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
|
// Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
#pragma once
#include "rise_wand.h"
#include <vespa/searchlib/queryeval/wand/wand_parts.h>
#include <cmath>
using search::queryeval::wand::TermFrequencyScorer;
namespace rise {
template <typename Scorer, typename Cmp>
RiseWand<Scorer, Cmp>::RiseWand(const Terms &terms, uint32_t n)
: _numStreams(0),
_streams(),
_lastPivotIdx(0),
_streamDocIds(new docid_t[terms.size()]),
_streamIndices(new uint16_t[terms.size()]),
_streamIndicesAux(new uint16_t[terms.size()]),
_streamComparator(_streamDocIds),
_scorer(),
_n(n),
_limit(1),
_streamScores(new score_t[terms.size()]),
_scores(),
_terms(terms)
{
for (size_t i = 0; i < terms.size(); ++i) {
_terms[i].maxScore = _scorer.calculateMaxScore(terms[i]);
_streamScores[i] = _terms[i].maxScore;
_streams.push_back(terms[i].search);
}
_numStreams = _streams.size();
if (_numStreams == 0) {
setAtEnd();
}
for (uint32_t i=0; i<_numStreams; ++i) {
_streamIndices[i] = i;
}
for (uint32_t i=0; i<_numStreams; ++i) {
_streamDocIds[i] = _streams[i]->getDocId();
}
std::sort(_streamIndices, _streamIndices+_numStreams, _streamComparator);
}
template <typename Scorer, typename Cmp>
RiseWand<Scorer, Cmp>::~RiseWand()
{
for (auto * stream : _streams) {
delete stream;
}
delete [] _streamScores;
delete [] _streamIndicesAux;
delete [] _streamIndices;
delete [] _streamDocIds;
}
template <typename Scorer, typename Cmp>
void
RiseWand<Scorer, Cmp>::next()
{
// We do not check whether the stream is already at the end
// here based on the assumption that application won't call
// next() for streams that are already at the end, or atleast
// won't do this frequently.
uint32_t pivotIdx;
docid_t pivotDocId = search::endDocId;
score_t threshold = _limit;
while (true) {
if (!_findPivotFeatureIdx(threshold, pivotIdx)) {
setAtEnd();
return;
}
pivotDocId = _streamDocIds[_streamIndices[pivotIdx]];
if (_streamDocIds[_streamIndices[0]] == _streamDocIds[_streamIndices[pivotIdx]]) {
// Found candidate. All cursors before (*_streams)[pivotIdx] point to
// the same doc and this doc is the candidate for full evaluation.
setDocId(pivotDocId);
// Advance pivotIdx sufficiently so that all instances of pivotDocId are included
while (pivotIdx < _numStreams-1 && _streamDocIds[_streamIndices[pivotIdx+1]] == pivotDocId) {
++pivotIdx;
}
_lastPivotIdx = pivotIdx;
return; // scoring and threshold adjustment is done in doUnpack
} else { // not all cursors upto the pivot are aligned at the same doc yet
// decreases pivotIdx to the first stream pointing at the pivotDocId
while (pivotIdx && _streamDocIds[_streamIndices[pivotIdx-1]] == pivotDocId) {
--pivotIdx;
}
_moveStreamsToDocAndSort(pivotIdx, pivotDocId);
}
} /* while (true) */
}
template <typename Scorer, typename Cmp>
bool
RiseWand<Scorer, Cmp>::_findPivotFeatureIdx(const score_t threshold, uint32_t &pivotIdx)
{
uint32_t idx;
score_t accumUB = 0;
for (idx=0;
!Cmp()(accumUB, threshold) && idx < _numStreams;
++idx) {
accumUB += _streamScores[_streamIndices[idx]];
}
if( Cmp()(accumUB, threshold) ) {
pivotIdx = idx - 1;
return true;
}
return false;
}
template <typename Scorer, typename Cmp>
void
RiseWand<Scorer, Cmp>::_moveStreamsAndSort(const uint32_t numStreamsToMove)
{
for (uint32_t i=0; i<numStreamsToMove; ++i) {
_streams[_streamIndices[i]]->seek(_streams[_streamIndices[i]]->getDocId() + 1);
_streamDocIds[_streamIndices[i]] = _streams[_streamIndices[i]]->getDocId();
}
_sortMerge(numStreamsToMove);
}
template <typename Scorer, typename Cmp>
void
RiseWand<Scorer, Cmp>::_moveStreamsToDocAndSort(const uint32_t numStreamsToMove, const docid_t desiredDocId)
{
for (uint32_t i=0; i<numStreamsToMove; ++i) {
_streams[_streamIndices[i]]->seek(desiredDocId);
_streamDocIds[_streamIndices[i]] = _streams[_streamIndices[i]]->getDocId();
}
_sortMerge(numStreamsToMove);
}
template <typename Scorer, typename Cmp>
void RiseWand<Scorer, Cmp>::_sortMerge(const uint32_t numStreamsToMove)
{
for (uint32_t i=0; i<numStreamsToMove; ++i) {
_streamIndicesAux[i] = _streamIndices[i];
}
std::sort(_streamIndicesAux, _streamIndicesAux+numStreamsToMove, _streamComparator);
uint16_t j=numStreamsToMove, k=0, i=0;
while (i < numStreamsToMove && j < _numStreams) {
if (_streamComparator(_streamIndicesAux[i], _streamIndices[j])) {
_streamIndices[k++] = _streamIndicesAux[i++];
}
else {
_streamIndices[k++] = _streamIndices[j++];
}
}
if (j == _numStreams) {
while (i < numStreamsToMove) {
_streamIndices[k++] = _streamIndicesAux[i++];
}
}
while (_numStreams &&
_streamDocIds[_streamIndices[_numStreams-1]] == search::endDocId) {
--_numStreams;
}
}
template <typename Scorer, typename Cmp>
void
RiseWand<Scorer, Cmp>::doSeek(uint32_t docid)
{
if (getDocId() != beginId() && (docid - 1) == getDocId()) {
_moveStreamsAndSort(_lastPivotIdx + 1);
} else {
_moveStreamsToDocAndSort(_numStreams, docid);
}
next();
}
template <typename Scorer, typename Cmp>
void
RiseWand<Scorer, Cmp>::doUnpack(uint32_t docid)
{
score_t score = 0;
for (size_t i = 0; i <= _lastPivotIdx; ++i) {
score += _scorer.calculateScore(_terms[_streamIndices[i]], docid);
}
if (_scores.size() < _n || _scores.front() < score) {
_scores.push(score);
if (_scores.size() > _n) {
_scores.pop_front();
}
if (_scores.size() == _n) {
_limit = _scores.front();
}
}
}
} // namespace rise
|
