Spaces:
Runtime error
Runtime error
File size: 3,653 Bytes
27d7338 | 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 | """Tests for reward shaping in RewardEngine."""
from __future__ import annotations
from env.models import CodeReviewAction, GroundTruthBug, ReviewComment
from env.reward_engine import RewardEngine
def test_add_comment_near_real_bug_positive() -> None:
"""Near-bug comment yields positive reward."""
gt = [GroundTruthBug(line_number=10, severity="major", category="bug", description="x")]
engine = RewardEngine(task_id="easy", ground_truth=gt, max_steps=8)
action = CodeReviewAction(operation="add_comment", line_number=10, severity="major", category="bug", message="x")
outcome = engine.compute(
action,
comments_so_far=[ReviewComment(line_number=10, severity="major", category="bug", message="x", step_added=1)],
correctly_identified_bug_lines=set(),
step_number=1,
steps_used_after_this=1,
)
assert outcome.reward > 0.0
def test_add_comment_on_red_herring_is_minus_point_two() -> None:
"""Flagging red herring yields -0.20."""
gt = [GroundTruthBug(line_number=10, severity="nit", category="style", description="trap", is_red_herring=True)]
engine = RewardEngine(task_id="hard", ground_truth=gt, max_steps=25)
action = CodeReviewAction(operation="add_comment", line_number=10, severity="nit", category="style", message="trap")
outcome = engine.compute(
action,
comments_so_far=[ReviewComment(line_number=10, severity="nit", category="style", message="trap", step_added=1)],
correctly_identified_bug_lines=set(),
step_number=1,
steps_used_after_this=1,
)
assert outcome.reward == -0.20
def test_add_comment_false_positive_is_minus_point_one() -> None:
"""False positive yields -0.10."""
gt = [GroundTruthBug(line_number=10, severity="major", category="bug", description="x")]
engine = RewardEngine(task_id="easy", ground_truth=gt, max_steps=8)
action = CodeReviewAction(operation="add_comment", line_number=100, severity="minor", category="style", message="nope")
outcome = engine.compute(
action,
comments_so_far=[ReviewComment(line_number=100, severity="minor", category="style", message="nope", step_added=1)],
correctly_identified_bug_lines=set(),
step_number=1,
steps_used_after_this=1,
)
assert outcome.reward == -0.10
def test_approve_with_unfound_critical_bugs_is_minus_point_five() -> None:
"""Approving with remaining critical/major bugs yields -0.50."""
gt = [GroundTruthBug(line_number=10, severity="critical", category="security", description="x")]
engine = RewardEngine(task_id="medium", ground_truth=gt, max_steps=15)
action = CodeReviewAction(operation="approve", summary="ok")
outcome = engine.compute(
action,
comments_so_far=[],
correctly_identified_bug_lines=set(),
step_number=1,
steps_used_after_this=1,
)
assert outcome.reward == -0.50
def test_efficiency_bonus_triggers() -> None:
"""Efficiency bonus triggers when under 60% steps and score > 0.8."""
gt = [GroundTruthBug(line_number=10, severity="major", category="bug", description="x")]
engine = RewardEngine(task_id="easy", ground_truth=gt, max_steps=10)
comments = [ReviewComment(line_number=10, severity="major", category="bug", message="x", step_added=1)]
action = CodeReviewAction(operation="done")
outcome = engine.compute(
action,
comments_so_far=comments,
correctly_identified_bug_lines={10},
step_number=2,
steps_used_after_this=2,
)
assert outcome.final_score == 0.999
assert outcome.reward == 1.099
|