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
|
// Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
#include "juniper_query_adapter.h"
#include "i_query_term_filter.h"
#include "juniper_dfw_query_item.h"
#include "juniper_dfw_term_visitor.h"
#include <vespa/searchlib/fef/properties.h>
#include <vespa/searchlib/parsequery/stackdumpiterator.h>
#include <vespa/searchlib/queryeval/split_float.h>
#include <vespa/searchlib/query/query_normalization.h>
namespace search::docsummary {
JuniperQueryAdapter::JuniperQueryAdapter(const QueryNormalization * normalization, const IQueryTermFilter *query_term_filter,
vespalib::stringref buf, const search::fef::Properties & highlightTerms)
: _query_normalization(normalization),
_query_term_filter(query_term_filter),
_buf(buf),
_highlightTerms(highlightTerms)
{
}
JuniperQueryAdapter::~JuniperQueryAdapter() = default;
// TODO: put this functionality into the stack dump iterator
bool
JuniperQueryAdapter::skipItem(search::SimpleQueryStackDumpIterator *iterator) const
{
uint32_t skipCount = iterator->getArity();
while (skipCount > 0) {
if (!iterator->next()) {
return false; // stack too small
}
skipCount = skipCount - 1 + iterator->getArity();
}
return true;
}
bool
JuniperQueryAdapter::Traverse(juniper::IQueryVisitor *v) const
{
bool rc = true;
search::SimpleQueryStackDumpIterator iterator(_buf);
JuniperDFWQueryItem item(&iterator);
if (_highlightTerms.numKeys() > 0) {
v->VisitAND(&item, 2);
}
while (rc && iterator.next()) {
bool isSpecialToken = iterator.hasSpecialTokenFlag();
bool prefix_like = false;
switch (iterator.getType()) {
case search::ParseItem::ITEM_OR:
case search::ParseItem::ITEM_WEAK_AND:
case search::ParseItem::ITEM_EQUIV:
case search::ParseItem::ITEM_WORD_ALTERNATIVES:
if (!v->VisitOR(&item, iterator.getArity()))
rc = skipItem(&iterator);
break;
case search::ParseItem::ITEM_AND:
if (!v->VisitAND(&item, iterator.getArity()))
rc = skipItem(&iterator);
break;
case search::ParseItem::ITEM_NOT:
if (!v->VisitANDNOT(&item, iterator.getArity()))
rc = skipItem(&iterator);
break;
case search::ParseItem::ITEM_RANK:
if (!v->VisitRANK(&item, iterator.getArity()))
rc = skipItem(&iterator);
break;
case search::ParseItem::ITEM_PREFIXTERM:
case search::ParseItem::ITEM_SUBSTRINGTERM:
prefix_like = true;
[[fallthrough]];
case search::ParseItem::ITEM_TERM:
case search::ParseItem::ITEM_EXACTSTRINGTERM:
case search::ParseItem::ITEM_PURE_WEIGHTED_STRING:
{
vespalib::string term = iterator.getTerm();
if (_query_normalization) {
vespalib::string index = iterator.getIndexName();
if (index.empty()) {
index = SimpleQueryStackDumpIterator::DEFAULT_INDEX;
}
Normalizing normalization = _query_normalization->normalizing_mode(index);
TermType termType = ParseItem::toTermType(iterator.getType());
v->visitKeyword(&item, QueryNormalization::optional_fold(term, termType, normalization),
prefix_like, isSpecialToken);
} else {
v->visitKeyword(&item, term, prefix_like, isSpecialToken);
}
}
break;
case search::ParseItem::ITEM_NUMTERM:
{
vespalib::string term = iterator.getTerm();
queryeval::SplitFloat splitter(term);
if (splitter.parts() > 1) {
if (v->VisitPHRASE(&item, splitter.parts())) {
for (size_t i = 0; i < splitter.parts(); ++i) {
v->visitKeyword(&item, splitter.getPart(i), false, false);
}
}
} else if (splitter.parts() == 1) {
v->visitKeyword(&item, splitter.getPart(0), false, false);
} else {
v->visitKeyword(&item, term, false, true);
}
}
break;
case search::ParseItem::ITEM_PHRASE:
if (!v->VisitPHRASE(&item, iterator.getArity()))
rc = skipItem(&iterator);
break;
case search::ParseItem::ITEM_ANY:
if (!v->VisitANY(&item, iterator.getArity()))
rc = skipItem(&iterator);
break;
case search::ParseItem::ITEM_NEAR:
if (!v->VisitNEAR(&item, iterator.getArity(),iterator.getNearDistance()))
rc = skipItem(&iterator);
break;
case search::ParseItem::ITEM_ONEAR:
if (!v->VisitWITHIN(&item, iterator.getArity(),iterator.getNearDistance()))
rc = skipItem(&iterator);
break;
case search::ParseItem::ITEM_TRUE:
case search::ParseItem::ITEM_FALSE:
if (!v->VisitOther(&item, iterator.getArity())) {
rc = skipItem(&iterator);
}
break;
// Unhandled items are just ignored by juniper
case search::ParseItem::ITEM_WAND:
case search::ParseItem::ITEM_WEIGHTED_SET:
case search::ParseItem::ITEM_DOT_PRODUCT:
case search::ParseItem::ITEM_PURE_WEIGHTED_LONG:
case search::ParseItem::ITEM_SUFFIXTERM:
case search::ParseItem::ITEM_REGEXP:
case search::ParseItem::ITEM_PREDICATE_QUERY:
case search::ParseItem::ITEM_SAME_ELEMENT:
case search::ParseItem::ITEM_NEAREST_NEIGHBOR:
case search::ParseItem::ITEM_GEO_LOCATION_TERM:
case search::ParseItem::ITEM_FUZZY:
case search::ParseItem::ITEM_STRING_IN:
case search::ParseItem::ITEM_NUMERIC_IN:
if (!v->VisitOther(&item, iterator.getArity())) {
rc = skipItem(&iterator);
}
break;
default:
rc = false;
}
}
if (_highlightTerms.numKeys() > 1) {
v->VisitAND(&item, _highlightTerms.numKeys());
}
JuniperDFWTermVisitor tv(v);
_highlightTerms.visitProperties(tv);
return rc;
}
bool
JuniperQueryAdapter::UsefulIndex(const juniper::QueryItem* item) const
{
if (_query_term_filter == nullptr) {
return true;
}
auto index = item->get_index();
return _query_term_filter->use_view(index);
}
}
|
