wordle!!

app.py

@@ -15,34 +15,47 @@ class WordleEnv:
         self._max = max_guesses
 
     def reset(self) -> str:
+        """Reset the environment and return the initial observation."""
         self._secret = random.choice(list(self.dictionary))
         self._n = 0
         self._obs = "⬜" * 4
         return self._obs
 
     def step(self, action: str) -> tuple[str, float, bool]:
+        """
+        Take an action (a 4-letter word) and return (observation, reward, done).
+        If 
+        When done is True, the episode has ended and reset() should be called to start a new episode.
+        """
+        if self._n >= self._max:
+            return "The game is over. Please reset.", -1.0, True
+
         guess: str = str(action)
         guess = guess.strip().lower()
 
         if len(guess) != 4 or not guess.isalpha():
-            raise ValueError("Action must be a 4-letter lowercase word.")
-
-        if self._n >= self._max:
-            raise RuntimeError("Episode is done. Call reset() to start a new episode.")
+            return "Invalid guess. Must be a 4-letter word.", -1.0, False
 
         self._n += 1
-        secret = self._secret
+
+        # Compute feedback
         feedback: list[str] = []
         for i, ch in enumerate(guess):
-            if ch == secret[i]:
+            if ch == self._secret[i]:
                 feedback.append("🟩")
-            elif ch in secret:
+            elif ch in self._secret:
                 feedback.append("🟨")
             else:
                 feedback.append("⬜")
         self._obs = "".join(feedback)
-        done = guess == secret or self._n >= self._max
-        reward = 1.0 if guess == secret else 0.0
+
+        # Check for success or timeout and compute reward
+        success = guess == self._secret
+        timeout = self._n >= self._max
+        done = success or timeout
+        reward = 1.0 if success else 0.0
+        if done and not success:
+            self._obs += f" Game over. The word was '{self._secret}'."
         return self._obs, reward, done
 
 

my_lib.py

@@ -42,6 +42,7 @@ def get_demo(env_cls: type) -> gr.Blocks:
     sessions = {}  # just a dict now
 
     def init_env() -> str:
+        """Initialize a new environment instance and return a session ID."""
         session_id = str(uuid.uuid4())
         env = env_cls()
         sessions[session_id] = env