datasketches 0.2.3 → 0.2.4
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGELOG.md +4 -0
- data/README.md +7 -7
- data/ext/datasketches/theta_wrapper.cpp +20 -4
- data/lib/datasketches/version.rb +1 -1
- data/vendor/datasketches-cpp/CMakeLists.txt +22 -3
- data/vendor/datasketches-cpp/MANIFEST.in +3 -0
- data/vendor/datasketches-cpp/README.md +76 -9
- data/vendor/datasketches-cpp/cmake/DataSketchesConfig.cmake.in +10 -0
- data/vendor/datasketches-cpp/common/CMakeLists.txt +14 -13
- data/vendor/datasketches-cpp/cpc/CMakeLists.txt +15 -35
- data/vendor/datasketches-cpp/cpc/include/cpc_sketch_impl.hpp +5 -3
- data/vendor/datasketches-cpp/cpc/include/cpc_union_impl.hpp +8 -6
- data/vendor/datasketches-cpp/cpc/test/cpc_sketch_allocation_test.cpp +17 -0
- data/vendor/datasketches-cpp/fi/CMakeLists.txt +5 -15
- data/vendor/datasketches-cpp/fi/include/frequent_items_sketch_impl.hpp +4 -2
- data/vendor/datasketches-cpp/hll/CMakeLists.txt +33 -56
- data/vendor/datasketches-cpp/hll/include/HllSketch-internal.hpp +6 -4
- data/vendor/datasketches-cpp/kll/CMakeLists.txt +9 -19
- data/vendor/datasketches-cpp/kll/include/kll_sketch_impl.hpp +4 -2
- data/vendor/datasketches-cpp/req/CMakeLists.txt +8 -21
- data/vendor/datasketches-cpp/req/include/req_sketch_impl.hpp +4 -2
- data/vendor/datasketches-cpp/sampling/CMakeLists.txt +5 -9
- data/vendor/datasketches-cpp/sampling/include/var_opt_sketch_impl.hpp +13 -7
- data/vendor/datasketches-cpp/sampling/include/var_opt_union_impl.hpp +8 -6
- data/vendor/datasketches-cpp/setup.py +1 -1
- data/vendor/datasketches-cpp/theta/CMakeLists.txt +26 -45
- data/vendor/datasketches-cpp/theta/include/compact_theta_sketch_parser_impl.hpp +89 -22
- data/vendor/datasketches-cpp/theta/include/theta_helpers.hpp +15 -0
- data/vendor/datasketches-cpp/theta/include/theta_intersection_base_impl.hpp +6 -6
- data/vendor/datasketches-cpp/theta/include/theta_set_difference_base_impl.hpp +2 -2
- data/vendor/datasketches-cpp/theta/include/theta_sketch.hpp +32 -15
- data/vendor/datasketches-cpp/theta/include/theta_sketch_impl.hpp +146 -51
- data/vendor/datasketches-cpp/theta/include/theta_union.hpp +6 -1
- data/vendor/datasketches-cpp/theta/include/theta_union_base.hpp +3 -1
- data/vendor/datasketches-cpp/theta/include/theta_union_base_impl.hpp +8 -2
- data/vendor/datasketches-cpp/theta/include/theta_union_impl.hpp +8 -5
- data/vendor/datasketches-cpp/theta/include/theta_update_sketch_base.hpp +5 -4
- data/vendor/datasketches-cpp/theta/include/theta_update_sketch_base_impl.hpp +33 -9
- data/vendor/datasketches-cpp/theta/test/CMakeLists.txt +1 -0
- data/vendor/datasketches-cpp/theta/test/theta_compact_empty_from_java_v1.sk +0 -0
- data/vendor/datasketches-cpp/theta/test/theta_compact_empty_from_java_v2.sk +0 -0
- data/vendor/datasketches-cpp/theta/test/theta_compact_estimation_from_java_v1.sk +0 -0
- data/vendor/datasketches-cpp/theta/test/theta_compact_estimation_from_java_v2.sk +0 -0
- data/vendor/datasketches-cpp/theta/test/theta_compact_exact_from_java.sk +0 -0
- data/vendor/datasketches-cpp/theta/test/theta_setop_test.cpp +445 -0
