From 912e3d42cb6b8021b015db3b3086e67ea7ac2a83 Mon Sep 17 00:00:00 2001
From: Brandon Rozek <rozekbrandon@gmail.com>
Date: Fri, 13 Sep 2019 19:48:24 -0400
Subject: [PATCH] Added new OpenAI Baseline Wrappers

---
 rltorch/env/wrappers.py | 132 +++++++++++++++++++++++++++++++++++++++-
 1 file changed, 131 insertions(+), 1 deletion(-)

diff --git a/rltorch/env/wrappers.py b/rltorch/env/wrappers.py
index 065b931..2bd5b97 100644
--- a/rltorch/env/wrappers.py
+++ b/rltorch/env/wrappers.py
@@ -3,6 +3,111 @@ import torch
 from gym import spaces
 import cv2
 from collections import deque
+import numpy as np
+
+class EpisodicLifeEnv(gym.Wrapper):
+    def __init__(self, env=None):
+        """Make end-of-life == end-of-episode, but only reset on true game over.
+        Done by DeepMind for the DQN and co. since it helps value estimation.
+        """
+        super(EpisodicLifeEnv, self).__init__(env)
+        self.lives = 0
+        self.was_real_done = True
+        self.was_real_reset = False
+
+    def step(self, action):
+        obs, reward, done, info = self.env.step(action)
+        self.was_real_done = done
+        # check current lives, make loss of life terminal,
+        # then update lives to handle bonus lives
+        lives = self.env.unwrapped.ale.lives()
+        if lives < self.lives and lives > 0:
+            # for Qbert somtimes we stay in lives == 0 condtion for a few frames
+            # so its important to keep lives > 0, so that we only reset once
+            # the environment advertises done.
+            done = True
+        self.lives = lives
+        return obs, reward, done, info
+
+    def reset(self):
+        """Reset only when lives are exhausted.
+        This way all states are still reachable even though lives are episodic,
+        and the learner need not know about any of this behind-the-scenes.
+        """
+        if self.was_real_done:
+            obs = self.env.reset()
+            self.was_real_reset = True
+        else:
+            # no-op step to advance from terminal/lost life state
+            obs, _, _, _ = self.env.step(0)
+            self.was_real_reset = False
+        self.lives = self.env.unwrapped.ale.lives()
+        return obs
+
+
+class NoopResetEnv(gym.Wrapper):
+    def __init__(self, env=None, noop_max=30):
+        """Sample initial states by taking random number of no-ops on reset.
+        No-op is assumed to be action 0.
+        """
+        super(NoopResetEnv, self).__init__(env)
+        self.noop_max = noop_max
+        self.override_num_noops = None
+        assert env.unwrapped.get_action_meanings()[0] == 'NOOP'
+
+    def step(self, action):
+        return self.env.step(action)
+
+    def reset(self):
+        """ Do no-op action for a number of steps in [1, noop_max]."""
+        self.env.reset()
+        if self.override_num_noops is not None:
+            noops = self.override_num_noops
+        else:
+            noops = np.random.randint(1, self.noop_max + 1)
+        assert noops > 0
+        obs = None
+        for _ in range(noops):
+            obs, _, done, _ = self.env.step(0)
+            if done:
+                obs = self.env.reset()
+        return obs
+
+
+class MaxAndSkipEnv(gym.Wrapper):
+    def __init__(self, env=None, skip=4):
+        """Return only every `skip`-th frame"""
+        super(MaxAndSkipEnv, self).__init__(env)
+        # most recent raw observations (for max pooling across time steps)
+        self._obs_buffer = deque(maxlen=2)
+        self._skip = skip
+
+    def step(self, action):
+        total_reward = 0.0
+        done = None
+        for _ in range(self._skip):
+            obs, reward, done, info = self.env.step(action)
+            self._obs_buffer.append(obs)
+            total_reward += reward
+            if done:
+                break
+
+        max_frame = np.max(np.stack(self._obs_buffer), axis=0)
+
+        return max_frame, total_reward, done, info
+
+    def reset(self):
+        """Clear past frame buffer and init. to first obs. from inner env."""
+        self._obs_buffer.clear()
+        obs = self.env.reset()
+        self._obs_buffer.append(obs)
+        return obs
+
+class ClippedRewardsWrapper(gym.RewardWrapper):
+    def reward(self, reward):
+        """Change all the positive rewards to 1, negative to -1 and keep zero."""
+        return np.sign(reward)
+
 
 # Mostly derived from OpenAI baselines
 class FireResetEnv(gym.Wrapper):
@@ -126,4 +231,29 @@ class TorchWrap(gym.Wrapper):
 
   def _convert(self, frame):
     frame = torch.from_numpy(frame).unsqueeze(0).float()
-    return frame
\ No newline at end of file
+    return frame
+
+
+
+class ProcessFrame84(gym.ObservationWrapper):
+    def __init__(self, env=None):
+        super(ProcessFrame84, self).__init__(env)
+        self.observation_space = spaces.Box(low=0, high=255, shape=(84, 84, 1), dtype=np.uint8)
+
+    def observation(self, obs):
+        return ProcessFrame84.process(obs)
+
+    @staticmethod
+    def process(frame):
+        if frame.size == 210 * 160 * 3:
+            img = np.reshape(frame, [210, 160, 3]).astype(np.float32)
+        elif frame.size == 250 * 160 * 3:
+            img = np.reshape(frame, [250, 160, 3]).astype(np.float32)
+        else:
+            assert False, "Unknown resolution."
+        img = img[:, :, 0] * 0.299 + img[:, :, 1] * 0.587 + img[:, :, 2] * 0.114
+        resized_screen = cv2.resize(img, (84, 110), interpolation=cv2.INTER_AREA)
+        x_t = resized_screen[18:102, :]
+        x_t = np.reshape(x_t, [84, 84])
+        return x_t.astype(np.uint8)
+