added gym wordle package to edit

.gitignore

@@ -1,4 +1,3 @@
 **/data/*
-**/*.zip
-**/__pycache__
 /env
+**/*.zip

Gym-Wordle-main/.gitignore

@@ -0,0 +1,129 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+pip-wheel-metadata/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+.python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/

Gym-Wordle-main/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2022 David Kraemer
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

Gym-Wordle-main/README.md

@@ -0,0 +1,78 @@
+# Gym-Wordle
+
+An OpenAI gym compatible environment for training agents to play Wordle.
+
+<p align='center'>
+  <img src="https://user-images.githubusercontent.com/8514041/152437216-d78e85f6-8049-4cb9-ae61-3c015a8a0e4f.gif"><br/>
+  <em>User-input demo of the environment</em>
+</p>
+
+## Installation
+
+My goal is for a minimalist package that lets you install quickly and get on
+with your research. Installation is just a simple call to `pip`:
+
+```
+$ pip install gym_wordle
+```
+
+### Requirements
+
+In keeping with my desire to have a minimalist package, there are only three
+major requirements:
+
+* `numpy`
+* `gym`
+* `sty`, a lovely little package for stylizing text in terminals
+
+## Usage
+
+The basic flow for training agents with the `Wordle-v0` environment is the same
+as with gym environments generally:
+
+```Python
+import gym
+import gym_wordle
+
+eng = gym.make("Wordle-v0")
+
+done = False
+while not done:
+    action = ...  # RL magic
+    state, reward, done, info = env.step(action)
+```
+
+If you're like millions of other people, you're a Wordle-obsessive in your own
+right. I have good news for you! The `Wordle-v0` environment currently has an
+implemented `render` method, which allows you to see a human-friendly version
+of the game. And it isn't so hard to set up the environment to play for
+yourself. Here's an example script:
+
+```Python
+from gym_wordle.utils import play
+
+play()
+```
+
+## Documentation
+
+Coming soon!
+
+## Examples
+
+Coming soon!
+
+## Citing
+
+If you decide to use this project in your work, please consider a citation!
+
+```bibtex
+@misc{gym_wordle,
+  author = {Kraemer, David},
+  title = {An Environment for Reinforcement Learning with Wordle},
+  year = {2022},
+  publisher = {GitHub},
+  journal = {GitHub repository},
+  howpublished = {\url{https://github.com/DavidNKraemer/Gym-Wordle}},
+}
+```

Gym-Wordle-main/gym-wordle.toml

@@ -0,0 +1,7 @@
+[build-system]
+
+requires = [
+  "setuptools>=42",
+  "wheel"
+]
+build-backend = "setuptools.build_meta"

Gym-Wordle-main/gym_wordle/utils.py

@@ -91,3 +91,4 @@ def play():
         env.render()
     
     print(f"The word was {solution}")
+

Gym-Wordle-main/gym_wordle/wordle.py

@@ -1,16 +1,18 @@
-import gymnasium as gym
+import gym
 import numpy as np
 import numpy.typing as npt
 from sty import fg, bg, ef, rs
 
-from collections import Counter, defaultdict
+from collections import Counter 
 from gym_wordle.utils import to_english, to_array, get_words
 from typing import Optional 
 
+
 class WordList(gym.spaces.Discrete):
     """Super class for defining a space of valid words according to a specified
     list.
 
+    TODO: Fix these paragraphs
     The space is a subclass of gym.spaces.Discrete, where each element
     corresponds to an index of a valid word in the word list. The obfuscation
     is necessary for more direct implementation of RL algorithms, which expect
@@ -71,9 +73,10 @@ class SolutionList(WordList):
 
     Of course, the set of solutions is a strict subset of the set of guesses.
 
+    Reference: https://fivethirtyeight.com/features/when-the-riddler-met-wordle/
+
     This class represents the set of solution words.
     """
-
     def __init__(self, **kwargs):
         """
         Args:
@@ -85,7 +88,7 @@ class SolutionList(WordList):
 
 class WordleObsSpace(gym.spaces.Box):
     """Implementation of the state (observation) space in terms of gym
-    primitives, in this case, gym.spaces.Box.
+    primatives, in this case, gym.spaces.Box.
 
     The Wordle observation space can be thought of as a 6x5 array with two
     channels:
@@ -98,11 +101,20 @@ class WordleObsSpace(gym.spaces.Box):
     where there are 6 rows, one for each turn in the game, and 5 columns, since
     the solution will always be a word of length 5.
 
