@@ -15,12 +15,28 @@
15
15
  # specific language governing permissions and limitations
16
16
  # under the License.
17
17
 
18
- [tox]
19
- envlist = py3
20
- isolated_build = true
18
+ add_library(density INTERFACE)
21
19
 
22
- [testenv]
23
- deps = pytest
24
- numpy
25
- changedir = python/tests
26
- commands = pytest
20
+ add_library(${PROJECT_NAME}::DENSITY ALIAS density)
21
+
22
+ if (BUILD_TESTS)
23
+ add_subdirectory(test)
24
+ endif()
25
+
26
+ target_include_directories(density
27
+ INTERFACE
28
+ $<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}/include>
29
+ $<INSTALL_INTERFACE:$<INSTALL_PREFIX>/include>
30
+ )
31
+
32
+ target_link_libraries(density INTERFACE common)
33
+ target_compile_features(density INTERFACE cxx_std_11)
34
+
35
+ install(TARGETS density
36
+ EXPORT ${PROJECT_NAME}
37
+ )
38
+
39
+ install(FILES
40
+ include/density_sketch.hpp
41
+ include/density_sketch_impl.hpp
42
+ DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}/DataSketches")
@@ -0,0 +1,256 @@
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 DENSITY_SKETCH_HPP_
21
+ #define DENSITY_SKETCH_HPP_
22
+
23
+ #include <type_traits>
24
+ #include <vector>
25
+ #include <functional>
26
+ #include <numeric>
27
+ #include <cmath>
28
+
29
+ #include "common_defs.hpp"
30
+
31
+ namespace datasketches {
32
+
33
+ template<typename T>
34
+ struct gaussian_kernel {
35
+ T operator()(const std::vector<T>& v1, const std::vector<T>& v2) const {
36
+ return exp(-std::inner_product(v1.begin(), v1.end(), v2.begin(), 0.0, std::plus<T>(), [](T a, T b){return (a-b)*(a-b);}));
37
+ }
38
+ };
39
+
40
+ /**
41
+ * Density sketch.
42
+ *
43
+ * Builds a coreset from the given set of input points. Provides density estimate at a given point.
44
+ *
45
+ * Based on the following paper:
46
+ * Zohar Karnin, Edo Liberty "Discrepancy, Coresets, and Sketches in Machine Learning"
47
+ * https://proceedings.mlr.press/v99/karnin19a/karnin19a.pdf
48
+ *
49
+ * Inspired by the following implementation:
50
+ * https://github.com/edoliberty/streaming-quantiles/blob/f688c8161a25582457b0a09deb4630a81406293b/gde.py
51
+ */
52
+ template<
53
+ typename T,
54
+ typename Kernel = gaussian_kernel<T>,
55
+ typename Allocator = std::allocator<T>
56
+ >
57
+ class density_sketch {
58
+ static_assert(std::is_floating_point<T>::value, "Floating point type expected");
59
+
60
+ public:
61
+ using Vector = std::vector<T, Allocator>;
62
+ using Level = std::vector<Vector, typename std::allocator_traits<Allocator>::template rebind_alloc<Vector>>;
63
+ using Levels = std::vector<Level, typename std::allocator_traits<Allocator>::template rebind_alloc<Level>>;
64
+
65
+ /**
66
+ * Constructor
67
+ * @param k controls the size and error of the sketch.
68
+ * @param dim dimension of the input domain
69
+ * @param kernel to use by this instance
70
+ * @param allocator to use by this instance
71
+ */
72
+ density_sketch(uint16_t k, uint32_t dim, const Kernel& kernel = Kernel(), const Allocator& allocator = Allocator());
73
+
74
+ /**
75
+ * Returns configured parameter K
76
+ * @return parameter K
77
+ */
78
+ uint16_t get_k() const;
79
+
80
+ /**
81
+ * Returns configured dimensions
82
+ * @return dimensions
83
+ */
84
+ uint32_t get_dim() const;
85
+
86
+ /**
87
+ * Returns true if this sketch is empty.
88
+ * @return empty flag
89
+ */
90
+ bool is_empty() const;
91
+
92
+ /**
93
+ * Returns the length of the input stream (number of points observed by this sketch).
94
+ * @return stream length
95
+ */
96
+ uint64_t get_n() const;
97
+
98
+ /**
99
+ * Returns the number of retained points in the sketch.
100
+ * @return number of retained points
101
+ */
102
+ uint32_t get_num_retained() const;
103
+
104
+ /**
105
+ * Returns true if this sketch is in estimation mode.
106
+ * @return estimation mode flag
107
+ */
108
+ bool is_estimation_mode() const;
109
+
110
+ /**
111
+ * Updates this sketch with a given point.
112
+ * @param point given point
113
+ */
114
+ template<typename FwdVector>
115
+ void update(FwdVector&& point);
116
+
117
+ /**
118
+ * Merges another sketch into this one.
119
+ * @param other sketch to merge into this one
120
+ */
121
+ template<typename FwdSketch>
122
+ void merge(FwdSketch&& other);
123
+
124
+ /**
125
+ * Density estimate at a given point
126
+ * @return density estimate at a given point
127
+ */
128
+ T get_estimate(const std::vector<T>& point) const;
129
+
130
+ /**
131
+ * Returns an instance of the allocator for this sketch.