- data/vendor/datasketches-cpp/theta/test/theta_sketch_test.cpp +400 -0
- data/vendor/datasketches-cpp/theta/test/theta_union_test.cpp +23 -11
- data/vendor/datasketches-cpp/tuple/CMakeLists.txt +18 -33
- data/vendor/datasketches-cpp/tuple/include/array_of_doubles_sketch.hpp +1 -1
- data/vendor/datasketches-cpp/tuple/include/array_of_doubles_sketch_impl.hpp +3 -3
- data/vendor/datasketches-cpp/tuple/include/array_of_doubles_union.hpp +1 -1
- data/vendor/datasketches-cpp/tuple/include/array_of_doubles_union_impl.hpp +3 -3
- data/vendor/datasketches-cpp/tuple/include/tuple_sketch.hpp +13 -9
- data/vendor/datasketches-cpp/tuple/include/tuple_sketch_impl.hpp +33 -14
- data/vendor/datasketches-cpp/tuple/include/tuple_union.hpp +6 -1
- data/vendor/datasketches-cpp/tuple/include/tuple_union_impl.hpp +8 -3
- data/vendor/datasketches-cpp/tuple/test/array_of_doubles_sketch_test.cpp +16 -0
- data/vendor/datasketches-cpp/tuple/test/tuple_sketch_test.cpp +46 -8
- data/vendor/datasketches-cpp/tuple/test/tuple_union_test.cpp +7 -0
- metadata +11 -6
- 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
@@ -20,8 +20,8 @@
|
|
20
20
|
namespace datasketches {
|
21
21
|
|
22
22
|
template<typename A>
|
23
|
-
array_of_doubles_union_alloc<A>::array_of_doubles_union_alloc(uint8_t lg_cur_size, uint8_t lg_nom_size, resize_factor rf, uint64_t theta, uint64_t seed, const Policy& policy, const A& allocator):
|
24
|
-
Base(lg_cur_size, lg_nom_size, rf, theta, seed, policy, allocator)
|
23
|
+
array_of_doubles_union_alloc<A>::array_of_doubles_union_alloc(uint8_t lg_cur_size, uint8_t lg_nom_size, resize_factor rf, float p, uint64_t theta, uint64_t seed, const Policy& policy, const A& allocator):
|
24
|
+
Base(lg_cur_size, lg_nom_size, rf, p, theta, seed, policy, allocator)
|
25
25
|
{}
|
26
26
|
|
27
27
|
template<typename A>
|
@@ -37,7 +37,7 @@ tuple_base_builder<builder, Policy, A>(policy, allocator) {}
|
|
37
37
|
|
38
38
|
template<typename A>
|
39
39
|
array_of_doubles_union_alloc<A> array_of_doubles_union_alloc<A>::builder::build() const {
|
40
|
-
return array_of_doubles_union_alloc<A>(this->starting_lg_size(), this->lg_k_, this->rf_, this->starting_theta(), this->seed_, this->policy_, this->allocator_);
|
40
|
+
return array_of_doubles_union_alloc<A>(this->starting_lg_size(), this->lg_k_, this->rf_, this->p_, this->starting_theta(), this->seed_, this->policy_, this->allocator_);
|
41
41
|
}
|
42
42
|
|
43
43
|
} /* namespace datasketches */
|
@@ -153,8 +153,7 @@ public:
|
|
153
153
|
virtual const_iterator end() const = 0;
|
154
154
|
|
155
155
|
protected:
|
156
|
-
|
157
|
-
virtual void print_specifics(ostrstream& os) const = 0;
|
156
|
+
virtual void print_specifics(std::ostringstream& os) const = 0;
|
158
157
|
|
159
158
|
static uint16_t get_seed_hash(uint64_t seed);
|
160
159
|
|
@@ -325,6 +324,11 @@ public:
|
|
325
324
|
*/
|
326
325
|
void trim();
|
327
326
|
|
327
|
+
/**
|
328
|
+
* Reset the sketch to the initial empty state
|
329
|
+
*/
|
330
|
+
void reset();
|
331
|
+
|
328
332
|
/**
|
329
333
|
* Converts this sketch to a compact sketch (ordered or unordered).
