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
|
// Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
#include "utf8stringfieldsearcherbase.h"
#include "tokenizereader.h"
#include <cassert>
using search::streaming::QueryTerm;
using search::streaming::QueryTermList;
using search::byte;
namespace vsm {
size_t
UTF8StringFieldSearcherBase::matchTermRegular(const FieldRef & f, QueryTerm & qt)
{
termcount_t words(0);
const cmptype_t * term;
termsize_t tsz = qt.term(term);
if ( f.size() >= _buf->size()) {
_buf->reserve(f.size() + 1);
}
cmptype_t * fn = _buf->data();
TokenizeReader reader(reinterpret_cast<const byte *> (f.data()), f.size(), fn);
while ( reader.hasNext() ) {
size_t fl = reader.tokenize(normalize_mode());
if ((tsz <= fl) && (prefix() || qt.isPrefix() || (tsz == fl))) {
const cmptype_t *tt=term, *et=term+tsz;
for (const cmptype_t *fnt=fn; (tt < et) && (*tt == *fnt); tt++, fnt++);
if (tt == et) {
addHit(qt, words);
}
}
words++;
}
return words;
}
size_t
UTF8StringFieldSearcherBase::matchTermExact(const FieldRef & f, QueryTerm & qt)
{
const cmptype_t * term;
termsize_t tsz = qt.term(term);
const cmptype_t * eterm = term+tsz;
if ( f.size() >= _buf->size()) {
_buf->reserve(f.size() + 1);
}
cmptype_t * fn = _buf->data();
if (tsz <= f.size()) {
bool equal(true);
Normalizing norm_mode = normalize_mode();
TokenizeReader reader(reinterpret_cast<const byte *> (f.data()), f.size(), fn);
while (equal && reader.hasNext() && (term < eterm)) {
reader.normalize(reader.next(), norm_mode);
size_t len = reader.complete();
for (size_t i(0); i < len; i++) {
equal = (term[i] == fn[i]);
}
term += len;
}
if (equal && (term == eterm) && (qt.isPrefix() || ! reader.hasNext())) {
addHit(qt,0);
}
}
return 1;
}
size_t
UTF8StringFieldSearcherBase::matchTermSubstring(const FieldRef & f, QueryTerm & qt)
{
if (qt.termLen() == 0) { return 0; }
const byte * n = reinterpret_cast<const byte *> (f.data());
const cmptype_t * term;
termsize_t tsz = qt.term(term);
if ( f.size() >= _buf->size()) {
_buf->reserve(f.size() + 1);
}
cmptype_t * fntemp = &(*_buf.get())[0];
BufferWrapper wrapper(fntemp);
size_t fl = skipSeparators(n, f.size(), wrapper);
const cmptype_t * fn(fntemp);
const cmptype_t * fe = fn + fl;
const cmptype_t * fre = fe - tsz;
termcount_t words(0);
for(words = 0; fn <= fre; ) {
const cmptype_t *tt=term, *et=term+tsz, *fnt=fn;
for (; (tt < et) && (*tt == *fnt); tt++, fnt++);
if (tt == et) {
fn = fnt;
addHit(qt, words);
} else {
if ( ! Fast_UnicodeUtil::IsWordChar(*fn++) ) {
words++;
for(; (fn < fre) && ! Fast_UnicodeUtil::IsWordChar(*fn) ; fn++ );
}
}
}
return words + 1; // we must also count the last word
}
size_t
UTF8StringFieldSearcherBase::matchTermSuffix(const FieldRef & f, QueryTerm & qt)
{
termcount_t words = 0;
const cmptype_t * term;
termsize_t tsz = qt.term(term);
if (f.size() >= _buf->size()) {
_buf->reserve(f.size() + 1);
}
cmptype_t * dstbuf = _buf->data();
TokenizeReader reader(reinterpret_cast<const byte *> (f.data()), f.size(), dstbuf);
while ( reader.hasNext() ) {
size_t tokenlen = reader.tokenize(normalize_mode());
if (matchTermSuffix(term, tsz, dstbuf, tokenlen)) {
addHit(qt, words);
}
words++;
}
return words;
}
UTF8StringFieldSearcherBase::UTF8StringFieldSearcherBase(FieldIdT fId) :
StrChrFieldSearcher(fId)
{
}
UTF8StringFieldSearcherBase::~UTF8StringFieldSearcherBase() = default;
void
UTF8StringFieldSearcherBase::prepare(search::streaming::QueryTermList& qtl,
const SharedSearcherBuf& buf,
const vsm::FieldPathMapT& field_paths,
search::fef::IQueryEnvironment& query_env)
{
StrChrFieldSearcher::prepare(qtl, buf, field_paths, query_env);
_buf = buf;
}
bool
UTF8StringFieldSearcherBase::matchTermSuffix(const cmptype_t * term, size_t termlen,
const cmptype_t * word, size_t wordlen)
{
if ((termlen <= wordlen)) {
const cmptype_t * titr = term + termlen - 1;
const cmptype_t * witr = word + wordlen - 1;
bool hit = true;
// traverse the term and the word back to front
for (; titr >= term; --titr, --witr) {
if (*titr != *witr) {
hit = false;
break;
}
}
return hit;
}
return false;
}
bool
UTF8StringFieldSearcherBase::isSeparatorCharacter(ucs4_t c)
{
return ((c < 0x20) && (c != '\n') && (c != '\t'));
}
template <typename T>
size_t
UTF8StringFieldSearcherBase::skipSeparators(const search::byte * p, size_t sz, T & dstbuf) {
const search::byte * e(p+sz);
const search::byte * b(p);
for(; p < e; ) {
ucs4_t c(*p);
const search::byte * oldP(p);
if (c < 128) {
p++;
if (!isSeparatorCharacter(c)) {
dstbuf.onCharacter(Fast_NormalizeWordFolder::lowercase_and_fold_ascii(c), (oldP - b));
}
} else {
c = Fast_UnicodeUtil::GetUTF8CharNonAscii(p);
const char *repl = Fast_NormalizeWordFolder::ReplacementString(c);
if (repl != nullptr) {
size_t repllen = strlen(repl);
if (repllen > 0) {
ucs4_t * buf = dstbuf.getBuf();
ucs4_t * newBuf = Fast_UnicodeUtil::ucs4copy(buf, repl);
if (dstbuf.hasOffsets()) {
for (; buf < newBuf; ++buf) {
dstbuf.incBuf(1);
dstbuf.onOffset(oldP - b);
}
} else {
dstbuf.incBuf(newBuf - buf);
}
}
} else {
c = Fast_NormalizeWordFolder::lowercase_and_fold(c);
dstbuf.onCharacter(c, (oldP - b));
}
if (c == Fast_UnicodeUtil::_BadUTF8Char) {
_badUtf8Count++;
}
}
}
assert(dstbuf.valid());
return dstbuf.size();
}
template unsigned long UTF8StringFieldSearcherBase::skipSeparators<UTF8StringFieldSearcherBase::BufferWrapper>(unsigned char const*, unsigned long, UTF8StringFieldSearcherBase::BufferWrapper&);
template unsigned long UTF8StringFieldSearcherBase::skipSeparators<UTF8StringFieldSearcherBase::OffsetWrapper>(unsigned char const*, unsigned long, UTF8StringFieldSearcherBase::OffsetWrapper&);
}
|
