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
|
// Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
#include <vespa/eval/eval/test/reference_evaluation.h>
#include <vespa/eval/eval/function.h>
#include <vespa/eval/eval/tensor_spec.h>
#include <vespa/eval/eval/test/eval_spec.h>
#include <vespa/vespalib/gtest/gtest.h>
using namespace vespalib;
using namespace vespalib::eval;
using namespace vespalib::eval::test;
//-----------------------------------------------------------------------------
TensorSpec ref_eval(const Function &fun, const std::vector<TensorSpec> ¶ms) {
return ReferenceEvaluation::eval(fun, params);
}
TensorSpec ref_eval(std::shared_ptr<const Function> fun, const std::vector<TensorSpec> ¶ms) {
return ref_eval(*fun, params);
}
TensorSpec ref_eval(const vespalib::string &expr, const std::vector<TensorSpec> ¶ms) {
return ref_eval(*Function::parse(expr), params);
}
TensorSpec make_val(const vespalib::string &expr) {
return ref_eval(*Function::parse(expr), {});
}
//-----------------------------------------------------------------------------
struct MyEvalTest : EvalSpec::EvalTest {
size_t pass_cnt = 0;
size_t fail_cnt = 0;
bool print_pass = false;
bool print_fail = false;
void next_expression(const std::vector<vespalib::string> &,
const vespalib::string &) override {}
void handle_case(const std::vector<vespalib::string> ¶m_names,
const std::vector<double> ¶m_values,
const vespalib::string &expression,
double expected_result) override
{
auto function = Function::parse(param_names, expression);
ASSERT_FALSE(function->has_error());
std::vector<TensorSpec> params;
for (double param: param_values) {
params.push_back(TensorSpec("double").add({}, param));
}
auto eval_result = ref_eval(function, params);
ASSERT_EQ(eval_result.type(), "double");
double result = eval_result.as_double();
if (is_same(expected_result, result)) {
print_pass && fprintf(stderr, "verifying: %s -> %g ... PASS\n",
as_string(param_names, param_values, expression).c_str(),
expected_result);
++pass_cnt;
} else {
print_fail && fprintf(stderr, "verifying: %s -> %g ... FAIL: got %g\n",
as_string(param_names, param_values, expression).c_str(),
expected_result, result);
++fail_cnt;
}
}
};
//-----------------------------------------------------------------------------
TEST(ReferenceEvaluationTest, reference_evaluation_passes_all_eval_spec_tests) {
MyEvalTest test;
EvalSpec spec;
test.print_fail = true;
spec.add_all_cases();
spec.each_case(test);
EXPECT_GT(test.pass_cnt, 1000);
EXPECT_EQ(test.fail_cnt, 0);
}
//-----------------------------------------------------------------------------
// 'make_val' will be used to generate tensor specs for inputs and
// expected outputs for other tests. In the production evaluation
// pipeline this kind of tensor create will be converted to a constant
// value when converting the Function to a TensorFunction. With the
// reference evaluation the Function is evaluated directly with no
// constant folding.
TEST(ReferenceEvaluationTest, constant_create_expression_works) {
auto expect = TensorSpec("tensor(x{},y[2])")
.add({{"x", "a"}, {"y", 0}}, 1.0)
.add({{"x", "a"}, {"y", 1}}, 2.0);
auto result = make_val("tensor(x{},y[2]):{a:[1,2]}");
EXPECT_EQ(result, expect);
}
//-----------------------------------------------------------------------------
TEST(ReferenceEvaluationTest, parameter_expression_works) {
auto a = make_val("tensor(x[2]):[1,2]");
auto b = make_val("tensor(x[2]):[3,4]");
auto fun_a = Function::parse({"a", "b"}, "a");
auto fun_b = Function::parse({"a", "b"}, "b");
EXPECT_EQ(ref_eval(fun_a, {a, b}), a);
EXPECT_EQ(ref_eval(fun_b, {a, b}), b);
}