|
330
334
|
* @param ordered optional flag to specify if ordered sketch should be produced
|
@@ -342,10 +346,9 @@ protected:
|
|
342
346
|
tuple_map map_;
|
343
347
|
|
344
348
|
// for builder
|
345
|
-
update_tuple_sketch(uint8_t lg_cur_size, uint8_t lg_nom_size, resize_factor rf, uint64_t theta, uint64_t seed, const Policy& policy, const Allocator& allocator);
|
349
|
+
update_tuple_sketch(uint8_t lg_cur_size, uint8_t lg_nom_size, resize_factor rf, float p, uint64_t theta, uint64_t seed, const Policy& policy, const Allocator& allocator);
|
346
350
|
|
347
|
-
|
348
|
-
virtual void print_specifics(ostrstream& os) const;
|
351
|
+
virtual void print_specifics(std::ostringstream& os) const;
|
349
352
|
};
|
350
353
|
|
351
354
|
// compact sketch
|
@@ -367,9 +370,11 @@ public:
|
|
367
370
|
using vector_bytes = std::vector<uint8_t, AllocBytes>;
|
368
371
|
using comparator = compare_by_key<ExtractKey>;
|
369
372
|
|
370
|
-
static const uint8_t
|
373
|
+
static const uint8_t SERIAL_VERSION_LEGACY = 1;
|
374
|
+
static const uint8_t SERIAL_VERSION = 3;
|
371
375
|
static const uint8_t SKETCH_FAMILY = 9;
|
372
|
-
static const uint8_t SKETCH_TYPE =
|
376
|
+
static const uint8_t SKETCH_TYPE = 1;
|
377
|
+
static const uint8_t SKETCH_TYPE_LEGACY = 5;
|
373
378
|
enum flags { IS_BIG_ENDIAN, IS_READ_ONLY, IS_EMPTY, IS_COMPACT, IS_ORDERED };
|
374
379
|
|
375
380
|
// Instances of this type can be obtained:
|
@@ -473,8 +478,7 @@ protected:
|
|
473
478
|
bool destroy_;
|
474
479
|
};
|
475
480
|
|
476
|
-
|
477
|
-
virtual void print_specifics(ostrstream& os) const;
|
481
|
+
virtual void print_specifics(std::ostringstream& os) const;
|
478
482
|
|
479
483
|
};
|
480
484
|
|
@@ -53,7 +53,9 @@ double tuple_sketch<S, A>::get_upper_bound(uint8_t num_std_devs) const {
|
|
53
53
|
|
54
54
|
template<typename S, typename A>
|
55
55
|
string<A> tuple_sketch<S, A>::to_string(bool detail) const {
|
56
|
-
|
56
|
+
// Using a temporary stream for implementation here does not comply with AllocatorAwareContainer requirements.
|
57
|
+
// The stream does not support passing an allocator instance, and alternatives are complicated.
|
58
|
+
std::ostringstream os;
|
57
59
|
os << "### Tuple sketch summary:" << std::endl;
|
58
60
|
os << " num retained entries : " << get_num_retained() << std::endl;
|
59
61
|
os << " seed hash : " << get_seed_hash() << std::endl;
|
@@ -74,15 +76,15 @@ string<A> tuple_sketch<S, A>::to_string(bool detail) const {
|
|
74
76
|
}
|
75
77
|
os << "### End retained entries" << std::endl;
|
76
78
|
}
|
77
|
-
return os.str();
|
79
|
+
return string<A>(os.str().c_str(), get_allocator());
|
78
80
|
}
|
79
81
|
|
80
82
|
// update sketch
|
81
83
|
|
82
84
|
template<typename S, typename U, typename P, typename A>
|
83
|
-
update_tuple_sketch<S, U, P, A>::update_tuple_sketch(uint8_t lg_cur_size, uint8_t lg_nom_size, resize_factor rf, uint64_t theta, uint64_t seed, const P& policy, const A& allocator):
|
85
|
+
update_tuple_sketch<S, U, P, A>::update_tuple_sketch(uint8_t lg_cur_size, uint8_t lg_nom_size, resize_factor rf, float p, uint64_t theta, uint64_t seed, const P& policy, const A& allocator):
|
84
86
|
policy_(policy),
|
85
|
-
map_(lg_cur_size, lg_nom_size, rf, theta, seed, allocator)
|
87
|
+
map_(lg_cur_size, lg_nom_size, rf, p, theta, seed, allocator)
|
86
88
|
{}
|
87
89
|
|
88
90
|
template<typename S, typename U, typename P, typename A>
|
@@ -97,12 +99,12 @@ bool update_tuple_sketch<S, U, P, A>::is_empty() const {
|
|
97
99
|
|
98
100
|
template<typename S, typename U, typename P, typename A>
|
99
101
|
bool update_tuple_sketch<S, U, P, A>::is_ordered() const {
|
100
|
-
return false
|
102
|
+
return map_.num_entries_ > 1 ? false : true;;
