Skip to content

armgit / camel

Public Imported from GitHub
0

🐫 CAMEL: The first and the best multi-agent framework. Finding the Scaling Law of Agents. https://www.camel-ai.org

enhance: Adds test for reset Method for BaseGenerator dataset PR3319 (#3323)

Wendong-Fan0a0c8897b453
diff --git a/test/environments/test_single_step_env.py b/test/environments/test_single_step_env.py
index dc3a552b5..b1c6af3e1 100644
--- a/test/environments/test_single_step_env.py
+++ b/test/environments/test_single_step_env.py
@@ -895,6 +895,9 @@ async def test_basegenerator_reset_reproducibility_with_identical_seed():
env = MockSingleStepEnv(dataset=dataset, verifier=mock_verifier)
await env.setup()
+ # Pre-populate dataset with all data to ensure sampling from full pool
+ await dataset.generate_new(PREDEFINED_DATA_SIZE)
+
# First reset with seed=42
observations1 = await env.reset(batch_size=RESET_BATCH_SIZE, seed=42)
questions1 = [obs.question for obs in observations1]
@@ -904,7 +907,6 @@ async def test_basegenerator_reset_reproducibility_with_identical_seed():
questions2 = [obs.question for obs in observations2]
# Same seed should produce same sampling order
- # given the large batch size)
assert (
questions1 == questions2
), "Same seed should produce same sampling order"
@@ -920,11 +922,14 @@ async def test_basegenerator_reset_variability_with_different_seeds():
RESET_BATCH_SIZE = 64
predefined_data = get_dummy_predefined_data(PREDEFINED_DATA_SIZE)
-
+ mock_verifier = create_mock_verifier()
dataset = MockGenerator(predefined_data, seed=42)
- env = MockSingleStepEnv(dataset=dataset, verifier=create_mock_verifier())
+ env = MockSingleStepEnv(dataset=dataset, verifier=mock_verifier)
await env.setup()
+ # Pre-populate dataset with all data to ensure sampling from full pool
+ await dataset.generate_new(PREDEFINED_DATA_SIZE)
+
observations1 = await env.reset(batch_size=RESET_BATCH_SIZE, seed=42)
questions1 = [obs.question for obs in observations1]
@@ -932,7 +937,6 @@ async def test_basegenerator_reset_variability_with_different_seeds():
questions2 = [obs.question for obs in observations2]
# Different seeds should produce different orders
- # given the large batch size
assert (
questions1 != questions2
), "Different seeds should produce different sampling orders"