1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
|
// Copyright Yahoo. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
#include "emptysearch.h"
#include "nearest_neighbor_blueprint.h"
#include "nearest_neighbor_iterator.h"
#include "nns_index_iterator.h"
#include <vespa/eval/eval/fast_value.h>
#include <vespa/searchlib/fef/termfieldmatchdataarray.h>
#include <vespa/searchlib/tensor/dense_tensor_attribute.h>
#include <vespa/searchlib/tensor/distance_function_factory.h>
#include <vespa/log/log.h>
LOG_SETUP(".searchlib.queryeval.nearest_neighbor_blueprint");
using vespalib::eval::CellType;
using vespalib::eval::FastValueBuilderFactory;
using vespalib::eval::TypedCells;
using vespalib::eval::Value;
using vespalib::eval::ValueType;
namespace search::queryeval {
namespace {
template<typename LCT, typename RCT>
std::unique_ptr<Value>
convert_cells(const ValueType &new_type, std::unique_ptr<Value> old_value)
{
auto old_cells = old_value->cells().typify<LCT>();
auto builder = FastValueBuilderFactory::get().create_value_builder<RCT>(new_type);
auto new_cells = builder->add_subspace();
assert(old_cells.size() == new_cells.size());
auto p = new_cells.begin();
for (LCT value : old_cells) {
RCT conv(value);
*p++ = conv;
}
return builder->build(std::move(builder));
}
struct ConvertCellsSelector
{
template <typename LCT, typename RCT>
static auto invoke(const ValueType &new_type, std::unique_ptr<Value> old_value) {
return convert_cells<LCT, RCT>(new_type, std::move(old_value));
}
auto operator() (CellType from, CellType to, std::unique_ptr<Value> old_value) const {
using MyTypify = vespalib::eval::TypifyCellType;
ValueType new_type = old_value->type().cell_cast(to);
return vespalib::typify_invoke<2,MyTypify,ConvertCellsSelector>(from, to, new_type, std::move(old_value));
}
};
vespalib::string
to_string(NearestNeighborBlueprint::Algorithm algorithm)
{
using NNBA = NearestNeighborBlueprint::Algorithm;
switch (algorithm) {
case NNBA::BRUTE_FORCE: return "brute force";
case NNBA::BRUTE_FORCE_FALLBACK: return "brute force fallback";
case NNBA::INDEX_TOP_K: return "index top k";
case NNBA::INDEX_TOP_K_WITH_FILTER: return "index top k using global filter";
}
return "unknown";
}
} // namespace <unnamed>
NearestNeighborBlueprint::NearestNeighborBlueprint(const queryeval::FieldSpec& field,
const tensor::ITensorAttribute& attr_tensor,
std::unique_ptr<Value> query_tensor,
uint32_t target_num_hits,
bool approximate,
uint32_t explore_additional_hits,
double distance_threshold,
double global_filter_lower_limit,
double global_filter_upper_limit)
: ComplexLeafBlueprint(field),
_attr_tensor(attr_tensor),
_query_tensor(std::move(query_tensor)),
_target_num_hits(target_num_hits),
_approximate(approximate),
_explore_additional_hits(explore_additional_hits),
_distance_threshold(std::numeric_limits<double>::max()),
_global_filter_lower_limit(global_filter_lower_limit),
_global_filter_upper_limit(global_filter_upper_limit),
_fallback_dist_fun(),
_distance_heap(target_num_hits),
_found_hits(),
_algorithm(Algorithm::BRUTE_FORCE),
_global_filter(GlobalFilter::create()),
_global_filter_hits(),
_global_filter_hit_ratio()
{
CellType attr_ct = _attr_tensor.getTensorType().cell_type();
_fallback_dist_fun = search::tensor::make_distance_function(_attr_tensor.distance_metric(), attr_ct);
_dist_fun = _fallback_dist_fun.get();
assert(_dist_fun);
auto nns_index = _attr_tensor.nearest_neighbor_index();
if (nns_index) {
_dist_fun = nns_index->distance_function();
assert(_dist_fun);
}
auto query_ct = _query_tensor->cells().type;
CellType required_ct = _dist_fun->expected_cell_type();
if (query_ct != required_ct) {
ConvertCellsSelector converter;
_query_tensor = converter(query_ct, required_ct, std::move(_query_tensor));
}
if (distance_threshold < std::numeric_limits<double>::max()) {
_distance_threshold = _dist_fun->convert_threshold(distance_threshold);
_distance_heap.set_distance_threshold(_distance_threshold);
}