|
101
103
|
}
|
102
104
|
|
103
105
|
template<typename S, typename U, typename P, typename A>
|
104
106
|
uint64_t update_tuple_sketch<S, U, P, A>::get_theta64() const {
|
105
|
-
return map_.theta_;
|
107
|
+
return is_empty() ? theta_constants::MAX_THETA : map_.theta_;
|
106
108
|
}
|
107
109
|
|
108
110
|
template<typename S, typename U, typename P, typename A>
|
@@ -212,6 +214,11 @@ void update_tuple_sketch<S, U, P, A>::trim() {
|
|
212
214
|
map_.trim();
|
213
215
|
}
|
214
216
|
|
217
|
+
template<typename S, typename U, typename P, typename A>
|
218
|
+
void update_tuple_sketch<S, U, P, A>::reset() {
|
219
|
+
map_.reset();
|
220
|
+
}
|
221
|
+
|
215
222
|
template<typename S, typename U, typename P, typename A>
|
216
223
|
auto update_tuple_sketch<S, U, P, A>::begin() -> iterator {
|
217
224
|
return iterator(map_.entries_, 1 << map_.lg_cur_size_, 0);
|
@@ -238,7 +245,7 @@ compact_tuple_sketch<S, A> update_tuple_sketch<S, U, P, A>::compact(bool ordered
|
|
238
245
|
}
|
239
246
|
|
240
247
|
template<typename S, typename U, typename P, typename A>
|
241
|
-
void update_tuple_sketch<S, U, P, A>::print_specifics(
|
248
|
+
void update_tuple_sketch<S, U, P, A>::print_specifics(std::ostringstream& os) const {
|
242
249
|
os << " lg nominal size : " << (int) map_.lg_nom_size_ << std::endl;
|
243
250
|
os << " lg current size : " << (int) map_.lg_cur_size_ << std::endl;
|
244
251
|
os << " resize factor : " << (1 << map_.rf_) << std::endl;
|
@@ -250,7 +257,7 @@ template<typename S, typename A>
|
|
250
257
|
compact_tuple_sketch<S, A>::compact_tuple_sketch(bool is_empty, bool is_ordered, uint16_t seed_hash, uint64_t theta,
|
251
258
|
std::vector<Entry, AllocEntry>&& entries):
|
252
259
|
is_empty_(is_empty),
|
253
|
-
is_ordered_(is_ordered),
|
260
|
+
is_ordered_(is_ordered || (entries.size() <= 1ULL)),
|
254
261
|
seed_hash_(seed_hash),
|
255
262
|
theta_(theta),
|
256
263
|
entries_(std::move(entries))
|
@@ -437,9 +444,15 @@ compact_tuple_sketch<S, A> compact_tuple_sketch<S, A>::deserialize(std::istream&
|
|
437
444
|
read<uint8_t>(is); // unused
|
438
445
|
const auto flags_byte = read<uint8_t>(is);
|
439
446
|
const auto seed_hash = read<uint16_t>(is);
|
440
|
-
|
447
|
+
if (serial_version != SERIAL_VERSION && serial_version != SERIAL_VERSION_LEGACY) {
|
448
|
+
throw std::invalid_argument("serial version mismatch: expected " + std::to_string(SERIAL_VERSION) + " or "
|
449
|
+
+ std::to_string(SERIAL_VERSION_LEGACY) + ", actual " + std::to_string(serial_version));
|
450
|
+
}
|
441
451
|
checker<true>::check_sketch_family(family, SKETCH_FAMILY);
|
442
|
-
|
452
|
+
if (type != SKETCH_TYPE && type != SKETCH_TYPE_LEGACY) {
|
453
|
+
throw std::invalid_argument("sketch type mismatch: expected " + std::to_string(SKETCH_TYPE) + " or "
|
454
|
+
+ std::to_string(SKETCH_TYPE_LEGACY) + ", actual " + std::to_string(type));
|
455
|
+
}
|
443
456
|
const bool is_empty = flags_byte & (1 << flags::IS_EMPTY);
|
444
457
|
if (!is_empty) checker<true>::check_seed_hash(seed_hash, compute_seed_hash(seed));
|
445
458
|
|
@@ -492,9 +505,15 @@ compact_tuple_sketch<S, A> compact_tuple_sketch<S, A>::deserialize(const void* b
|
|
492
505
|
ptr += copy_from_mem(ptr, flags_byte);
|
493
506
|
uint16_t seed_hash;
|
494
507
|
ptr += copy_from_mem(ptr, seed_hash);
|
495
|
-
|
508
|
+
if (serial_version != SERIAL_VERSION && serial_version != SERIAL_VERSION_LEGACY) {
|
509
|
+
throw std::invalid_argument("serial version mismatch: expected " + std::to_string(SERIAL_VERSION) + " or "
|
510
|
+
+ std::to_string(SERIAL_VERSION_LEGACY) + ", actual " + std::to_string(serial_version));
|
511
|
+
}
|
496
512
|
checker<true>::check_sketch_family(family, SKETCH_FAMILY);
|
