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
|
# ---------------------------------------------------------
# Copyright (c) Microsoft Corporation. All rights reserved.
# ---------------------------------------------------------
import pytest
from unittest.mock import MagicMock, patch
from azure.core.credentials import TokenCredential
from azure.ai.evaluation import AzureOpenAIModelConfiguration
from azure.ai.evaluation._evaluators._relevance import RelevanceEvaluator
@pytest.fixture(scope="session")
def test_model_config() -> AzureOpenAIModelConfiguration:
"""Mock model configuration for tests."""
return AzureOpenAIModelConfiguration(
azure_endpoint="https://test.openai.azure.com",
api_key="test-key",
api_version="2024-02-15-preview",
azure_deployment="test-deployment",
)
@pytest.fixture(scope="function")
def mock_credential():
"""Mock credential for tests."""
return MagicMock(spec=TokenCredential)
@pytest.mark.unittest
class TestConversationThresholdBehavior:
"""Test threshold behavior in conversation evaluators."""
@patch("azure.ai.evaluation._evaluators._relevance._relevance." "RelevanceEvaluator.__call__")
def test_relevance_evaluator_with_conversation(self, mock_call, test_model_config):
"""Test relevance evaluator with conversation input."""
mock_result = {"relevance": 4.0, "relevance_result": "PASS", "evaluation_per_turn": {"relevance": [4.0]}}
mock_call.return_value = mock_result
evaluator = RelevanceEvaluator(model_config=test_model_config, threshold=3)
conversation = {
"messages": [
{"role": "user", "content": "What is the weather like?"},
{"role": "assistant", "content": "It's sunny today."},
]
}
result = evaluator(conversation=conversation)
assert result["relevance"] > 0.0
assert "evaluation_per_turn" in result
@pytest.mark.unittest
class TestMultipleEvaluatorThresholds:
"""Test multiple evaluators with different thresholds."""
@patch("azure.ai.evaluation._evaluators._relevance._relevance." "RelevanceEvaluator.__call__")
def test_evaluators_with_different_thresholds(self, mock_call, test_model_config):
"""Test that evaluators can have different thresholds."""
mock_result = {"relevance": 4.0, "relevance_result": "PASS", "evaluation_per_turn": {"relevance": [4.0]}}
mock_call.return_value = mock_result
evaluator1 = RelevanceEvaluator(model_config=test_model_config, threshold=2)
evaluator2 = RelevanceEvaluator(model_config=test_model_config, threshold=4)
conversation = {
"messages": [
{"role": "user", "content": "What is the time?"},
{"role": "assistant", "content": "It is 3pm."},
]
}
result1 = evaluator1(conversation=conversation)
result2 = evaluator2(conversation=conversation)
assert evaluator1._threshold != evaluator2._threshold
assert "relevance" in result1
assert "relevance" in result2
assert "evaluation_per_turn" in result1
assert "evaluation_per_turn" in result2
@patch("azure.ai.evaluation._evaluators._relevance._relevance." "RelevanceEvaluator.__call__")
def test_threshold_comparison_behavior(self, mock_call, test_model_config):
"""Test how different thresholds affect evaluation results."""
mock_result_high = {"relevance": 4.5, "relevance_result": "PASS", "evaluation_per_turn": {"relevance": [4.5]}}
mock_result_low = {"relevance": 2.5, "relevance_result": "FAIL", "evaluation_per_turn": {"relevance": [2.5]}}
mock_call.side_effect = [mock_result_high, mock_result_low]
strict_evaluator = RelevanceEvaluator(model_config=test_model_config, threshold=4)
lenient_evaluator = RelevanceEvaluator(model_config=test_model_config, threshold=2)
conversation = {
"messages": [
{"role": "user", "content": "What is the capital of France?"},
{"role": "assistant", "content": "Paris is the capital of France."},
]
}
strict_result = strict_evaluator(conversation=conversation)
lenient_result = lenient_evaluator(conversation=conversation)
assert strict_evaluator._threshold > lenient_evaluator._threshold
assert strict_result["relevance"] > lenient_result["relevance"]
assert "evaluation_per_turn" in strict_result
assert "evaluation_per_turn" in lenient_result
@pytest.mark.unittest
class TestEvaluatorsCombinedWithSample:
"""Test evaluators with sample conversations and thresholds."""
def test_sample_evaluators_threshold_setup(self, test_model_config):
"""Test setting up evaluators with different thresholds."""
evaluators = [
RelevanceEvaluator(model_config=test_model_config, threshold=threshold) for threshold in [1, 3, 5]
]
for i, evaluator in enumerate(evaluators):
assert evaluator._threshold == [1, 3, 5][i]
@patch("azure.ai.evaluation._evaluators._relevance._relevance." "RelevanceEvaluator.__call__")
def test_sample_evaluators_passing_thresholds(self, mock_call, test_model_config):
"""Test evaluators with passing threshold values."""
mock_result = {"relevance": 4.0, "relevance_result": "PASS", "evaluation_per_turn": {"relevance": [4.0]}}
mock_call.return_value = mock_result
evaluator = RelevanceEvaluator(model_config=test_model_config, threshold=3)
conversations = [
{
"messages": [
{"role": "user", "content": "What is 2+2?"},
{"role": "assistant", "content": "2+2 equals 4."},
]
},
{
"messages": [
{"role": "user", "content": "Who wrote Romeo and Juliet?"},
{"role": "assistant", "content": "Shakespeare was the author."},
]
},
]
for conversation in conversations:
result = evaluator(conversation=conversation)
assert "relevance" in result
assert result["relevance"] == 4.0
assert "evaluation_per_turn" in result
|
