Spaces:
Sleeping
Sleeping
File size: 7,402 Bytes
8ddfad6 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 | """
Example usage of the Document Classification OpenEnv Environment
Demonstrates various ways to interact with the environment
"""
import numpy as np
from environment import DocumentClassificationEnv
from grading import BaselineAgent, AgentGrader
def example_basic_usage():
"""Example 1: Basic environment usage"""
print("=" * 70)
print("Example 1: Basic Environment Usage")
print("=" * 70)
# Create environment
env = DocumentClassificationEnv(task_difficulty="easy")
print(f"✓ Created environment: {env.task_difficulty}")
# Reset and get initial observation
obs, info = env.reset()
print(f"✓ Reset environment")
print(f" Initial document: {obs['document_id']}")
print(f" Content preview: {obs['content'][:60]}...")
# Take a few steps
total_reward = 0
for step in range(5):
action = env.action_space.sample() # Random action
obs, reward, done, truncated, info = env.step(action)
total_reward += reward
print(f"\nStep {step + 1}:")
print(f" Action: {env.CATEGORY_MAPS['easy'][action]}")
print(f" Reward: {reward:.3f}")
print(f" Episode Accuracy: {info['episode_accuracy']:.3f}")
if done:
print(f" Episode Complete!")
break
print(f"\nTotal Reward: {total_reward:.3f}")
def example_state_inspection():
"""Example 2: Inspecting environment state"""
print("\n" + "=" * 70)
print("Example 2: Environment State Inspection")
print("=" * 70)
env = DocumentClassificationEnv(task_difficulty="medium")
obs, _ = env.reset()
# Take a few steps
for _ in range(3):
action = env.action_space.sample()
obs, reward, done, _, info = env.step(action)
# Get and inspect state
state = env.state()
print("Environment State:")
print(f" Current document index: {state['current_document_index']}")
print(f" Total documents: {state['total_documents']}")
print(f" Episode reward: {state['episode_reward_total']:.3f}")
print(f" Accuracy so far: {state['episode_accuracy']:.3f}")
print(f" Avg processing time: {state['average_processing_time_ms']:.2f}ms")
print(f"\nCurrent Observation:")
print(f" Document ID: {state['current_observation']['document_id']}")
print(f" Word count: {state['current_observation']['word_count'][0]}")
print(f" Has urgency: {bool(state['current_observation']['has_urgency_markers'][0])}")
def example_baseline_agent():
"""Example 3: Using the baseline agent"""
print("\n" + "=" * 70)
print("Example 3: Baseline Agent Performance")
print("=" * 70)
for difficulty in ["easy", "medium", "hard"]:
print(f"\n{difficulty.upper()} Task:")
env = DocumentClassificationEnv(task_difficulty=difficulty)
agent = BaselineAgent(difficulty)
obs, _ = env.reset()
total_reward = 0
correct = 0
total = 0
for step in range(20): # Run 20 steps
action = agent.decide(obs)
obs, reward, done, truncated, info = env.step(action)
total_reward += reward
if info['is_correct']:
correct += 1
total += 1
if done:
break
accuracy = correct / total if total > 0 else 0
print(f" Accuracy: {accuracy:.3f} ({correct}/{total})")
print(f" Total Reward: {total_reward:.3f}")
print(f" Avg Reward: {total_reward / total:.3f}")
def example_agent_grading():
"""Example 4: Grading an agent"""
print("\n" + "=" * 70)
print("Example 4: Agent Grading System")
print("=" * 70)
difficulty = "easy"
# Create grader
grader = AgentGrader(difficulty)
# Create a simple custom agent
def custom_agent(obs):
"""A simple custom agent that prefers 'Support' category"""
# 70% of the time, choose Support (index 2 in easy)
if np.random.random() < 0.7:
return 2 # Support
else:
return np.random.randint(0, 5) # Random action
# Grade the custom agent
score, metrics = grader.grade_agent(custom_agent, verbose=True)
print(f"Final Score: {score:.4f}")
def example_difficulty_comparison():
"""Example 5: Comparing difficulty levels"""
print("\n" + "=" * 70)
print("Example 5: Difficulty Level Comparison")
print("=" * 70)
difficulties = ["easy", "medium", "hard"]
print(f"\n{'Difficulty':<10} {'Categories':<12} {'Documents':<12} {'Time Limit':<15} {'Features':<10}")
print("-" * 60)
for diff in difficulties:
env = DocumentClassificationEnv(task_difficulty=diff)
config = env.task_config
time_limit = config.get("time_limit", "None")
time_str = f"{time_limit}s" if time_limit else "None"
print(f"{diff:<10} {config['num_categories']:<12} {config['num_documents']:<12} {time_str:<15} {config['feature_dim']:<10}")
def example_seed_reproducibility():
"""Example 6: Demonstrating reproducibility with seeds"""
print("\n" + "=" * 70)
print("Example 6: Seed-based Reproducibility")
print("=" * 70)
# Create two environments with the same seed
env1 = DocumentClassificationEnv(task_difficulty="easy", seed=42)
env2 = DocumentClassificationEnv(task_difficulty="easy", seed=42)
obs1, _ = env1.reset(seed=42)
obs2, _ = env2.reset(seed=42)
print("Comparing two environments with seed=42:")
print(f" Env1 doc ID: {obs1['document_id']}")
print(f" Env2 doc ID: {obs2['document_id']}")
print(f" Match: {obs1['document_id'] == obs2['document_id']}")
print(f"\n Env1 content: {obs1['content'][:40]}...")
print(f" Env2 content: {obs2['content'][:40]}...")
print(f" Match: {obs1['content'] == obs2['content']}")
print(f"\n Feature vectors equal: {np.allclose(obs1['features'], obs2['features'])}")
def example_episode_summary():
"""Example 7: Getting episode summary"""
print("\n" + "=" * 70)
print("Example 7: Episode Summary")
print("=" * 70)
env = DocumentClassificationEnv(task_difficulty="easy")
obs, _ = env.reset()
# Run a full episode
done = False
while not done:
action = env.action_space.sample()
obs, reward, done, truncated, info = env.step(action)
# Get summary from info
summary = info.get("episode_summary", {})
print("Episode Summary:")
for key, value in summary.items():
if isinstance(value, float):
print(f" {key}: {value:.4f}")
else:
print(f" {key}: {value}")
def main():
"""Run all examples"""
try:
example_basic_usage()
example_state_inspection()
example_baseline_agent()
example_agent_grading()
example_difficulty_comparison()
example_seed_reproducibility()
example_episode_summary()
print("\n" + "=" * 70)
print("✓ All examples completed successfully!")
print("=" * 70)
except Exception as e:
print(f"\n✗ Error running examples: {e}")
import traceback
traceback.print_exc()
if __name__ == "__main__":
main()
|