-    For simplicity, and compatibility with stable_baselines algorithms,
+    For simplicity, and compatibility with the stable_baselines algorithms,
     this multichannel is modeled as a 6x10 array, where the two channels are
     horizontally appended (along columns). Thus each row in the observation
-    should be interpreted as c0 c1 c2 c3 c4 f0 f1 f2 f3 f4 when the word is
+    should be interpreted as 
-    c0...c4 and its associated flags are f0...f4.
+
+        c0 c1 c2 c3 c4 f0 f1 f2 f3 f4
+
+    when the word is c0...c4 and its associated flags are f0...f4.
+
+    While the superclass method `sample` is available to the WordleObsSpace, it
+    should be emphasized that the output of `sample` will (almost surely) not
+    correspond to a real game configuration, because the sampling is not out of
+    possible game configurations. Instead, the Box superclass just samples the
+    integer array space uniformly.
     """
 
     def __init__(self, **kwargs):
@@ -119,11 +131,20 @@ class WordleObsSpace(gym.spaces.Box):
 
 
 class GuessList(WordList):
-    """Space for *guess* words to the Wordle environment.
+    """Space for *solution* words to the Wordle environment.
+
+    In the game Wordle, there are two different collections of words:
+
+        * "guesses", which the game accepts as valid words to use to guess the
+          answer.
+        * "solutions", which the game uses to choose solutions from.
+
+    Of course, the set of solutions is a strict subset of the set of guesses.
+
+    Reference: https://fivethirtyeight.com/features/when-the-riddler-met-wordle/
 
     This class represents the set of guess words.
     """
-
     def __init__(self, **kwargs):
         """
         Args:
@@ -134,9 +155,10 @@ class GuessList(WordList):
 
 
 class WordleEnv(gym.Env):
+
     metadata = {'render.modes': ['human']}
 
-    # Character flag codes
+    # character flag codes
     no_char = 0
     right_pos = 1
     wrong_pos = 2
@@ -145,6 +167,7 @@ class WordleEnv(gym.Env):
     def __init__(self):
         super().__init__()
 
+        self.seed()
         self.action_space = GuessList()
         self.solution_space = SolutionList()
 
@@ -159,7 +182,6 @@ class WordleEnv(gym.Env):
 
         self.n_rounds = 6
         self.n_letters = 5
-        self.info = {'correct': False, 'guesses': defaultdict(int)}
 
     def _highlighter(self, char: str, flag: int) -> str:
         """Terminal renderer functionality. Properly highlights a character
@@ -180,43 +202,36 @@ class WordleEnv(gym.Env):
         front, back = self._highlights[flag]
         return front + char + back
 
-    def reset(self, seed=None, options=None):
+    def reset(self):
-        """Reset the environment to an initial state and returns an initial
-        observation.
-
-        Note: The observation space instance should be a Box space.
-
-        Returns:
-            state (object): The initial observation of the space.
-        """
         self.round = 0
         self.solution = self.solution_space.sample()
 
-        self.state = np.zeros((self.n_rounds, 2 * self.n_letters), dtype=np.int64)
+        self.state = np.zeros((self.n_rounds, 2 * self.n_letters),
+                              dtype=np.int64)
 
-        self.info = {'correct': False, 'guesses': defaultdict(int)}
+        return self.state
-
-        return self.state, self.info
 
     def render(self, mode: str ='human'):
         """Renders the Wordle environment.
 
         Currently supported render modes:
+
         - human: renders the Wordle game to the terminal.
 
         Args:
-            mode: the mode to render with.
+            mode: the mode to render with
         """
         if mode == 'human':
-            for row in self.state:
+            for row in self.states:
                 text = ''.join(map(
                     self._highlighter, 
                     to_english(row[:self.n_letters]).upper(), 
                     row[self.n_letters:]
                 ))
+
                 print(text)
         else:
-            super().render(mode=mode)
+            super(WordleEnv, self).render(mode=mode)
                 
     def step(self, action):
         """Run one step of the Wordle game. Every game must be previously
@@ -224,31 +239,38 @@ class WordleEnv(gym.Env):
 
         Args:
             action: Word guessed by the agent.
-
         Returns:
             state (object): Wordle game state after the guess.
-            reward (float): Reward associated with the guess.
+            reward (float): Reward associated with the guess (-1 for incorrect,
-            done (bool): Whether the game has ended.
+              0 for correct)
-            info (dict): Auxiliary diagnostic information.
+            done (bool): Whether the game has ended (by a correct guess or
+              after six guesses).
+            info (dict): Auxiliary diagnostic information (empty).
         """
         assert self.action_space.contains(action), 'Invalid word!'
 
