-
Notifications
You must be signed in to change notification settings - Fork 2
/
pointrangequery.go
439 lines (370 loc) · 10.8 KB
/
pointrangequery.go
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
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
package search
import (
"bytes"
"context"
"encoding/base64"
"errors"
"fmt"
"github.com/geange/lucene-go/core/types"
"io"
"math"
"github.com/bits-and-blooms/bitset"
"github.com/geange/lucene-go/core/index"
)
var _ Query = &PointRangeQuery{}
type PointRangeQuery struct {
field string
numDims int
bytesPerDim int
lowerPoint []byte
upperPoint []byte
}
func NewPointRangeQuery(field string, lowerPoint []byte, upperPoint []byte, numDims int) (*PointRangeQuery, error) {
if numDims <= 0 {
return nil, errors.New("numDims must be positive")
}
if len(lowerPoint) == 0 {
return nil, errors.New("lowerPoint has length of zero")
}
if len(lowerPoint)%numDims != 0 {
return nil, errors.New("lowerPoint is not a fixed multiple of numDims")
}
if len(lowerPoint) != len(upperPoint) {
return nil, errors.New("lowerPoint's length not equal upperPoint's length")
}
return &PointRangeQuery{
field: field,
numDims: numDims,
bytesPerDim: len(lowerPoint) / numDims,
lowerPoint: lowerPoint,
upperPoint: upperPoint,
}, nil
}
func (p *PointRangeQuery) String(field string) string {
sb := new(bytes.Buffer)
if p.field != field {
sb.WriteString(p.field)
sb.WriteString(":")
}
// print ourselves as "range per dimension"
for i := 0; i < p.numDims; i++ {
if i > 0 {
sb.WriteString(".")
}
startOffset := p.bytesPerDim * i
sb.WriteString("[")
fromDimData := p.lowerPoint[startOffset : startOffset+p.bytesPerDim]
from := fmt.Sprintf("idx=%d value=%s", i, base64.StdEncoding.EncodeToString(fromDimData))
sb.WriteString(from)
sb.WriteString(" TO ")
toDimData := p.upperPoint[startOffset : startOffset+p.bytesPerDim]
to := fmt.Sprintf("idx=%d value=%s", i, base64.StdEncoding.EncodeToString(toDimData))
sb.WriteString(to)
sb.WriteString("]")
}
return sb.String()
}
func (p *PointRangeQuery) CreateWeight(searcher *IndexSearcher, scoreMode ScoreMode, boost float64) (Weight, error) {
// We don't use RandomAccessWeight here: it's no good to approximate with "match all docs".
// This is an inverted structure and should be used in the first pass:
return p.newPrQueryWeight(boost, scoreMode), nil
}
type prQueryWeight struct {
*ConstantScoreWeight
p *PointRangeQuery
scoreMode ScoreMode
}
func (p *PointRangeQuery) newPrQueryWeight(boost float64, scoreMode ScoreMode) *prQueryWeight {
weight := &prQueryWeight{
ConstantScoreWeight: nil,
p: p,
scoreMode: scoreMode,
}
weight.ConstantScoreWeight = NewConstantScoreWeight(boost, p, weight)
return weight
}
func (r *prQueryWeight) matches(packedValue []byte) bool {
for dim := 0; dim < r.p.numDims; dim++ {
fromIndex := dim * r.p.bytesPerDim
toIndex := fromIndex + r.p.bytesPerDim
if bytes.Compare(packedValue[fromIndex:toIndex], r.p.lowerPoint[fromIndex:toIndex]) < 0 {
// Doc's value is too low, in this dimension
return false
}
if bytes.Compare(packedValue[fromIndex:toIndex], r.p.upperPoint[fromIndex:toIndex]) > 0 {
// Doc's value is too high, in this dimension
return false
}
}
return true
}
func (r *prQueryWeight) relate(minPackedValue, maxPackedValue []byte) types.Relation {
crosses := false
for dim := 0; dim < r.p.numDims; dim++ {
offset := dim * r.p.bytesPerDim
toIndex := offset + r.p.bytesPerDim
if bytes.Compare(minPackedValue[offset:toIndex], r.p.upperPoint[offset:toIndex]) > 0 ||
bytes.Compare(maxPackedValue[offset:toIndex], r.p.lowerPoint[offset:toIndex]) < 0 {
return types.CELL_OUTSIDE_QUERY
}
crosses = crosses || (bytes.Compare(minPackedValue[offset:toIndex], r.p.lowerPoint[offset:toIndex]) < 0 ||
bytes.Compare(maxPackedValue[offset:toIndex], r.p.upperPoint[offset:toIndex]) > 0)
}
if crosses {
return types.CELL_CROSSES_QUERY
}
return types.CELL_INSIDE_QUERY
}
func (r *prQueryWeight) getIntersectVisitor(result *DocIdSetBuilder) types.IntersectVisitor {
return &prQueryVisitor{
weight: r,
addr: nil,
result: result,
}
}
var _ types.IntersectVisitor = &prQueryVisitor{}
type prQueryVisitor struct {
weight *prQueryWeight
addr BulkAdder
result *DocIdSetBuilder
}
func (p *prQueryVisitor) Visit(ctx context.Context, docID int) error {
p.addr.Add(docID)
return nil
}
func (p *prQueryVisitor) VisitLeaf(ctx context.Context, docID int, packedValue []byte) error {
if p.weight.matches(packedValue) {
return p.Visit(nil, docID)
}
return nil
}
func (p *prQueryVisitor) VisitIterator(iterator types.DocValuesIterator, packedValue []byte) error {
if p.weight.matches(packedValue) {
for {
doc, err := iterator.NextDoc()
if err != nil {
if errors.Is(err, io.EOF) {
return nil
}
return err
}
err = p.Visit(nil, doc)
if err != nil {
return err
}
}
}
return nil
}
func (p *prQueryVisitor) Compare(minPackedValue, maxPackedValue []byte) types.Relation {
return p.weight.relate(minPackedValue, maxPackedValue)
}
func (p *prQueryVisitor) Grow(count int) {
p.addr = p.result.Grow(count)
}
func (r *prQueryWeight) getInverseIntersectVisitor(result *bitset.BitSet, cost []int64) types.IntersectVisitor {
panic("")
}
var _ types.IntersectVisitor = &invPrQueryVisitor{}
type invPrQueryVisitor struct {
result *bitset.BitSet
cost []int64
weight *prQueryWeight
}
func (r *invPrQueryVisitor) Visit(ctx context.Context, docID int) error {
r.result.Clear(uint(docID))
r.cost[0]--
return nil
}
func (r *invPrQueryVisitor) VisitLeaf(ctx context.Context, docID int, packedValue []byte) error {
if r.weight.matches(packedValue) == false {
return r.Visit(nil, docID)
}
return nil
}
func (r *invPrQueryVisitor) VisitIterator(iterator types.DocValuesIterator, packedValue []byte) error {
if r.weight.matches(packedValue) == false {
for {
doc, err := iterator.NextDoc()
if err != nil {
if errors.Is(err, io.EOF) {
return nil
}
return err
}
err = r.Visit(nil, doc)
if err != nil {
return err
}
}
}
return nil
}
func (r *invPrQueryVisitor) Compare(minPackedValue, maxPackedValue []byte) types.Relation {
relation := r.weight.relate(minPackedValue, maxPackedValue)
switch relation {
case types.CELL_INSIDE_QUERY:
// all points match, skip this subtree
return types.CELL_OUTSIDE_QUERY
case types.CELL_OUTSIDE_QUERY:
// none of the points match, clear all documents
return types.CELL_INSIDE_QUERY
default:
return relation
}
}
func (r *invPrQueryVisitor) Grow(count int) {
return
}
func (r *prQueryWeight) ScorerSupplier(ctx index.LeafReaderContext) (ScorerSupplier, error) {
reader, ok := ctx.Reader().(index.LeafReader)
if !ok {
return nil, errors.New("get reader error")
}
field := r.p.field
values, exist := reader.GetPointValues(field)
if !exist {
return nil, nil
}
dimensions, err := values.GetNumIndexDimensions()
if err != nil {
return nil, err
}
if dimensions != r.p.numDims {
return nil, fmt.Errorf("field=%s numIndexDimensions not equal", field)
}
bytesPerDimension, err := values.GetBytesPerDimension()
if err != nil {
return nil, err
}
if bytesPerDimension != r.p.bytesPerDim {
return nil, fmt.Errorf("field=%s bytesPerDim not equal", field)
}
allDocsMatch := false
if values.GetDocCount() == reader.MaxDoc() {
fieldPackedLower, err := values.GetMinPackedValue()