132
+ * @return allocator
133
+ */
134
+ Allocator get_allocator() const;
135
+
136
+ /**
137
+ * This method serializes the sketch into a given stream in a binary form
138
+ * @param os output stream
139
+ */
140
+ void serialize(std::ostream& os) const;
141
+
142
+ using vector_bytes = std::vector<uint8_t, typename std::allocator_traits<Allocator>::template rebind_alloc<uint8_t>>;
143
+
144
+ /**
145
+ * This method serializes the sketch as a vector of bytes.
146
+ * An optional header can be reserved in front of the sketch.
147
+ * It is an uninitialized space of a given size.
148
+ * This header is used in Datasketches PostgreSQL extension.
149
+ * @param header_size_bytes space to reserve in front of the sketch
150
+ */
151
+ vector_bytes serialize(unsigned header_size_bytes = 0) const;
152
+
153
+ /**
154
+ * This method deserializes a sketch from a given stream.
155
+ * @param is input stream
156
+ * @param kernel the kernel function to use for this sketch
157
+ * @param allocator the memory allocator to use with this sketch
158
+ * @return an instance of the sketch
159
+ */
160
+ static density_sketch deserialize(std::istream& is,
161
+ const Kernel& kernel=Kernel(), const Allocator& allocator = Allocator());
162
+
163
+ /**
164
+ * This method deserializes a sketch from a given array of bytes.
165
+ * @param bytes pointer to the array of bytes
166
+ * @param size the size of the array
167
+ * @param kernel the kernel function to use for this sketch
168
+ * @param allocator the memory allocator to use with this sketch
169
+ * @return an instance of the sketch
170
+ */
171
+ static density_sketch deserialize(const void* bytes, size_t size,
172
+ const Kernel& kernel=Kernel(), const Allocator& allocator = Allocator());
173
+
174
+ /**
175
+ * Prints a summary of the sketch.
176
+ * @param print_levels if true include information about levels
177
+ * @param print_items if true include sketch data
178
+ */
179
+ string<Allocator> to_string(bool print_levels = false, bool print_items = false) const;
180
+
181
+ class const_iterator;
182
+
183
+ /**
184
+ * Iterator pointing to the first item in the sketch.
185
+ * If the sketch is empty, the returned iterator must not be dereferenced or incremented.
186
+ * @return iterator pointing to the first item in the sketch
187
+ */
188
+ const_iterator begin() const;
189
+
190
+ /**
191
+ * Iterator pointing to the past-the-end item in the sketch.
192
+ * The past-the-end item is the hypothetical item that would follow the last item.
193
+ * It does not point to any item, and must not be dereferenced or incremented.
194
+ * @return iterator pointing to the past-the-end item in the sketch
195
+ */
196
+ const_iterator end() const;
197
+
198
+ private:
199
+ enum flags { RESERVED0, RESERVED1, IS_EMPTY };
200
+ static const uint8_t PREAMBLE_INTS_SHORT = 3;
201
+ static const uint8_t PREAMBLE_INTS_LONG = 6;
202
+ static const uint8_t FAMILY_ID = 19;
203
+ static const uint8_t SERIAL_VERSION = 1;
204
+ static const size_t LEVELS_ARRAY_START = 5;
205
+
206
+ Allocator allocator_;
207
+ Kernel kernel_;
208
+ uint16_t k_;
209
+ uint32_t dim_;
210
+ uint32_t num_retained_;
211
+ uint64_t n_;
212
+ Levels levels_;
213
+
214
+ void compact();
215
+ void compact_level(unsigned height);
216
+
217
+ static void check_k(uint16_t k);
218
+ static void check_serial_version(uint8_t serial_version);
219
+ static void check_family_id(uint8_t family_id);
220
+ static void check_header_validity(uint8_t preamble_ints, uint8_t flags_byte, uint8_t serial_version);
221
+
222
+ density_sketch(uint16_t k, uint32_t dim, uint32_t num_retained, uint64_t n, Levels&& levels,
223
+ const Kernel& kernel = Kernel());
224
+ };
225
+
226
+ template<typename T, typename K, typename A>
227
+ class density_sketch<T, K, A>::const_iterator {
228
+ public:
229
+ using Vector = density_sketch<T, K, A>::Vector;
230
+ using iterator_category = std::input_iterator_tag;
231
+ using value_type = std::pair<const Vector&, const uint64_t>;
232
+ using difference_type = void;
233
+ using pointer = return_value_holder<value_type>;
234
+ using reference = const value_type;
235
+ const_iterator& operator++();
236
+ const_iterator& operator++(int);
237
+ bool operator==(const const_iterator& other) const;
238
+ bool operator!=(const const_iterator& other) const;
239
+ const value_type operator*() const;
240
+ const return_value_holder<value_type> operator->() const;
241
+ private:
242
+ using LevelsIterator = typename density_sketch<T, K, A>::Levels::const_iterator;
243
+ using LevelIterator = typename density_sketch<T, K, A>::Level::const_iterator;
244
+ LevelsIterator levels_it_;
245
+ LevelsIterator levels_end_;
246
+ LevelIterator level_it_;
247
+ unsigned height_;
248
+ friend class density_sketch<T, K, A>;
249
+ const_iterator(LevelsIterator begin, LevelsIterator end);
250
+ };
251
+
252
+ } /* namespace datasketches */
253
+
254
+ #include "density_sketch_impl.hpp"
255
+
256
+ #endif