Deva22467/govt-scheme-openenv

1from env import SchemeEnv2from run_agent import RuleBasedBaselineAgent3 4def run_validation():5    print("Running Formal Trajectory Validation...")6    env = SchemeEnv(seed=42)7    agent = RuleBasedBaselineAgent()8    9    total_reward = 0.010    episodes = 511    12    for i in range(episodes):13        obs = env.reset()14        done = False15        episode_reward = 0.016        17        while not done:18            action = agent.act(obs)19            obs, reward, done, info = env.step(action)20            episode_reward += reward21            assert reward >= 0.0, f"Negative Step Output: {reward}"22            23        print(f"Episode {i+1}: Task [{getattr(obs, 'task', 'None')}], Cumulative Reward: {episode_reward:.2f}, Final Stage: {info.get('stage')}")24        25        assert 0.0 <= episode_reward <= 1.0, f"Reward {episode_reward} out of bounds!"26        assert done is True, "Episode lock failed to trigger."27        assert "eligible_score" in info, "Missing logging schema."28        29        total_reward += episode_reward30        31    avg_reward = total_reward / episodes32    print("-" * 30)33    print(f"Validation Completed!")34    print(f"Final Average Reward: {avg_reward:.2f}")35    assert 0.0 <= avg_reward <= 1.0, "Score calculation failed bounds."36 37if __name__ == "__main__":38    run_validation()39