def after_step()

in gym-compete/gym_compete/new_envs/agents/humanoid_kicker.py [0:0]

• 11 lines of code
• 1 McCabe index (conditional complexity)

    def after_step(self, action):
        action = np.clip(action, self.action_space.low, self.action_space.high)
        # print(action)
        _, done, rinfo = super(HumanoidKicker, self).after_step(action)
        ball_xy = self.get_ball_qpos()[:2]
        my_xy = self.get_qpos()[:2]
        ball_dist = np.sqrt(np.sum((my_xy - ball_xy)**2))
        rinfo['reward_goal_dist'] = np.asscalar(ball_dist)
        reward = rinfo['reward_forward'] - rinfo['reward_ctrl'] - rinfo['reward_contact'] + rinfo['reward_survive'] - rinfo['reward_goal_dist']
        rinfo['reward_move'] = reward
        assert np.isfinite(reward), (rinfo, action)
        return reward, done, rinfo