Add correct char reward as a constant

wordle_env/const.py

@@ -1,3 +1,4 @@
 WORDLE_CHARS = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ'
 WORDLE_N = 5
 REWARD = 10
+CHAR_REWARD = 0.1

wordle_env/state.py

@@ -12,10 +12,10 @@ where status has codes
  [0, 0, 1] - char is definitely in this spot
 """
 import collections
-from typing import List
+from typing import List, Tuple
 import numpy as np
 
-from .const import WORDLE_CHARS, WORDLE_N
+from .const import CHAR_REWARD, WORDLE_CHARS, WORDLE_N
 
 
 WordleState = np.ndarray
@@ -138,7 +138,7 @@ def update_mask(state: WordleState, word: str, goal_word: str) -> WordleState:
     return update_from_mask(state, word, mask)
 
 
-def update(state: WordleState, word: str, goal_word: str) -> WordleState:
+def update(state: WordleState, word: str, goal_word: str) -> Tuple(WordleState, float):
     state = state.copy()
     reward = 0
     state[0] -= 1
@@ -149,7 +149,7 @@ def update(state: WordleState, word: str, goal_word: str) -> WordleState:
         if goal_word[i] == c:
             # char at position i = yes, all other chars at position i == no
             if state[offset + 3 * i:offset + 3 * i + 3][2] == 0:
-                reward += 0.1
+                reward += CHAR_REWARD
             state[offset + 3 * i:offset + 3 * i + 3] = [0, 0, 1]
             for ocint in range(len(WORDLE_CHARS)):
                 if ocint != cint: