-
Notifications
You must be signed in to change notification settings - Fork 29
/
ferret.go
488 lines (474 loc) · 13.2 KB
/
ferret.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
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
// Copyright 2013 Mark Canning
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
//
// Author: Mark Canning
// Developed at: Tamber, Inc. (http://www.tamber.com/).
// Package ferret implements a fast in-memory substring search engine
package ferret // import "github.com/argusdusty/Ferret"
import "sort"
// InvertedSuffix implements the data structure for substring searches
type InvertedSuffix struct {
WordIndex []int // WordIndex and SuffixIndex are sorted by Words[WordIndex[i]][SuffixIndex[i]:]
SuffixIndex []int // WordIndex and SuffixIndex are sorted by Words[WordIndex[i]][SuffixIndex[i]:]
Words [][]byte // Words is the list of words (in []byte form) to perform substring searches over
Results []string // Results is the string value of the words. Used as a return value
Values []interface{} // Values is some data mapped to the words. Can be used for sorting, or as a return value
Converter func(string) []byte // Converter converts an inserted word/query to a byte array to search for/with
}
// A wrapper type used to sort the three arrays according to sort.sort
type sortWrapper struct {
WordIndex []int
SuffixIndex []int
Words [][]byte
}
func (SW *sortWrapper) Swap(i, j int) {
SW.WordIndex[i], SW.WordIndex[j] = SW.WordIndex[j], SW.WordIndex[i]
SW.SuffixIndex[i], SW.SuffixIndex[j] = SW.SuffixIndex[j], SW.SuffixIndex[i]
}
func (SW *sortWrapper) Len() int {
return len(SW.WordIndex)
}
// Equivalent to:
// bytes.Compare(S.Words[S.WordIndex[i]][S.SuffixIndex[i]:], S.Words[S.WordIndex[j]][S.SuffixIndex[j]:]) <= 0
// but faster
func (SW *sortWrapper) Less(i, j int) bool {
x := SW.WordIndex[i]
y := SW.WordIndex[j]
a := SW.Words[x]
b := SW.Words[y]
pa := SW.SuffixIndex[i]
pb := SW.SuffixIndex[j]
na := len(a)
nb := len(b)
for pa < na && pb < nb {
wa := a[pa]
wb := b[pb]
if wa < wb {
return true
}
if wb < wa {
return false
}
pa++
pb++
}
return pa == na
}
// New creates an inverted suffix from a dictionary of byte arrays, mapping data, and a string->[]byte converter
func New(Words, Results []string, Data []interface{}, Converter func(string) []byte) *InvertedSuffix {
CharCount := 0
NewWords := make([][]byte, len(Words))
for i, Word := range Words {
NewWord := Converter(Word)
NewWords[i] = NewWord
CharCount += len(NewWord)
}
WordIndex := make([]int, 0, CharCount)
SuffixIndex := make([]int, 0, CharCount)
for i, NewWord := range NewWords {
for j := 0; j < len(NewWord); j++ {
WordIndex = append(WordIndex, i)
SuffixIndex = append(SuffixIndex, j)
}
}
sort.Sort(&sortWrapper{WordIndex, SuffixIndex, NewWords})
Suffixes := &InvertedSuffix{WordIndex, SuffixIndex, NewWords, Results, Data, Converter}
return Suffixes
}
// Insert adds a word to the dictionary that IS was built on.
// This is pretty slow, because of linear-time insertion into an array,
// so stick to New when you can
func (IS *InvertedSuffix) Insert(Word, Result string, Data interface{}) {
Query := IS.Converter(Word)
low, high := IS.Search(Query)
for k := low; k < high; k++ {
if IS.Results[IS.WordIndex[k]] == Word {
IS.Values[IS.WordIndex[k]] = Data
return
}
}
i := len(IS.Words)
IS.Words = append(IS.Words, Query)
Length := len(Query)
IS.Results = append(IS.Results, Result)
IS.Values = append(IS.Values, Data)
for j := 0; j < Length; j++ {
k, _ := IS.Search(Query[j:])
IS.WordIndex = append(IS.WordIndex, 0)
copy(IS.WordIndex[k+1:], IS.WordIndex[k:])
IS.WordIndex[k] = i
IS.SuffixIndex = append(IS.SuffixIndex, 0)
copy(IS.SuffixIndex[k+1:], IS.SuffixIndex[k:])
IS.SuffixIndex[k] = j
}
}
// Search performs an exact substring search for the query in the word dictionary
// Returns the boundaries (low/high) of sorted suffixes which have the query as a prefix
// This is a low-level interface. I wouldn't recommend using this yourself
func (IS *InvertedSuffix) Search(Query []byte) (int, int) {
low := 0
high := len(IS.WordIndex)
n := len(Query)
for a := 0; a < n; a++ {
c := Query[a]
oldlow := low
oldhigh := high
i := low
j := high
// Raise the lower-bound
for i < j {
h := (i + j) >> 1
Index := IS.WordIndex[h]
Word := IS.Words[Index]
Length := len(Word)
d := IS.SuffixIndex[h] + a
if d >= Length {
i = h + 1
} else {
e := Word[d]
if e < c {
i = h + 1
} else {
j = h
if e > c {
high = h
}
}
}
}
low = i
if low == high {
break
}
j = high
if low == oldlow && high == oldhigh {
// nothing to do here: Word[IS.SuffixIndex[(i + j) >> 1] + a] == c
continue
}
// Lower the upper-bound
for i < j {
h := (i + j) >> 1
Index := IS.WordIndex[h]
Word := IS.Words[Index]
Length := len(Word)
d := IS.SuffixIndex[h] + a
if d >= Length {
i = h + 1
} else {
e := Word[d]
if e <= c {
i = h + 1
if e < c {
low = i
}
} else {
j = h
}
}
}
high = j
if low == high {
break
}
}
return low, high
}
// Query returns the strings which contain the query, and their stored values unsorted
// Input:
// Word: The substring to search for.
