wehub-resource-sync
34 lines
869 B
Python
34 lines
869 B
Python
from deepeval.evaluate import evaluate
|
|
from deepeval.test_case import LLMTestCase, ConversationalTestCase, Turn
|
|
|
|
|
|
def test_single_turn_evaluate():
|
|
evaluate(
|
|
test_cases=[
|
|
LLMTestCase(
|
|
input="What is the capital of France?", actual_output="Paris"
|
|
)
|
|
],
|
|
metric_collection="single_turn_test",
|
|
)
|
|
|
|
|
|
def test_multi_turn_evaluate():
|
|
evaluate(
|
|
test_cases=[
|
|
ConversationalTestCase(
|
|
turns=[
|
|
Turn(
|
|
role="user",
|
|
content="What is the capital of France?",
|
|
),
|
|
Turn(
|
|
role="assistant",
|
|
content="Paris",
|
|
),
|
|
]
|
|
)
|
|
],
|
|
metric_collection="multi_turn_test",
|
|
)
|