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
|
// Copyright Yahoo. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
#include "nns_index_iterator.h"
#include <vespa/searchlib/tensor/nearest_neighbor_index.h>
#include <cmath>
using Neighbor = search::tensor::NearestNeighborIndex::Neighbor;
namespace search::queryeval {
/**
* Search iterator for K nearest neighbor matching,
* where the actual search is done up front and this class
* just iterates over a vector held by the blueprint.
**/
class NeighborVectorIterator : public NnsIndexIterator
{
private:
fef::TermFieldMatchData &_tfmd;
const std::vector<Neighbor> &_hits;
const search::tensor::BoundDistanceFunction &_dist_fun;
uint32_t _idx;
double _last_abstract_dist;
public:
NeighborVectorIterator(fef::TermFieldMatchData &tfmd,
const std::vector<Neighbor> &hits,
const search::tensor::BoundDistanceFunction &dist_fun)
: _tfmd(tfmd),
_hits(hits),
_dist_fun(dist_fun),
_idx(0),
_last_abstract_dist(0.0)
{}
void initRange(uint32_t begin_id, uint32_t end_id) override {
SearchIterator::initRange(begin_id, end_id);
_idx = 0;
}
void doSeek(uint32_t docId) override {
while (_idx < _hits.size()) {
uint32_t hit_id = _hits[_idx].docid;
if (hit_id < docId) {
++_idx;
} else if (hit_id < getEndId()) {
setDocId(hit_id);
_last_abstract_dist = _hits[_idx].distance;
return;
} else {
_idx = _hits.size();
}
}
setAtEnd();
}
void doUnpack(uint32_t docId) override {
double score = _dist_fun.to_rawscore(_last_abstract_dist);
_tfmd.setRawScore(docId, score);
}
Trinary is_strict() const override { return Trinary::True; }
};
std::unique_ptr<NnsIndexIterator>
NnsIndexIterator::create(
fef::TermFieldMatchData &tfmd,
const std::vector<Neighbor> &hits,
const search::tensor::BoundDistanceFunction &dist_fun)
{
return std::make_unique<NeighborVectorIterator>(tfmd, hits, dist_fun);
}
} // namespace
|