if err != nil {
return nil, err
}
fieldPackedUpper, err := values.GetMaxPackedValue()
if err != nil {
return nil, err
}
allDocsMatch = true
for i := 0; i < r.p.numDims; i++ {
offset := i * r.p.bytesPerDim
toIndex := offset + r.p.bytesPerDim
if bytes.Compare(r.p.lowerPoint[offset:toIndex], fieldPackedLower[offset:toIndex]) > 0 ||
bytes.Compare(r.p.upperPoint[offset:toIndex], fieldPackedUpper[offset:toIndex]) < 0 {
allDocsMatch = false
break
}
}
}
if allDocsMatch {
// all docs have a value and all points are within bounds, so everything matches
return &allDocsScorerSupplier{
reader: reader,
weight: r,
scoreMode: r.scoreMode,
}, nil
} else {
result := NewDocIdSetBuilderV2(reader.MaxDoc(), values, field)
return ¬AllDocsScorerSupplier{
result: result,
visitor: r.getIntersectVisitor(result),
reader: reader,
values: values,
cost: -1,
weight: r,
scoreMode: r.scoreMode,
}, nil
}
}
var _ ScorerSupplier = &allDocsScorerSupplier{}
type allDocsScorerSupplier struct {
reader index.LeafReader
weight *prQueryWeight
scoreMode ScoreMode
}
func (r *allDocsScorerSupplier) Get(leadCost int64) (Scorer, error) {
return NewConstantScoreScorer(r.weight, r.weight.Score(), r.scoreMode, types.DocIdSetIteratorAll(r.reader.MaxDoc()))
}
func (a *allDocsScorerSupplier) Cost() int64 {
return int64(a.reader.MaxDoc())
}
var _ ScorerSupplier = ¬AllDocsScorerSupplier{}
type notAllDocsScorerSupplier struct {
result *DocIdSetBuilder
visitor types.IntersectVisitor
reader index.LeafReader
values types.PointValues
cost int64
weight *prQueryWeight
scoreMode ScoreMode
}
func (r *notAllDocsScorerSupplier) Get(leadCost int64) (Scorer, error) {
if r.values.GetDocCount() == r.reader.MaxDoc() &&
r.values.GetDocCount() == r.values.Size() &&
r.Cost() > int64(r.reader.MaxDoc()/2) {
// If all docs have exactly one value and the cost is greater
// than half the leaf size then maybe we can make things faster
// by computing the set of documents that do NOT match the range
result := bitset.New(uint(r.reader.MaxDoc()))
for i := range result.Bytes() {
result.Bytes()[i] = math.MaxUint64
}
cost := []int64{int64(r.reader.MaxDoc())}
err := r.values.Intersect(nil, r.weight.getInverseIntersectVisitor(result, cost))
if err != nil {
return nil, err
}
iterator := index.NewBitSetIterator(result, cost[0])
return NewConstantScoreScorer(r.weight, r.weight.Score(), r.scoreMode, iterator)
}
err := r.values.Intersect(nil, r.visitor)
if err != nil {
return nil, err
}
iterator := r.result.Build().Iterator()
return NewConstantScoreScorer(r.weight, r.weight.Score(), r.scoreMode, iterator)
}
func (r *notAllDocsScorerSupplier) Cost() int64 {
if r.cost == -1 {
// Computing the cost may be expensive, so only do it if necessary
cost, _ := r.values.EstimateDocCount(r.visitor)
r.cost = int64(cost)
//assert cost >= 0;
}
return r.cost
}
func (r *prQueryWeight) Scorer(ctx index.LeafReaderContext) (Scorer, error) {
scorerSupplier, err := r.ScorerSupplier(ctx)
if err != nil {
return nil, err
}
if scorerSupplier == nil {
return nil, nil
}
return scorerSupplier.Get(math.MaxInt32)
}
func (r *prQueryWeight) IsCacheable(ctx index.LeafReaderContext) bool {
return true
}
func (p *PointRangeQuery) Rewrite(reader index.IndexReader) (Query, error) {
return p, nil
}
func (p *PointRangeQuery) Visit(visitor QueryVisitor) (err error) {
if visitor.AcceptField(p.field) {
err := visitor.VisitLeaf(p)
if err != nil {
return err
}
}
return nil
}
func copyOfSubArray(bs []byte, from, to int) []byte {
newBytes := make([]byte, to-from)
copy(newBytes, bs[from:])
return newBytes
}