uint32_t est_hits = _attr_tensor.get_num_docs();
setEstimate(HitEstimate(est_hits, false));
set_want_global_filter(nns_index && _approximate);
}
NearestNeighborBlueprint::~NearestNeighborBlueprint() = default;
void
NearestNeighborBlueprint::set_global_filter(const GlobalFilter &global_filter)
{
_global_filter = global_filter.shared_from_this();
auto nns_index = _attr_tensor.nearest_neighbor_index();
LOG(debug, "set_global_filter with: %s / %s / %s",
(_approximate ? "approximate" : "exact"),
(nns_index ? "nns_index" : "no_index"),
(_global_filter->has_filter() ? "has_filter" : "no_filter"));
if (_approximate && nns_index) {
uint32_t est_hits = _attr_tensor.get_num_docs();
if (_global_filter->has_filter()) {
uint32_t max_hits = _global_filter->filter()->countTrueBits();
LOG(debug, "set_global_filter getNumDocs: %u / max_hits %u", est_hits, max_hits);
double max_hit_ratio = static_cast<double>(max_hits) / est_hits;
if (max_hit_ratio < _global_filter_lower_limit) {
_approximate = false;
_algorithm = Algorithm::BRUTE_FORCE_FALLBACK;
LOG(debug, "too many hits filtered out, using brute force implementation");
} else {
est_hits = std::min(est_hits, max_hits);
}
_global_filter_hits = max_hits;
_global_filter_hit_ratio = max_hit_ratio;
}
if (_approximate) {
est_hits = std::min(est_hits, _target_num_hits);
setEstimate(HitEstimate(est_hits, false));
perform_top_k();
LOG(debug, "perform_top_k found %zu hits", _found_hits.size());
}
}
}
void
NearestNeighborBlueprint::perform_top_k()
{
auto nns_index = _attr_tensor.nearest_neighbor_index();
if (_approximate && nns_index) {
auto lhs = _query_tensor->cells();
uint32_t k = _target_num_hits;
if (_global_filter->has_filter()) {
auto filter = _global_filter->filter();
_found_hits = nns_index->find_top_k_with_filter(k, lhs, *filter, k + _explore_additional_hits, _distance_threshold);
_algorithm = Algorithm::INDEX_TOP_K_WITH_FILTER;
} else {
_found_hits = nns_index->find_top_k(k, lhs, k + _explore_additional_hits, _distance_threshold);
_algorithm = Algorithm::INDEX_TOP_K;
}
}
}
std::unique_ptr<SearchIterator>
NearestNeighborBlueprint::createLeafSearch(const search::fef::TermFieldMatchDataArray& tfmda, bool strict) const
{
assert(tfmda.size() == 1);
fef::TermFieldMatchData &tfmd = *tfmda[0]; // always search in only one field
if (! _found_hits.empty()) {
return NnsIndexIterator::create(tfmd, _found_hits, _dist_fun);
}
const Value &qT = *_query_tensor;
return NearestNeighborIterator::create(strict, tfmd, qT, _attr_tensor,
_distance_heap, _global_filter->filter(), _dist_fun);
}
void
NearestNeighborBlueprint::visitMembers(vespalib::ObjectVisitor& visitor) const
{
ComplexLeafBlueprint::visitMembers(visitor);
visitor.visitString("attribute_tensor", _attr_tensor.getTensorType().to_spec());
visitor.visitString("query_tensor", _query_tensor->type().to_spec());
visitor.visitInt("target_num_hits", _target_num_hits);
visitor.visitInt("explore_additional_hits", _explore_additional_hits);
visitor.visitBool("approximate", _approximate);
visitor.visitBool("has_index", _attr_tensor.nearest_neighbor_index());
visitor.visitString("algorithm", to_string(_algorithm));
visitor.visitInt("top_k_hits", _found_hits.size());
visitor.openStruct("global_filter", "GlobalFilter");
visitor.visitBool("is_set", (_global_filter != nullptr));
visitor.visitBool("has_filter", (_global_filter && _global_filter->has_filter()));
visitor.visitFloat("lower_limit", _global_filter_lower_limit);
visitor.visitFloat("upper_limit", _global_filter_upper_limit);
if (_global_filter_hits.has_value()) {
visitor.visitInt("hits", _global_filter_hits.value());
}
if (_global_filter_hit_ratio.has_value()) {
visitor.visitFloat("hit_ratio", _global_filter_hit_ratio.value());
}
visitor.closeStruct();
}
bool
NearestNeighborBlueprint::always_needs_unpack() const
{
return true;
}
std::ostream&
operator<<(std::ostream& out, NearestNeighborBlueprint::Algorithm algorithm)
{
out << to_string(algorithm);
return out;
}
}
|