// ResultsLimit: Limit the results to some number of values. Set to -1 for no limit
func (IS *InvertedSuffix) Query(Word string, ResultsLimit int) ([]string, []interface{}) {
Query := IS.Converter(Word)
if ResultsLimit == 0 {
return []string{}, []interface{}{}
}
if ResultsLimit < 0 {
ResultsLimit = 0
}
Results := make([]string, 0, ResultsLimit)
Values := make([]interface{}, 0, ResultsLimit)
low, high := IS.Search(Query)
a := 0
used := make(map[int]bool, 0)
for k := low; k < high; k++ {
x := IS.WordIndex[k]
if _, ok := used[x]; ok {
continue
}
used[x] = true
Results = append(Results, IS.Results[x])
Values = append(Values, IS.Values[x])
a++
if a == ResultsLimit {
return Results, Values
}
}
return Results, Values
}
// SortedQuery returns the strings which contain the query sorted
// The function sorter produces a value to sort by (largest first)
// Input:
// Word: The substring to search for.
// ResultsLimit: Limit the results to some number of values. Set to -1 for no limit
// Sorter: Takes (Result, Value, Length, Index (where Query begins in Result)) (string, []byte, int, int)
// and produces a value (float64) to sort by (largest first).
func (IS *InvertedSuffix) SortedQuery(Word string, ResultsLimit int, Sorter func(string, interface{}, int, int) float64) ([]string, []interface{}, []float64) {
Query := IS.Converter(Word)
if ResultsLimit == 0 {
return []string{}, []interface{}{}, []float64{}
}
if ResultsLimit < 0 {
ResultsLimit = 0
}
Results := make([]string, 0, ResultsLimit)
Values := make([]interface{}, 0, ResultsLimit)
Scores := make([]float64, 0, ResultsLimit)
if ResultsLimit == 0 {
ResultsLimit = -1
}
low, high := IS.Search(Query)
a := 0
used := make(map[int]float64, 0)
for k := low; k < high; k++ {
x := IS.WordIndex[k]
w := IS.Results[x]
v := IS.Values[x]
s := Sorter(w, v, len(IS.Words[x]), IS.SuffixIndex[k])
if ps, ok := used[x]; ok && ps >= s {
continue
}
used[x] = s
i := 0
j := a
for i < j {
h := (i + j) >> 1
if Scores[h] > s {
i = h + 1
} else {
j = h
}
}
if a == ResultsLimit {
if i < a {
copy(Results[i+1:], Results[i:a-1])
Results[i] = w
copy(Values[i+1:], Values[i:a-1])
Values[i] = v
copy(Scores[i+1:], Scores[i:a-1])
Scores[i] = s
}
} else {
a++
if i < a {
Results = append(Results, "")
copy(Results[i+1:], Results[i:])
Results[i] = w
Values = append(Values, nil)
copy(Values[i+1:], Values[i:a-1])
Values[i] = v
Scores = append(Scores, 0.0)
copy(Scores[i+1:], Scores[i:a-1])
Scores[i] = s
} else {
Results = append(Results, w)
Values = append(Values, v)
Scores = append(Scores, s)
}
}
}
return Results, Values, Scores
}
// ErrorCorrectingQuery returns the strings which contain the query
// Unsorted, I think it's partially sorted alphabetically
// Will search for all substrings defined by ErrorCorrection
// if no results are found on the initial query
// Input:
// Query: The substring to search for.