+        # transform the action, solution indices to their words
         action = self.action_space[action]  
         solution = self.solution_space[self.solution]
 
+        # populate the word chars into the row (character channel)
         self.state[self.round][:self.n_letters] = action
 
+        # populate the flag characters into the row (flag channel)
         counter = Counter()
         for i, char in enumerate(action):
-            flag_i = i + self.n_letters
+            flag_i = i + self.n_letters  # starts at 5
             counter[char] += 1
 
-            if char == solution[i]:
+            if char == solution[i]:  # character is in correct position
-                self.state[self.round, flag_i] = self.right_pos
+                self.state[self.round, i] = self.right_pos
             elif counter[char] <= (char == solution).sum():  
-                self.state[self.round, flag_i] = self.wrong_pos
+                # current character has been seen within correct number of
+                # occurrences 
+                self.state[self.round, i] = self.wrong_pos
             else:
-                self.state[self.round, flag_i] = self.wrong_char
+                # wrong character, or "correct" character too many times
+                self.state[self.round, i] = self.wrong_char
 
         self.round += 1
 
@@ -257,29 +279,8 @@ class WordleEnv(gym.Env):
 
         done = correct or game_over
 
-        reward = 0
+        # Total reward equals -(number of incorrect guesses)
-        # correct spot
+        reward = 0. if correct else -1.
-        reward += np.sum(self.state[:, 5:] == 1) * 2
 
-        # correct letter not correct spot
+        return self.state, reward, done, {}
-        reward += np.sum(self.state[:, 5:] == 2) * 1
 
-        # incorrect letter
-        reward += np.sum(self.state[:, 5:] == 3) * -1
-
-        # guess same word as before
-        hashable_action = to_english(action)
-        if hashable_action in self.info['guesses']:
-            reward += -10 * self.info['guesses'][hashable_action]
-        else:  # guess different word
-            reward += 10
-
-        self.info['guesses'][hashable_action] += 1
-
-        # for game ending in win or loss
-        reward += 10 if correct else -10 if done else 0
-
-        self.info['correct'] = correct
-
-        # observation, reward, terminated, truncated, info
-        return self.state, reward, done, False, self.info

Gym-Wordle-main/setup.py

@@ -0,0 +1,35 @@
+from setuptools import setup, find_packages
+
+with open('README.md', 'r', encoding='utf-8') as fh:
+    long_description = fh.read()
+
+setup(
+    name='gym_wordle',
+    version='0.1.3',
+    author='David Kraemer',
+    author_email='david.kraemer@stonybrook.edu',
+    description='OpenAI gym environment for training agents on Wordle',
+    long_description=long_description,
+    long_description_content_type='text/markdown',
+    url='https://github.com/DavidNKraemer/Gym-Wordle',
+    packages=find_packages(
+        include=[
+            'gym_wordle',
+            'gym_wordle.*'
+        ]
+    ),
+    package_data={
+        'gym_wordle': ['dictionary/*']
+    },
+    python_requires='>=3.7',
+    classifiers=[
+        "Programming Language :: Python :: 3",
+        "License :: OSI Approved :: MIT License",
+        "Operating System :: OS Independent",
+    ],
+    install_requires=[
+        'numpy>=1.20',
+        'gym==0.19',
+        'sty==1.0',
+    ],
+)