497
|
-
|
513
|
+
if (type != SKETCH_TYPE && type != SKETCH_TYPE_LEGACY) {
|
514
|
+
throw std::invalid_argument("sketch type mismatch: expected " + std::to_string(SKETCH_TYPE) + " or "
|
515
|
+
+ std::to_string(SKETCH_TYPE_LEGACY) + ", actual " + std::to_string(type));
|
516
|
+
}
|
498
517
|
const bool is_empty = flags_byte & (1 << flags::IS_EMPTY);
|
499
518
|
if (!is_empty) checker<true>::check_seed_hash(seed_hash, compute_seed_hash(seed));
|
500
519
|
|
@@ -554,7 +573,7 @@ auto compact_tuple_sketch<S, A>::end() const -> const_iterator {
|
|
554
573
|
}
|
555
574
|
|
556
575
|
template<typename S, typename A>
|
557
|
-
void compact_tuple_sketch<S, A>::print_specifics(
|
576
|
+
void compact_tuple_sketch<S, A>::print_specifics(std::ostringstream&) const {}
|
558
577
|
|
559
578
|
// builder
|
560
579
|
|
@@ -568,7 +587,7 @@ tuple_base_builder<builder, P, A>(policy, allocator) {}
|
|
568
587
|
|
569
588
|
template<typename S, typename U, typename P, typename A>
|
570
589
|
auto update_tuple_sketch<S, U, P, A>::builder::build() const -> update_tuple_sketch {
|
571
|
-
return update_tuple_sketch(this->starting_lg_size(), this->lg_k_, this->rf_, this->starting_theta(), this->seed_, this->policy_, this->allocator_);
|
590
|
+
return update_tuple_sketch(this->starting_lg_size(), this->lg_k_, this->rf_, this->p_, this->starting_theta(), this->seed_, this->policy_, this->allocator_);
|
572
591
|
}
|
573
592
|
|
574
593
|
} /* namespace datasketches */
|
@@ -80,11 +80,16 @@ public:
|
|
80
80
|
*/
|
81
81
|
CompactSketch get_result(bool ordered = true) const;
|
82
82
|
|
83
|
+
/**
|
84
|
+
* Reset the union to the initial empty state
|
85
|
+
*/
|
86
|
+
void reset();
|
87
|
+
|
83
88
|
protected:
|
84
89
|
State state_;
|
85
90
|
|
86
91
|
// for builder
|
87
|
-
tuple_union(uint8_t lg_cur_size, uint8_t lg_nom_size, resize_factor rf, uint64_t theta, uint64_t seed, const Policy& policy, const Allocator& allocator);
|
92
|
+
tuple_union(uint8_t lg_cur_size, uint8_t lg_nom_size, resize_factor rf, float p, uint64_t theta, uint64_t seed, const Policy& policy, const Allocator& allocator);
|
88
93
|
};
|
89
94
|
|
90
95
|
template<typename S, typename P, typename A>
|
@@ -20,8 +20,8 @@
|
|
20
20
|
namespace datasketches {
|
21
21
|
|
22
22
|
template<typename S, typename P, typename A>
|
23
|
-
tuple_union<S, P, A>::tuple_union(uint8_t lg_cur_size, uint8_t lg_nom_size, resize_factor rf, uint64_t theta, uint64_t seed, const P& policy, const A& allocator):
|
24
|
-
state_(lg_cur_size, lg_nom_size, rf, theta, seed, internal_policy(policy), allocator)
|
23
|
+
tuple_union<S, P, A>::tuple_union(uint8_t lg_cur_size, uint8_t lg_nom_size, resize_factor rf, float p, uint64_t theta, uint64_t seed, const P& policy, const A& allocator):
|
24
|
+
state_(lg_cur_size, lg_nom_size, rf, p, theta, seed, internal_policy(policy), allocator)
|
25
25
|
{}
|
26
26
|
|
27
27
|
template<typename S, typename P, typename A>
|
@@ -35,13 +35,18 @@ auto tuple_union<S, P, A>::get_result(bool ordered) const -> CompactSketch {
|
|
35
35
|
return state_.get_result(ordered);
|
36
36
|
}
|
37
37
|
|
38
|
+
template<typename S, typename P, typename A>
|
39
|
+
void tuple_union<S, P, A>::reset() {
|
40
|
+
return state_.reset();
|
41
|
+
}
|
42
|
+
|
38
43
|
template<typename S, typename P, typename A>
|
39
44
|
tuple_union<S, P, A>::builder::builder(const P& policy, const A& allocator):
|
40
45
|
tuple_base_builder<builder, P, A>(policy, allocator) {}
|
41
46
|
|
42
47
|
template<typename S, typename P, typename A>
|
43
48
|
auto tuple_union<S, P, A>::builder::build() const -> tuple_union {
|
44
|
-
return tuple_union(this->starting_lg_size(), this->lg_k_, this->rf_, this->starting_theta(), this->seed_, this->policy_, this->allocator_);
