Algorithmic refactor (#383)

* Refactor/document algorithmic environments and add tests.

* test for 3 row addition

* Fix failing rollout test by reinserting quirk in reversedAddition env

* todo regarding addition3-v0

* Fix python 3 division issues

* typo fix

* Re-generate python3 rollout file to account for ReversedAddition bug fix

gym/envs/algorithmic/duplicated_input.py

@@ -1,26 +1,25 @@
 """
-Task is to return every second character from the input tape.
+Task is to return every nth character from the input tape.
 http://arxiv.org/abs/1511.07275
 """
-
+from __future__ import division
 import numpy as np
 from gym.envs.algorithmic import algorithmic_env
-from gym.envs.algorithmic.algorithmic_env import ha
 
-class DuplicatedInputEnv(algorithmic_env.AlgorithmicEnv):
+class DuplicatedInputEnv(algorithmic_env.TapeAlgorithmicEnv):
     def __init__(self, duplication=2, base=5):
         self.duplication = duplication
-        algorithmic_env.AlgorithmicEnv.__init__(self,
-                                                inp_dim=1,
-                                                base=base,
-                                                chars=True)
-    def set_data(self):
-        self.content = {}
-        self.target = {}
-        copies = int(self.total_len / self.duplication)
-        for i in range(copies):
-            val = self.np_random.randint(self.base)
-            self.target[i] = val
-            for d in range(self.duplication):
-                self.content[ha(np.array([i * self.duplication + d]))] = val
-        self.total_reward = self.total_len / self.duplication
+        super(DuplicatedInputEnv, self).__init__(base=base, chars=True)
+
+    def generate_input_data(self, size):
+        res = []
+        if size < self.duplication:
+            size = self.duplication
+        for i in range(size//self.duplication):
+            char = self.np_random.randint(self.base)
+            for _ in range(self.duplication):
+                res.append(char)
+        return res
+
+    def target_from_input_data(self, input_data):
+        return [input_data[i] for i in range(0, len(input_data), self.duplication)]