custom_env/agent.py

@@ -0,0 +1,91 @@
+import torch
+
+
+class Agent:
+
+    def __init__(self, ) -> None:
+        # BATCH_SIZE is the number of transitions sampled from the replay buffer
+        # GAMMA is the discount factor as mentioned in the previous section
+        # EPS_START is the starting value of epsilon
+        # EPS_END is the final value of epsilon
+        # EPS_DECAY controls the rate of exponential decay of epsilon, higher means a slower decay
+        # TAU is the update rate of the target network
+        # LR is the learning rate of the ``AdamW`` optimizer
+        self.batch_size = 128
+        self.gamma = 0.99
+        self.eps_start = 0.9
+        self.eps_end = 0.05
+        self.eps_decay = 1000
+        self.tau = 0.005
+        self.lr = 1e-4
+        self.n_actions = n_actions
+
+        policy_net = DQN(n_observations, n_actions).to(device)
+        target_net = DQN(n_observations, n_actions).to(device)
+        target_net.load_state_dict(policy_net.state_dict())
+
+        optimizer = optim.AdamW(policy_net.parameters(), lr=LR, amsgrad=True)
+        memory = ReplayMemory(10000)
+
+    def get_state(self, game):
+        pass
+
+    def select_action(state):
+        sample = random.random()
+        eps_threshold = EPS_END + (EPS_START - EPS_END) * \
+            math.exp(-1. * steps_done / EPS_DECAY)
+        steps_done += 1
+        if sample > eps_threshold:
+            with torch.no_grad():
+                # t.max(1) will return the largest column value of each row.
+                # second column on max result is index of where max element was
+                # found, so we pick action with the larger expected reward.
+                return policy_net(state).max(1).indices.view(1, 1)
+        else:
+            return torch.tensor([[env.action_space.sample()]], device=device, dtype=torch.long)
+
+    def optimize_model():
+        if len(memory) < BATCH_SIZE:
+            return
+        transitions = memory.sample(BATCH_SIZE)
+        # Transpose the batch (see https://stackoverflow.com/a/19343/3343043 for
+        # detailed explanation). This converts batch-array of Transitions
+        # to Transition of batch-arrays.
+        batch = Transition(*zip(*transitions))
+
+        # Compute a mask of non-final states and concatenate the batch elements
+        # (a final state would've been the one after which simulation ended)
+        non_final_mask = torch.tensor(tuple(map(lambda s: s is not None,
+                                            batch.next_state)), device=device, dtype=torch.bool)
+        non_final_next_states = torch.cat([s for s in batch.next_state
+                                           if s is not None])
+        state_batch = torch.cat(batch.state)
+        action_batch = torch.cat(batch.action)
+        reward_batch = torch.cat(batch.reward)
+
+        # Compute Q(s_t, a) - the model computes Q(s_t), then we select the
+        # columns of actions taken. These are the actions which would've been taken
+        # for each batch state according to policy_net
+        state_action_values = policy_net(state_batch).gather(1, action_batch)
+
+        # Compute V(s_{t+1}) for all next states.
+        # Expected values of actions for non_final_next_states are computed based
+        # on the "older" target_net; selecting their best reward with max(1).values
+        # This is merged based on the mask, such that we'll have either the expected
+        # state value or 0 in case the state was final.
+        next_state_values = torch.zeros(BATCH_SIZE, device=device)
+        with torch.no_grad():
+            next_state_values[non_final_mask] = target_net(non_final_next_states).max(1).values
+        # Compute the expected Q values
+        expected_state_action_values = (next_state_values * GAMMA) + reward_batch
+
+        # Compute Huber loss
+        criterion = nn.SmoothL1Loss()
+        loss = criterion(state_action_values, expected_state_action_values.unsqueeze(1))
+
+        # Optimize the model
+        optimizer.zero_grad()
+        loss.backward()
+        # In-place gradient clipping
+        torch.nn.utils.clip_grad_value_(policy_net.parameters(), 100)
+        optimizer.step()

custom_env/create_wordlist.py

@@ -0,0 +1,16 @@
+import pathlib
+import sys
+from string import ascii_letters
+
+in_path = pathlib.Path(sys.argv[1])
+out_path = pathlib.Path(sys.argv[2])
+
+words = sorted(
+    {
+        word.lower()
+        for word in in_path.read_text(encoding="utf-8").split()
+        if all(letter in ascii_letters for letter in word)
+    },
+    key=lambda word: (len(word), word),
+)
+out_path.write_text("\n".join(words))

custom_env/model.py

@@ -0,0 +1,44 @@
+import math
+import random
+import matplotlib
+import matplotlib.pyplot as plt
+from collections import namedtuple, deque
+from itertools import count
+
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import torch.nn.functional as F
+
+
+Transition = namedtuple('Transition',
+                        ('state', 'action', 'next_state', 'reward'))
+
+
+class ReplayMemory(object):
+
+    def __init__(self, capacity: int) -> None:
+        self.memory = deque([], maxlen=capacity)
+
+    def push(self, *args):
+        self.memory.append(Transition(*args))
+
+    def sample(self, batch_size):
+        return random.sample(self.memory, batch_size)
+
+    def __len__(self):
+        return len(self.memory)
+
+
+class DQN(nn.Module):
+
+    def __init__(self, n_observations: int, n_actions: int) -> None:
+        super(DQN, self).__init__()
+        self.layer1 = nn.Linear(n_observations, 128)
+        self.layer2 = nn.Linear(128, 128)
+        self.layer3 = nn.Linear(128, n_actions)
+
+    def forward(self, x):
+        x = F.relu(self.layer1(x))
+        x = F.relu(self.layer2(x))
+        return self.layer3(x)