|
49
|
+
return tuple_union(this->starting_lg_size(), this->lg_k_, this->rf_, this->p_, this->starting_theta(), this->seed_, this->policy_, this->allocator_);
|
45
50
|
}
|
46
51
|
|
47
52
|
} /* namespace datasketches */
|
@@ -36,6 +36,17 @@ const std::string inputPath = TEST_BINARY_INPUT_PATH;
|
|
36
36
|
const std::string inputPath = "test/";
|
37
37
|
#endif
|
38
38
|
|
39
|
+
TEST_CASE("aod sketch: reset", "[tuple_sketch]") {
|
40
|
+
auto update_sketch = update_array_of_doubles_sketch::builder().build();
|
41
|
+
std::vector<double> a = {1};
|
42
|
+
update_sketch.update(1, a);
|
43
|
+
REQUIRE(!update_sketch.is_empty());
|
44
|
+
REQUIRE(update_sketch.get_num_retained() == 1);
|
45
|
+
update_sketch.reset();
|
46
|
+
REQUIRE(update_sketch.is_empty());
|
47
|
+
REQUIRE(update_sketch.get_num_retained() == 0);
|
48
|
+
}
|
49
|
+
|
39
50
|
TEST_CASE("aod sketch: serialization compatibility with java - empty", "[tuple_sketch]") {
|
40
51
|
auto update_sketch = update_array_of_doubles_sketch::builder().build();
|
41
52
|
REQUIRE(update_sketch.is_empty());
|
@@ -263,6 +274,11 @@ TEST_CASE("aod union: half overlap", "[tuple_sketch]") {
|
|
263
274
|
u.update(update_sketch2);
|
264
275
|
auto result = u.get_result();
|
265
276
|
REQUIRE(result.get_estimate() == Approx(1500).margin(0.01));
|
277
|
+
|
278
|
+
u.reset();
|
279
|
+
result = u.get_result();
|
280
|
+
REQUIRE(result.is_empty());
|
281
|
+
REQUIRE(result.get_num_retained() == 0);
|
266
282
|
}
|
267
283
|
|
268
284
|
TEST_CASE("aod intersection: half overlap", "[tuple_sketch]") {
|
@@ -34,7 +34,6 @@ std::ostream& operator<<(std::ostream& os, const three_doubles& tuple) {
|
|
34
34
|
|
35
35
|
#include <catch.hpp>
|
36
36
|
#include <tuple_sketch.hpp>
|
37
|
-
//#include <test_type.hpp>
|
38
37
|
|
39
38
|
namespace datasketches {
|
40
39
|
|
@@ -43,9 +42,11 @@ TEST_CASE("tuple sketch float: builder", "[tuple_sketch]") {
|
|
43
42
|
builder.set_lg_k(10).set_p(0.5f).set_resize_factor(theta_constants::resize_factor::X2).set_seed(123);
|
44
43
|
auto sketch = builder.build();
|
45
44
|
REQUIRE(sketch.get_lg_k() == 10);
|
46
|
-
REQUIRE(sketch.get_theta() == 0
|
45
|
+
REQUIRE(sketch.get_theta() == 1.0); // empty sketch should have theta 1.0
|
47
46
|
REQUIRE(sketch.get_rf() == theta_constants::resize_factor::X2);
|
48
47
|
REQUIRE(sketch.get_seed_hash() == compute_seed_hash(123));
|
48
|
+
sketch.update(1, 0);
|
49
|
+
REQUIRE(sketch.get_theta() == 0.5); // theta = p
|
49
50
|
}
|
50
51
|
|
51
52
|
TEST_CASE("tuple sketch float: empty", "[tuple_sketch]") {
|
@@ -58,7 +59,7 @@ TEST_CASE("tuple sketch float: empty", "[tuple_sketch]") {
|
|
58
59
|
REQUIRE(update_sketch.get_upper_bound(1) == 0);
|
59
60
|
REQUIRE(update_sketch.get_theta() == 1);
|
60
61
|
REQUIRE(update_sketch.get_num_retained() == 0);
|
61
|
-
REQUIRE(
|
62
|
+
REQUIRE(update_sketch.is_ordered());
|
62
63
|
|
63
64
|
auto compact_sketch = update_sketch.compact();
|
64
65
|
std::cout << "sizeof(compact_tuple_sketch<float>)=" << sizeof(compact_sketch) << std::endl;
|
@@ -70,6 +71,33 @@ TEST_CASE("tuple sketch float: empty", "[tuple_sketch]") {
|
|
70
71
|
REQUIRE(compact_sketch.get_theta() == 1);
|
71
72
|
REQUIRE(compact_sketch.get_num_retained() == 0);
|
72
73
|
REQUIRE(compact_sketch.is_ordered());
|
74
|
+
|
75
|
+
// empty is forced to be ordered
|
76
|
+
REQUIRE(update_sketch.compact(false).is_ordered());
|
77
|
+
}
|
78
|
+
|
79
|
+
TEST_CASE("tuple sketch: single item", "[theta_sketch]") {
|
80
|
+
auto update_sketch = update_tuple_sketch<float>::builder().build();
|
81
|
+
update_sketch.update(1, 1.0f);
|
82
|
+
REQUIRE_FALSE(update_sketch.is_empty());
|
83
|
+
