agent-harness / tests /test_analysis.py
cuber12's picture
Publish agent harness research code and paper artifacts
d61821a verified
Raw
History Blame Contribute Delete
1.32 kB
from __future__ import annotations
import unittest
from agent_harness.analysis import summarize_rows
class AnalysisTests(unittest.TestCase):
def test_paired_treatment_summary(self) -> None:
rows = []
values = {
"H000": [(0.0, 0.5), (1.0, 0.5)],
"H001": [(1.0, 1.0), (1.0, 1.0)],
}
for harness, task_values in values.items():
for index, (recall, mrr) in enumerate(task_values, start=1):
rows.append(
{
"task_id": f"TASK_{index}",
"harness_id": harness,
"all_gold_in_top_10": recall == 1.0,
"file_recall_at_1": recall,
"file_recall_at_5": recall,
"file_recall_at_10": recall,
"mrr": mrr,
"ndcg_at_10": recall,
"query_seconds": 0.1,
}
)
summary = summarize_rows(rows)
self.assertEqual(summary["treatments"]["H001"]["all_gold_in_top_10_count"], 2)
self.assertEqual(
summary["paired_differences"]["H001_minus_H000"]["mean_delta_file_recall_at_10"],
0.5,
)
if __name__ == "__main__":
unittest.main()