custom_env/test2.ipynb

@@ -0,0 +1,61 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from string import ascii_letters, ascii_uppercase, ascii_lowercase"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'ABCDEFGHIJKLMNOPQRSTUVWXYZ'"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "ascii_uppercase"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "env",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.5"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

custom_env/wyrdl.py

@@ -0,0 +1,119 @@
+import contextlib
+import pathlib
+import random
+from string import ascii_letters, ascii_lowercase
+
+from rich.console import Console
+from rich.theme import Theme
+
+console = Console(width=40, theme=Theme({"warning": "red on yellow"}))
+
+NUM_LETTERS = 5
+NUM_GUESSES = 6
+WORDS_PATH = pathlib.Path(__file__).parent / "wordlist.txt"
+
+
+class Wordle:
+
+    def __init__(self) -> None:
+        self.word_list = WORDS_PATH.read_text(encoding="utf-8").split("\n")
+        self.n_guesses = 6
+        self.num_letters = 5
+        self.curr_word = None
+        self.reset()
+
+    def refresh_page(self, headline):
+        console.clear()
+        console.rule(f"[bold blue]:leafy_green: {headline} :leafy_green:[/]\n")
+
+    def start_game(self):
+        # get a new random word
+        word = self.get_random_word(self.word_list)
+
+        self.curr_word = word
+
+    def get_state(self):
+        return 
+
+    def action_to_word(self, action):
+        # Calculate the word from the array
+        word = ''
+        for i in range(0, len(ascii_lowercase), 26):
+            # Find the index of 1 in each block of 26
+            letter_index = action[i:i+26].index(1)
+            # Append the corresponding letter to the word
+            word += ascii_lowercase[letter_index]
+
+        return word
+
+    def play_guess(self, action):
+        # probably an array of length 26 * 5 for 26 letters and 5 positions
+        guess = action
+
+    def get_random_word(self, word_list):
+        if words := [
+            word.upper()
+            for word in word_list
+            if len(word) == NUM_LETTERS
+            and all(letter in ascii_letters for letter in word)
+        ]:
+            return random.choice(words)
+        else:
+            console.print(
+                f"No words of length {NUM_LETTERS} in the word list",
+                style="warning",
+            )
+            raise SystemExit()
+
+    def show_guesses(self, guesses, word):
+        letter_status = {letter: letter for letter in ascii_lowercase}
+        for guess in guesses:
+            styled_guess = []
+            for letter, correct in zip(guess, word):
+                if letter == correct:
+                    style = "bold white on green"
+                elif letter in word:
+                    style = "bold white on yellow"
+                elif letter in ascii_letters:
+                    style = "white on #666666"
+                else:
+                    style = "dim"
+                styled_guess.append(f"[{style}]{letter}[/]")
+                if letter != "_":
+                    letter_status[letter] = f"[{style}]{letter}[/]"
+
+            console.print("".join(styled_guess), justify="center")
+        console.print("\n" + "".join(letter_status.values()), justify="center")
+
+    def guess_word(self, previous_guesses):
+        guess = console.input("\nGuess word: ").upper()
+
+        if guess in previous_guesses:
+            console.print(f"You've already guessed {guess}.", style="warning")
+            return guess_word(previous_guesses)
+
+        if len(guess) != NUM_LETTERS:
+            console.print(
+                f"Your guess must be {NUM_LETTERS} letters.", style="warning"
+            )
+            return guess_word(previous_guesses)
+
+        if any((invalid := letter) not in ascii_letters for letter in guess):
+            console.print(
+                f"Invalid letter: '{invalid}'. Please use English letters.",
+                style="warning",
+            )
+            return guess_word(previous_guesses)
+
+        return guess
+
+    def reset(self, guesses, word, guessed_correctly, n_episodes):
+        refresh_page(headline=f"Game: {n_episodes}")
+
+        if guessed_correctly:
+            console.print(f"\n[bold white on green]Correct, the word is {word}[/]")
+        else:
+            console.print(f"\n[bold white on red]Sorry, the word was {word}[/]")
+
+    if __name__ == "__main__":
+        main()

dqn_wordle.ipynb

@@ -6,10 +6,9 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import gym_wordle\n",
+    "import gymnasium as gym\n",
-    "from stable_baselines3 import DQN, PPO, common\n",
+    "from stable_baselines3 import DQN\n",
-    "import numpy as np\n",
+    "import numpy as np"
-    "from tqdm import tqdm"
    ]
   },
   {
@@ -18,252 +17,68 @@
    "metadata": {},
    "outputs": [
     {
-     "name": "stdout",
+     "ename": "NameNotFound",
-     "output_type": "stream",
+     "evalue": "Environment `Wordle` doesn't exist.",
-     "text": [
+     "output_type": "error",
-      "<Monitor<WordleEnv instance>>\n"
+     "traceback": [
+      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[1;31mNameNotFound\u001b[0m                              Traceback (most recent call last)",
+      "Cell \u001b[1;32mIn[2], line 1\u001b[0m\n\u001b[1;32m----> 1\u001b[0m env \u001b[38;5;241m=\u001b[39m \u001b[43mgym\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mmake\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mWordle-v0\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[0;32m      3\u001b[0m \u001b[38;5;28mprint\u001b[39m(env)\n",
+      "File \u001b[1;32mc:\\Repository\\cse151b-final-project\\env\\Lib\\site-packages\\gymnasium\\envs\\registration.py:741\u001b[0m, in \u001b[0;36mmake\u001b[1;34m(id, max_episode_steps, autoreset, apply_api_compatibility, disable_env_checker, **kwargs)\u001b[0m\n\u001b[0;32m    738\u001b[0m     \u001b[38;5;28;01massert\u001b[39;00m \u001b[38;5;28misinstance\u001b[39m(\u001b[38;5;28mid\u001b[39m, \u001b[38;5;28mstr\u001b[39m)\n\u001b[0;32m    740\u001b[0m     \u001b[38;5;66;03m# The environment name can include an unloaded module in \"module:env_name\" style\u001b[39;00m\n\u001b[1;32m--> 741\u001b[0m     env_spec \u001b[38;5;241m=\u001b[39m \u001b[43m_find_spec\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mid\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[0;32m    743\u001b[0m \u001b[38;5;28;01massert\u001b[39;00m \u001b[38;5;28misinstance\u001b[39m(env_spec, EnvSpec)\n\u001b[0;32m    745\u001b[0m \u001b[38;5;66;03m# Update the env spec kwargs with the `make` kwargs\u001b[39;00m\n",
+      "File \u001b[1;32mc:\\Repository\\cse151b-final-project\\env\\Lib\\site-packages\\gymnasium\\envs\\registration.py:527\u001b[0m, in \u001b[0;36m_find_spec\u001b[1;34m(env_id)\u001b[0m\n\u001b[0;32m    521\u001b[0m     logger\u001b[38;5;241m.\u001b[39mwarn(\n\u001b[0;32m    522\u001b[0m         \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mUsing the latest versioned environment `\u001b[39m\u001b[38;5;132;01m{\u001b[39;00mnew_env_id\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m` \u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[0;32m    523\u001b[0m         \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124minstead of the unversioned environment `\u001b[39m\u001b[38;5;132;01m{\u001b[39;00menv_name\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m`.\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[0;32m    524\u001b[0m     )\n\u001b[0;32m    526\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m env_spec \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[1;32m--> 527\u001b[0m     \u001b[43m_check_version_exists\u001b[49m\u001b[43m(\u001b[49m\u001b[43mns\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mname\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mversion\u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m    528\u001b[0m     \u001b[38;5;28;01mraise\u001b[39;00m error\u001b[38;5;241m.\u001b[39mError(\n\u001b[0;32m    529\u001b[0m         \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mNo registered env with id: \u001b[39m\u001b[38;5;132;01m{\u001b[39;00menv_name\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m. Did you register it, or import the package that registers it? Use `gymnasium.pprint_registry()` to see all of the registered environments.\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[0;32m    530\u001b[0m     )\n\u001b[0;32m    532\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m env_spec\n",
+      "File \u001b[1;32mc:\\Repository\\cse151b-final-project\\env\\Lib\\site-packages\\gymnasium\\envs\\registration.py:393\u001b[0m, in \u001b[0;36m_check_version_exists\u001b[1;34m(ns, name, version)\u001b[0m\n\u001b[0;32m    390\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m get_env_id(ns, name, version) \u001b[38;5;129;01min\u001b[39;00m registry:\n\u001b[0;32m    391\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m\n\u001b[1;32m--> 393\u001b[0m \u001b[43m_check_name_exists\u001b[49m\u001b[43m(\u001b[49m\u001b[43mns\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mname\u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m    394\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m version \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[0;32m    395\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m\n",
+      "File \u001b[1;32mc:\\Repository\\cse151b-final-project\\env\\Lib\\site-packages\\gymnasium\\envs\\registration.py:370\u001b[0m, in \u001b[0;36m_check_name_exists\u001b[1;34m(ns, name)\u001b[0m\n\u001b[0;32m    367\u001b[0m namespace_msg \u001b[38;5;241m=\u001b[39m \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m in namespace \u001b[39m\u001b[38;5;132;01m{\u001b[39;00mns\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m\"\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m ns \u001b[38;5;28;01melse\u001b[39;00m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[0;32m    368\u001b[0m suggestion_msg \u001b[38;5;241m=\u001b[39m \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m Did you mean: `\u001b[39m\u001b[38;5;132;01m{\u001b[39;00msuggestion[\u001b[38;5;241m0\u001b[39m]\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m`?\u001b[39m\u001b[38;5;124m\"\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m suggestion \u001b[38;5;28;01melse\u001b[39;00m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m--> 370\u001b[0m \u001b[38;5;28;01mraise\u001b[39;00m error\u001b[38;5;241m.\u001b[39mNameNotFound(\n\u001b[0;32m    371\u001b[0m     \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mEnvironment `\u001b[39m\u001b[38;5;132;01m{\u001b[39;00mname\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m` doesn\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mt exist\u001b[39m\u001b[38;5;132;01m{\u001b[39;00mnamespace_msg\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m.\u001b[39m\u001b[38;5;132;01m{\u001b[39;00msuggestion_msg\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m\"\u001b[39m\n\u001b[0;32m    372\u001b[0m )\n",
+      "\u001b[1;31mNameNotFound\u001b[0m: Environment `Wordle` doesn't exist."
      ]
     }
    ],
    "source": [
-    "env = gym_wordle.wordle.WordleEnv()\n",
+    "env = gym.make(\"Wordle-v0\")\n",
-    "env = common.monitor.Monitor(env)\n",
     "\n",
     "print(env)"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 35,
    "metadata": {},
-   "outputs": [
+   "outputs": [],
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Using cuda device\n",
-      "Wrapping the env in a DummyVecEnv.\n",
-      "---------------------------------\n",
-      "| rollout/           |          |\n",
-      "|    ep_len_mean     | 6        |\n",
-      "|    ep_rew_mean     | 2.14     |\n",
-      "| time/              |          |\n",
-      "|    fps             | 750      |\n",
-      "|    iterations      | 1        |\n",
-      "|    time_elapsed    | 2        |\n",
-      "|    total_timesteps | 2048     |\n",
-      "---------------------------------\n",
-      "-----------------------------------------\n",
-      "| rollout/                |             |\n",
-      "|    ep_len_mean          | 6           |\n",
-      "|    ep_rew_mean          | 4.59        |\n",
-      "| time/                   |             |\n",
-      "|    fps                  | 625         |\n",
-      "|    iterations           | 2           |\n",
-      "|    time_elapsed         | 6           |\n",
-      "|    total_timesteps      | 4096        |\n",
-      "| train/                  |             |\n",
-      "|    approx_kl            | 0.022059526 |\n",
-      "|    clip_fraction        | 0.331       |\n",
-      "|    clip_range           | 0.2         |\n",
-      "|    entropy_loss         | -9.47       |\n",
-      "|    explained_variance   | -0.0118     |\n",
-      "|    learning_rate        | 0.0003      |\n",
-      "|    loss                 | 130         |\n",
-      "|    n_updates            | 10          |\n",
-      "|    policy_gradient_loss | -0.0851     |\n",
-      "|    value_loss           | 253         |\n",
-      "-----------------------------------------\n",
-      "-----------------------------------------\n",
-      "| rollout/                |             |\n",
-      "|    ep_len_mean          | 6           |\n",
-      "|    ep_rew_mean          | 5.86        |\n",
-      "| time/                   |             |\n",
-      "|    fps                  | 585         |\n",
-      "|    iterations           | 3           |\n",
-      "|    time_elapsed         | 10          |\n",
-      "|    total_timesteps      | 6144        |\n",
-      "| train/                  |             |\n",
-      "|    approx_kl            | 0.024416003 |\n",
-      "|    clip_fraction        | 0.462       |\n",