REQUIRE_FALSE(update_sketch.is_estimation_mode());
|
84
|
+
REQUIRE(update_sketch.get_theta() == 1.0);
|
85
|
+
REQUIRE(update_sketch.get_estimate() == 1.0);
|
86
|
+
REQUIRE(update_sketch.get_lower_bound(1) == 1.0);
|
87
|
+
REQUIRE(update_sketch.get_upper_bound(1) == 1.0);
|
88
|
+
REQUIRE(update_sketch.is_ordered()); // one item is ordered
|
89
|
+
|
90
|
+
auto compact_sketch = update_sketch.compact();
|
91
|
+
REQUIRE_FALSE(compact_sketch.is_empty());
|
92
|
+
REQUIRE_FALSE(compact_sketch.is_estimation_mode());
|
93
|
+
REQUIRE(compact_sketch.get_theta() == 1.0);
|
94
|
+
REQUIRE(compact_sketch.get_estimate() == 1.0);
|
95
|
+
REQUIRE(compact_sketch.get_lower_bound(1) == 1.0);
|
96
|
+
REQUIRE(compact_sketch.get_upper_bound(1) == 1.0);
|
97
|
+
REQUIRE(compact_sketch.is_ordered());
|
98
|
+
|
99
|
+
// single item is forced to be ordered
|
100
|
+
REQUIRE(update_sketch.compact(false).is_ordered());
|
73
101
|
}
|
74
102
|
|
75
103
|
TEST_CASE("tuple sketch float: exact mode", "[tuple_sketch]") {
|
@@ -78,14 +106,14 @@ TEST_CASE("tuple sketch float: exact mode", "[tuple_sketch]") {
|
|
78
106
|
update_sketch.update(2, 2.0f);
|
79
107
|
update_sketch.update(1, 1.0f);
|
80
108
|
// std::cout << update_sketch.to_string(true);
|
81
|
-
|
82
|
-
|
109
|
+
REQUIRE_FALSE(update_sketch.is_empty());
|
110
|
+
REQUIRE_FALSE(update_sketch.is_estimation_mode());
|
83
111
|
REQUIRE(update_sketch.get_estimate() == 2);
|
84
112
|
REQUIRE(update_sketch.get_lower_bound(1) == 2);
|
85
113
|
REQUIRE(update_sketch.get_upper_bound(1) == 2);
|
86
114
|
REQUIRE(update_sketch.get_theta() == 1);
|
87
115
|
REQUIRE(update_sketch.get_num_retained() == 2);
|
88
|
-
|
116
|
+
REQUIRE_FALSE(update_sketch.is_ordered());
|
89
117
|
int count = 0;
|
90
118
|
for (const auto& entry: update_sketch) {
|
91
119
|
REQUIRE(entry.second == 2);
|
@@ -95,8 +123,8 @@ TEST_CASE("tuple sketch float: exact mode", "[tuple_sketch]") {
|
|
95
123
|
|
96
124
|
auto compact_sketch = update_sketch.compact();
|
97
125
|
// std::cout << compact_sketch.to_string(true);
|
98
|
-
|
99
|
-
|
126
|
+
REQUIRE_FALSE(compact_sketch.is_empty());
|
127
|
+
REQUIRE_FALSE(compact_sketch.is_estimation_mode());
|
100
128
|
REQUIRE(compact_sketch.get_estimate() == 2);
|
101
129
|
REQUIRE(compact_sketch.get_lower_bound(1) == 2);
|
102
130
|
REQUIRE(compact_sketch.get_upper_bound(1) == 2);
|
@@ -151,6 +179,16 @@ TEST_CASE("tuple sketch float: exact mode", "[tuple_sketch]") {
|
|
151
179
|
++it;
|
152
180
|
}
|
153
181
|
}
|
182
|
+
|
183
|
+
update_sketch.reset();
|
184
|
+
REQUIRE(update_sketch.is_empty());
|
185
|
+
REQUIRE_FALSE(update_sketch.is_estimation_mode());
|
186
|
+
REQUIRE(update_sketch.get_estimate() == 0);
|
187
|
+
REQUIRE(update_sketch.get_lower_bound(1) == 0);
|
188
|
+
REQUIRE(update_sketch.get_upper_bound(1) == 0);
|
189
|
+
REQUIRE(update_sketch.get_theta() == 1);
|
190
|
+
REQUIRE(update_sketch.get_num_retained() == 0);
|
191
|
+
REQUIRE(update_sketch.is_ordered());
|
154
192
|
}
|
155
193
|
|
156
194
|
template<typename T>
|
@@ -81,6 +81,13 @@ TEST_CASE("tuple_union float: simple case", "[tuple union]") {
|
|
81
81
|
u.update(update_sketch2);
|
82
82
|
auto result = u.get_result();
|
83
83
|
REQUIRE(result.get_num_retained() == 3);
|
84
|
+
|
85
|
+
u.reset();
|
86
|
+
result = u.get_result();
|
87
|
+
REQUIRE(result.is_empty());
|
88
|
+
REQUIRE(result.get_num_retained() == 0);
|
89
|
+
REQUIRE(!result.is_estimation_mode());
|
90
|
+
REQUIRE(result.get_estimate() == 0);
|
84
91
|
}
|
85
92
|
|
86
93
|
TEST_CASE("tuple_union float: exact mode half overlap", "[tuple union]") {
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: datasketches
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.2.
