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
|
// Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
#include <vespa/vespalib/testkit/test_kit.h>
#include <vespa/eval/eval/value_type.h>
#include <vespa/searchlib/fef/feature_type.h>
#include <vespa/searchlib/fef/featurenameparser.h>
#include <vespa/searchlib/features/rankingexpressionfeature.h>
#include <vespa/searchlib/fef/test/dummy_dependency_handler.h>
#include <vespa/searchlib/fef/test/indexenvironment.h>
#include <vespa/searchlib/fef/test/queryenvironment.h>
using namespace search::features;
using namespace search::features::rankingexpression;
using namespace search::fef::test;
using namespace search::fef;
using namespace vespalib::eval;
using TypeMap = std::map<vespalib::string,vespalib::string>;
struct DummyExecutor : FeatureExecutor {
void execute(uint32_t) override {}
};
struct DummyExpression : IntrinsicExpression {
FeatureType type;
DummyExpression(const FeatureType &type_in) : type(type_in) {}
vespalib::string describe_self() const override { return "dummy"; }
const FeatureType &result_type() const override { return type; }
void prepare_shared_state(const QueryEnv &, IObjectStore &) const override {
}
FeatureExecutor &create_executor(const QueryEnv &, vespalib::Stash &stash) const override {
return stash.create<DummyExecutor>();
}
};
struct DummyReplacer : ExpressionReplacer {
vespalib::string trigger;
FeatureType type;
DummyReplacer(const vespalib::string trigger_in, const FeatureType &type_in)
: trigger(trigger_in),
type(type_in)
{}
IntrinsicExpression::UP maybe_replace(const vespalib::eval::Function &function,
const search::fef::IIndexEnvironment &) const override
{
for (size_t i = 0; i < function.num_params(); ++i) {
if (function.param_name(i) == trigger) {
return std::make_unique<DummyExpression>(type);
}
}
return IntrinsicExpression::UP(nullptr);
}
};
ExpressionReplacer::SP make_replacer() {
auto replacer = std::make_shared<ListExpressionReplacer>();
replacer->add(std::make_unique<NullExpressionReplacer>());
replacer->add(std::make_unique<DummyReplacer>("foo", FeatureType::number()));
replacer->add(std::make_unique<DummyReplacer>("bar", FeatureType::object(ValueType::from_spec("tensor(x[5])"))));
return replacer;
}
struct SetupResult {
vespalib::Stash stash;
IndexEnvironment index_env;
QueryEnvironment query_env;
RankingExpressionBlueprint rank;
DummyDependencyHandler deps;
bool setup_ok;
SetupResult(const TypeMap &object_inputs, const vespalib::string &expression,
const vespalib::string &expression_name = "");
~SetupResult();
};
SetupResult::SetupResult(const TypeMap &object_inputs,
const vespalib::string &expression,
const vespalib::string &expression_name)
: stash(), index_env(), query_env(&index_env), rank(make_replacer()), deps(rank), setup_ok(false)
{
rank.setName("self");
for (const auto &input: object_inputs) {
deps.define_object_input(input.first, ValueType::from_spec(input.second));
}
std::vector<vespalib::string> params;
if (expression_name.empty()) {
index_env.getProperties().add("self.rankingScript", expression);
} else {
index_env.addRankingExpression(expression_name, expression);
index_env.getProperties().add("self.expressionName", expression_name);
}
Blueprint &bp = rank;
setup_ok = bp.setup(index_env, params);
EXPECT_TRUE(!deps.accept_type_mismatch);
}
SetupResult::~SetupResult() = default;
void verify_output_type(const TypeMap &object_inputs,
const vespalib::string &expression, const FeatureType &expect,
const vespalib::string &expression_name = "")
{
SetupResult result(object_inputs, expression, expression_name);
EXPECT_TRUE(result.setup_ok);
EXPECT_EQUAL(1u, result.deps.output.size());
ASSERT_EQUAL(1u, result.deps.output_type.size());
if (expect.is_object()) {
EXPECT_EQUAL(expect.type(), result.deps.output_type[0].type());
} else {
EXPECT_TRUE(!result.deps.output_type[0].is_object());
}
}
void verify_setup_fail(const TypeMap &object_inputs,
const vespalib::string &expression)
{
SetupResult result(object_inputs, expression);
EXPECT_TRUE(!result.setup_ok);
EXPECT_EQUAL(0u, result.deps.output.size());
}
void verify_input_count(const vespalib::string &expression, size_t expect) {
SetupResult result({}, expression);
EXPECT_TRUE(result.setup_ok);
EXPECT_EQUAL(result.deps.input.size(), expect);
}
TEST("require that expression with only number inputs produce number output (compiled)") {
TEST_DO(verify_output_type({}, "a*b", FeatureType::number()));
}
TEST("require that expression with object input produces object output (interpreted)") {
TEST_DO(verify_output_type({{"b", "tensor(x{})"}}, "a*b", FeatureType::object(ValueType::from_spec("tensor(x{})"))));
}
TEST("require that scalar expressions are auto-unboxed (interpreted)") {
TEST_DO(verify_output_type({{"b", "tensor(x{})"}}, "reduce(a*b,sum)", FeatureType::number()));
}
TEST("require that ranking expression can resolve to concrete complex type") {
TEST_DO(verify_output_type({{"a", "tensor(x{},y{})"}, {"b", "tensor(y{},z{})"}}, "a*b",
FeatureType::object(ValueType::from_spec("tensor(x{},y{},z{})"))));
}
TEST("require that ranking expression can be external") {
TEST_DO(verify_output_type({}, "a*b", FeatureType::number(), "my_expr"));
TEST_DO(verify_output_type({{"b", "double"}}, "a*b", FeatureType::number(), "my_expr"));
TEST_DO(verify_output_type({{"a", "tensor(x{},y{})"}, {"b", "tensor(y{},z{})"}}, "a*b",
FeatureType::object(ValueType::from_spec("tensor(x{},y{},z{})")), "my_expr"));
}
TEST("require that setup fails for incompatible types") {
TEST_DO(verify_setup_fail({{"a", "tensor(x{},y{})"}, {"b", "tensor(y[10],z{})"}}, "a*b"));
}
TEST("require that replaced expressions have no inputs") {
TEST_DO(verify_input_count("a*b*c", 3u));
TEST_DO(verify_input_count("foo*b*c", 0u));
TEST_DO(verify_input_count("a*b*bar", 0u));
TEST_DO(verify_input_count("foo*b*bar", 0u));
}
TEST("require that replaced expressions override result type") {
TEST_DO(verify_output_type({{"b", "tensor(z{})"}}, "a*b*c",
FeatureType::object(ValueType::from_spec("tensor(z{})"))));
TEST_DO(verify_output_type({{"b", "tensor(z{})"}}, "foo*b*c",
FeatureType::number()));
TEST_DO(verify_output_type({{"b", "tensor(z{})"}}, "a*b*bar",
FeatureType::object(ValueType::from_spec("tensor(x[5])"))));
TEST_DO(verify_output_type({{"b", "tensor(z{})"}}, "foo*b*bar",
FeatureType::number()));
}
TEST_F("require that replaced expressions create the appropriate executor", SetupResult({}, "foo")) {
EXPECT_TRUE(f1.setup_ok);
FeatureExecutor &executor = f1.rank.createExecutor(f1.query_env, f1.stash);
EXPECT_TRUE(dynamic_cast<DummyExecutor*>(&executor) != nullptr);
}
TEST_MAIN() { TEST_RUN_ALL(); }
|