// ResultsLimit: Limit the results so you don't return your whole dictionary by accident. Set to -1 for no limit
// ErrorCorrection: Returns a list of alternate queries
func (IS *InvertedSuffix) ErrorCorrectingQuery(Word string, ResultsLimit int, ErrorCorrection func([]byte) [][]byte) ([]string, []interface{}) {
Query := IS.Converter(Word)
if ResultsLimit == 0 {
return []string{}, []interface{}{}
}
if ResultsLimit < 0 {
ResultsLimit = 0
}
Results := make([]string, 0, ResultsLimit)
Values := make([]interface{}, 0, ResultsLimit)
low, high := IS.Search(Query)
a := 0
used := make(map[int]bool, 0)
for k := low; k < high; k++ {
x := IS.WordIndex[k]
if _, ok := used[x]; ok {
continue
}
used[x] = true
Results = append(Results, IS.Results[x])
Values = append(Values, IS.Values[x])
a++
if a == ResultsLimit {
return Results, Values
}
}
if a != ResultsLimit {
for _, q := range ErrorCorrection(Query) {
low, high := IS.Search(q)
for k := low; k < high; k++ {
x := IS.WordIndex[k]
if _, ok := used[x]; ok {
continue
}
used[x] = true
Results = append(Results, IS.Results[x])
Values = append(Values, IS.Values[x])
a++
if a == ResultsLimit {
return Results, Values
}
}
}
}
return Results, Values
}
// SortedErrorCorrectingQuery returns the strings which contain the query
// Sorted. The function sorter produces a value to sort by (largest first)
// Will search for all substrings defined by ErrorCorrection
// if no results are found on the initial query
// Input:
// Query: The substring to search for.
// ResultsLimit: Limit the results so you don't return your whole dictionary by accident. Set to -1 for no limit
// ErrorCorrection: Returns a list of alternate queries
// Sorter: Takes (Result, Value, Length, Index (where Query begins in Result))
// (string, []byte, int, int), and produces a value (float64) to sort by (largest first).
func (IS *InvertedSuffix) SortedErrorCorrectingQuery(Word string, ResultsLimit int, ErrorCorrection func([]byte) [][]byte, Sorter func(string, interface{}, int, int) float64) ([]string, []interface{}, []float64) {
Query := IS.Converter(Word)
if ResultsLimit == 0 {
return []string{}, []interface{}{}, []float64{}
}
if ResultsLimit < 0 {
ResultsLimit = 0
}
Results := make([]string, 0, ResultsLimit)
Values := make([]interface{}, 0, ResultsLimit)
Scores := make([]float64, 0, ResultsLimit)
if ResultsLimit == 0 {
ResultsLimit = -1
}
low, high := IS.Search(Query)
a := 0
used := make(map[int]float64, 0)
for k := low; k < high; k++ {
x := IS.WordIndex[k]
w := IS.Results[x]
v := IS.Values[x]
s := Sorter(w, v, len(IS.Words[x]), IS.SuffixIndex[k])
if ps, ok := used[x]; ok && ps >= s {
continue
}
used[x] = s
i := 0
j := a
for i < j {
h := (i + j) >> 1
if Scores[h] > s {
i = h + 1
} else {
j = h
}
}
if a == ResultsLimit {
if i < a {
copy(Results[i+1:], Results[i:a-1])
Results[i] = w
copy(Values[i+1:], Values[i:a-1])
Values[i] = v
copy(Scores[i+1:], Scores[i:a-1])
Scores[i] = s
}
} else {
a++
if i < a {
Results = append(Results, "")
copy(Results[i+1:], Results[i:])
Results[i] = w
Values = append(Values, nil)
copy(Values[i+1:], Values[i:a-1])
Values[i] = v
Scores = append(Scores, 0.0)
copy(Scores[i+1:], Scores[i:a-1])
Scores[i] = s
} else {
Results = append(Results, w)
Values = append(Values, v)
Scores = append(Scores, s)
}
}
}
if a == 0 {
for _, q := range ErrorCorrection(Query) {
low, high := IS.Search(q)
for k := low; k < high; k++ {
x := IS.WordIndex[k]
w := IS.Results[x]
v := IS.Values[x]
s := Sorter(w, v, len(IS.Words[x]), IS.SuffixIndex[k])
if ps, ok := used[x]; ok && ps >= s {
continue
}
used[x] = s
i := 0
j := a
for i < j {
h := (i + j) >> 1
if Scores[h] > s {
i = h + 1
} else {
j = h
}
}
if a == ResultsLimit {
if i < a {
copy(Results[i+1:], Results[i:a-1])
Results[i] = w
copy(Values[i+1:], Values[i:a-1])
Values[i] = v
copy(Scores[i+1:], Scores[i:a-1])
Scores[i] = s
}
} else {
a++
if i < a {
Results = append(Results, "")
copy(Results[i+1:], Results[i:])
Results[i] = w
Values = append(Values, nil)
copy(Values[i+1:], Values[i:a-1])
Values[i] = v
Scores = append(Scores, 0.0)
copy(Scores[i+1:], Scores[i:a-1])
Scores[i] = s
} else {
Results = append(Results, w)
Values = append(Values, v)
Scores = append(Scores, s)
}
}
}
}
}
return Results, Values, Scores
}