|
4
|
+
version: 0.2.4
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Andrew Kane
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date: 2021-
|
11
|
+
date: 2021-12-28 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: rice
|
@@ -51,6 +51,7 @@ files:
|
|
51
51
|
- vendor/datasketches-cpp/MANIFEST.in
|
52
52
|
- vendor/datasketches-cpp/NOTICE
|
53
53
|
- vendor/datasketches-cpp/README.md
|
54
|
+
- vendor/datasketches-cpp/cmake/DataSketchesConfig.cmake.in
|
54
55
|
- vendor/datasketches-cpp/common/CMakeLists.txt
|
55
56
|
- vendor/datasketches-cpp/common/include/MurmurHash3.h
|
56
57
|
- vendor/datasketches-cpp/common/include/binomial_bounds.hpp
|
@@ -256,14 +257,18 @@ files:
|
|
256
257
|
- vendor/datasketches-cpp/theta/test/CMakeLists.txt
|
257
258
|
- vendor/datasketches-cpp/theta/test/theta_a_not_b_test.cpp
|
258
259
|
- vendor/datasketches-cpp/theta/test/theta_compact_empty_from_java.sk
|
260
|
+
- vendor/datasketches-cpp/theta/test/theta_compact_empty_from_java_v1.sk
|
261
|
+
- vendor/datasketches-cpp/theta/test/theta_compact_empty_from_java_v2.sk
|
259
262
|
- vendor/datasketches-cpp/theta/test/theta_compact_estimation_from_java.sk
|
263
|
+
- vendor/datasketches-cpp/theta/test/theta_compact_estimation_from_java_v1.sk
|
264
|
+
- vendor/datasketches-cpp/theta/test/theta_compact_estimation_from_java_v2.sk
|
265
|
+
- vendor/datasketches-cpp/theta/test/theta_compact_exact_from_java.sk
|
260
266
|
- vendor/datasketches-cpp/theta/test/theta_compact_single_item_from_java.sk
|
261
267
|
- vendor/datasketches-cpp/theta/test/theta_intersection_test.cpp
|
262
268
|
- vendor/datasketches-cpp/theta/test/theta_jaccard_similarity_test.cpp
|
269
|
+
- vendor/datasketches-cpp/theta/test/theta_setop_test.cpp
|
263
270
|
- vendor/datasketches-cpp/theta/test/theta_sketch_test.cpp
|
264
271
|
- vendor/datasketches-cpp/theta/test/theta_union_test.cpp
|
265
|
-
- vendor/datasketches-cpp/theta/test/theta_update_empty_from_java.sk
|
266
|
-
- vendor/datasketches-cpp/theta/test/theta_update_estimation_from_java.sk
|
267
272
|
- vendor/datasketches-cpp/tuple/CMakeLists.txt
|
268
273
|
- vendor/datasketches-cpp/tuple/include/array_of_doubles_a_not_b.hpp
|
269
274
|
- vendor/datasketches-cpp/tuple/include/array_of_doubles_a_not_b_impl.hpp
|
@@ -295,7 +300,7 @@ files:
|
|
295
300
|
- vendor/datasketches-cpp/tuple/test/tuple_sketch_allocation_test.cpp
|
296
301
|
- vendor/datasketches-cpp/tuple/test/tuple_sketch_test.cpp
|
297
302
|
- vendor/datasketches-cpp/tuple/test/tuple_union_test.cpp
|
298
|
-
homepage: https://github.com/ankane/datasketches
|
303
|
+
homepage: https://github.com/ankane/datasketches-ruby
|
299
304
|
licenses:
|
300
305
|
- Apache-2.0
|
301
306
|
metadata: {}
|
@@ -314,7 +319,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
|
|
314
319
|
- !ruby/object:Gem::Version
|
315
320
|
version: '0'
|
316
321
|
requirements: []
|
317
|
-
rubygems_version: 3.
|
322
|
+
rubygems_version: 3.3.3
|
318
323
|
signing_key:
|
319
324
|
specification_version: 4
|
320
325
|
summary: Sketch data structures for Ruby
|
Binary file
|
Binary file
|