Spaces:
Sleeping
Sleeping
File size: 4,379 Bytes
d43715d | 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 | """
Grader module for the Cat Rescue Reinforcement Learning environment.
Evaluates agent performance based on episode logs.
"""
class CatRescueGrader:
"""
Evaluates an episode log from the Cat Rescue environment and produces
structured grading results and human-readable summaries.
"""
def grade(self, episode_log: dict) -> dict:
"""
Grades an episode based on the provided episode log.
Args:
episode_log (dict): A dictionary containing:
- total_steps (int): Total steps taken by the agent.
- cats_rescued (int): Number of cats successfully rescued.
- total_cats (int): Total number of cats in the episode.
- total_reward (float): Total reward accumulated.
- hit_walls (int): Number of times the agent hit a wall.
- rescue_log (list of bool): True if rescued, one per cat.
Returns:
dict: A structured evaluation result containing:
- cats_rescued_count (int)
- all_cats_rescued (bool)
- rescue_details (list of dict)
- total_steps (int)
- wall_hits (int)
- total_reward (float)
- efficiency_score (float)
- passed (bool)
- grade_label (str)
"""
total_steps = episode_log.get("total_steps", 0)
cats_rescued = episode_log.get("cats_rescued", 0)
total_cats = episode_log.get("total_cats", 0)
total_reward = episode_log.get("total_reward", 0.0)
hit_walls = episode_log.get("hit_walls", 0)
rescue_log = episode_log.get("rescue_log", [])
all_cats_rescued = (cats_rescued == total_cats) and (total_cats > 0)
passed = all_cats_rescued
rescue_details = [{"cat_id": i, "rescued": rescued} for i, rescued in enumerate(rescue_log)]
efficiency_score = 0.0
if total_steps > 0:
efficiency_score = min(1.0, float(cats_rescued) / total_steps)
if passed and efficiency_score >= 0.8:
grade_label = "PERFECT"
elif passed and efficiency_score >= 0.5:
grade_label = "GOOD"
elif cats_rescued > 0 and not passed:
grade_label = "PARTIAL"
elif cats_rescued == 0:
grade_label = "FAIL"
else:
# Fallback if passed but efficiency < 0.5
grade_label = "PASS"
return {
"cats_rescued_count": cats_rescued,
"all_cats_rescued": all_cats_rescued,
"rescue_details": rescue_details,
"total_steps": total_steps,
"wall_hits": hit_walls,
"total_reward": total_reward,
"efficiency_score": efficiency_score,
"passed": passed,
"grade_label": grade_label
}
def summary(self, grade_result: dict) -> str:
"""
Generates a human-readable summary of the evaluation result.
Args:
grade_result (dict): The result dictionary produced by the grade method.
Returns:
str: A formatted summary string.
"""
status = "PASSED" if grade_result.get("passed") else "FAILED"
grade_label = grade_result.get("grade_label", "UNKNOWN")
cats_rescued = grade_result.get("cats_rescued_count", 0)
total_steps = grade_result.get("total_steps", 0)
efficiency = grade_result.get("efficiency_score", 0.0)
return (
f"--- Grader Summary ---\n"
f"Status: {status}\n"
f"Grade Label: {grade_label}\n"
f"Cats Rescued: {cats_rescued}\n"
f"Total Steps: {total_steps}\n"
f"Efficiency Score: {efficiency:.2f}\n"
f"----------------------"
)
if __name__ == "__main__":
sample_log = {
"total_steps": 5,
"cats_rescued": 3,
"total_cats": 3,
"total_reward": 15.5,
"hit_walls": 1,
"rescue_log": [True, True, True]
}
grader = CatRescueGrader()
result = grader.grade(sample_log)
print("Structured Result:")
import json
print(json.dumps(result, indent=2))
print("\nSummary:")
print(grader.summary(result))
|