def act()

in robosumo/policy_zoo/policy.py [0:0]

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

    def act(self, observation, stochastic=True):
        outputs = [self.sampled_action, self.vpred]
        feed_dict = {
            self.observation_ph: observation[None],
            self.stochastic_ph: stochastic,
        }
        a, v = tf.get_default_session().run(outputs, feed_dict)
        return a[0], {'vpred': v[0]}