TEST(ReferenceEvaluationTest, parameter_expression_will_pad_with_zero) {
auto a = TensorSpec("tensor(x[3])")
.add({{"x", 1}}, 5.0);
auto expect = make_val("tensor(x[3]):[0,5,0]");
EXPECT_EQ(ref_eval("a", {a}), expect);
}
TEST(ReferenceEvaluationTest, reduce_expression_works) {
auto a = make_val("tensor(x[2],y[2]):[[1,2],[3,4]]");
auto expect = make_val("tensor(x[2]):[3,7]");
EXPECT_EQ(ref_eval("reduce(a,sum,y)", {a}), expect);
}
TEST(ReferenceEvaluationTest, reduce_can_expand) {
auto a = make_val("tensor(x{},y[2]):{}");
auto expect = make_val("tensor(y[2]):[0,0]");
EXPECT_EQ(ref_eval("reduce(a,sum,x)", {a}), expect);
}
TEST(ReferenceEvaluationTest, map_expression_works) {
auto a = make_val("tensor(x[2]):[1,10]");
auto expect = make_val("tensor(x[2]):[5,23]");
EXPECT_EQ(ref_eval("map(a,f(x)(x*2+3))", {a}), expect);
}
TEST(ReferenceEvaluationTest, join_expression_works) {
auto a = make_val("tensor(x[2]):[1,2]");
auto b = make_val("tensor(y[2]):[3,4]");
auto expect = make_val("tensor(x[2],y[2]):[[4,5],[5,6]]");
EXPECT_EQ(ref_eval("join(a,b,f(x,y)(x+y))", {a, b}), expect);
}
TEST(ReferenceEvaluationTest, merge_expression_works) {
auto a = make_val("tensor(x{}):{a:1,b:2,c:3}");
auto b = make_val("tensor(x{}):{c:3,d:4}");
auto expect = make_val("tensor(x{}):{a:1,b:2,c:6,d:4}");
EXPECT_EQ(ref_eval("merge(a,b,f(x,y)(x+y))", {a, b}), expect);
}
TEST(ReferenceEvaluationTest, concat_expression_works) {
auto a = make_val("tensor(x[2]):[1,2]");
auto b = make_val("tensor(x[2]):[3,4]");
auto expect = make_val("tensor(x[4]):[1,2,3,4]");
EXPECT_EQ(ref_eval("concat(a,b,x)", {a, b}), expect);
}
TEST(ReferenceEvaluationTest, cell_cast_expression_works) {
auto a = make_val("tensor<double>(x[4]):[1,2,3,4]");
auto expect = make_val("tensor<float>(x[4]):[1,2,3,4]");
EXPECT_EQ(ref_eval("cell_cast(a,float)", {a}), expect);
}
TEST(ReferenceEvaluationTest, rename_expression_works) {
auto a = make_val("tensor(x[2]):[1,2]");
auto expect = make_val("tensor(y[2]):[1,2]");
EXPECT_EQ(ref_eval("rename(a,x,y)", {a}), expect);
}
TEST(ReferenceEvaluationTest, create_expression_works) {
auto a = make_val("5");
auto expect = make_val("tensor(x[3]):[5,10,15]");
EXPECT_EQ(ref_eval("tensor(x[3]):[a,2*a,3*a]", {a}), expect);
}
TEST(ReferenceEvaluationTest, tensor_create_will_pad_with_zero) {
auto a = make_val("5");
auto expect = make_val("tensor(x[3]):[0,5,0]");
EXPECT_EQ(ref_eval("tensor(x[3]):{{x:1}:a}", {a}), expect);
}
TEST(ReferenceEvaluationTest, lambda_expression_works) {
auto a = make_val("5");
auto expect = make_val("tensor(x[3]):[5,10,15]");
EXPECT_EQ(ref_eval("tensor(x[3])((x+1)*a)", {a}), expect);
}
TEST(ReferenceEvaluationTest, peek_expression_works) {
auto a = make_val("tensor(x{},y[2]):{a:[3,7]}");
auto b = make_val("1");
auto expect = make_val("7");
EXPECT_EQ(ref_eval("a{x:a,y:(b)}", {a, b}), expect);
}
TEST(ReferenceEvaluationTest, verbatim_peek_of_dense_dimension_works) {
auto a = make_val("tensor(x[4]):[1,2,3,4]");
auto expect = make_val("3");
EXPECT_EQ(ref_eval("a{x:2}", {a}), expect);
}
TEST(ReferenceEvaluationTest, out_of_bounds_peek_works) {
auto a = make_val("tensor(x[4]):[1,2,3,4]");
auto b = make_val("4");
auto expect = make_val("0");
EXPECT_EQ(ref_eval("a{x:(b)}", {a, b}), expect);
}
//-----------------------------------------------------------------------------
TEST(ReferenceEvaluationTest, compound_expression_works) {
auto a = make_val("10");
auto b = make_val("20");
auto expect = make_val("20");
EXPECT_EQ(ref_eval("reduce(concat(a,b,x)+5,avg,x)", {a, b}), expect);
}
//-----------------------------------------------------------------------------
GTEST_MAIN_RUN_ALL_TESTS()
|