-      "|    clip_range           | 0.2         |\n",
-      "|    entropy_loss         | -9.47       |\n",
-      "|    explained_variance   | 0.152       |\n",
-      "|    learning_rate        | 0.0003      |\n",
-      "|    loss                 | 85.2        |\n",
-      "|    n_updates            | 20          |\n",
-      "|    policy_gradient_loss | -0.0987     |\n",
-      "|    value_loss           | 218         |\n",
-      "-----------------------------------------\n",
-      "-----------------------------------------\n",
-      "| rollout/                |             |\n",
-      "|    ep_len_mean          | 6           |\n",
-      "|    ep_rew_mean          | 4.75        |\n",
-      "| time/                   |             |\n",
-      "|    fps                  | 566         |\n",
-      "|    iterations           | 4           |\n",
-      "|    time_elapsed         | 14          |\n",
-      "|    total_timesteps      | 8192        |\n",
-      "| train/                  |             |\n",
-      "|    approx_kl            | 0.026305672 |\n",
-      "|    clip_fraction        | 0.45        |\n",
-      "|    clip_range           | 0.2         |\n",
-      "|    entropy_loss         | -9.47       |\n",
-      "|    explained_variance   | 0.161       |\n",
-      "|    learning_rate        | 0.0003      |\n",
-      "|    loss                 | 144         |\n",
-      "|    n_updates            | 30          |\n",
-      "|    policy_gradient_loss | -0.105      |\n",
-      "|    value_loss           | 220         |\n",
-      "-----------------------------------------\n",
-      "----------------------------------------\n",
-      "| rollout/                |            |\n",
-      "|    ep_len_mean          | 6          |\n",
-      "|    ep_rew_mean          | 1.47       |\n",
-      "| time/                   |            |\n",
-      "|    fps                  | 554        |\n",
-      "|    iterations           | 5          |\n",
-      "|    time_elapsed         | 18         |\n",
-      "|    total_timesteps      | 10240      |\n",
-      "| train/                  |            |\n",
-      "|    approx_kl            | 0.02928267 |\n",
-      "|    clip_fraction        | 0.498      |\n",
-      "|    clip_range           | 0.2        |\n",
-      "|    entropy_loss         | -9.46      |\n",
-      "|    explained_variance   | 0.167      |\n",
-      "|    learning_rate        | 0.0003     |\n",
-      "|    loss                 | 127        |\n",
-      "|    n_updates            | 40         |\n",
-      "|    policy_gradient_loss | -0.116     |\n",
-      "|    value_loss           | 207        |\n",
-      "----------------------------------------\n",
-      "-----------------------------------------\n",
-      "| rollout/                |             |\n",
-      "|    ep_len_mean          | 6           |\n",
-      "|    ep_rew_mean          | 1.62        |\n",
-      "| time/                   |             |\n",
-      "|    fps                  | 546         |\n",
-      "|    iterations           | 6           |\n",
-      "|    time_elapsed         | 22          |\n",
-      "|    total_timesteps      | 12288       |\n",
-      "| train/                  |             |\n",
-      "|    approx_kl            | 0.028425258 |\n",
-      "|    clip_fraction        | 0.483       |\n",
-      "|    clip_range           | 0.2         |\n",
-      "|    entropy_loss         | -9.46       |\n",
-      "|    explained_variance   | 0.143       |\n",
-      "|    learning_rate        | 0.0003      |\n",
-      "|    loss                 | 109         |\n",
-      "|    n_updates            | 50          |\n",
-      "|    policy_gradient_loss | -0.117      |\n",
-      "|    value_loss           | 240         |\n",
-      "-----------------------------------------\n",
-      "-----------------------------------------\n",
-      "| rollout/                |             |\n",
-      "|    ep_len_mean          | 5.98        |\n",
-      "|    ep_rew_mean          | 6.14        |\n",
-      "| time/                   |             |\n",
-      "|    fps                  | 541         |\n",
-      "|    iterations           | 7           |\n",
-      "|    time_elapsed         | 26          |\n",
-      "|    total_timesteps      | 14336       |\n",
-      "| train/                  |             |\n",
-      "|    approx_kl            | 0.026178032 |\n",
-      "|    clip_fraction        | 0.453       |\n",
-      "|    clip_range           | 0.2         |\n",
-      "|    entropy_loss         | -9.46       |\n",
-      "|    explained_variance   | 0.174       |\n",
-      "|    learning_rate        | 0.0003      |\n",
-      "|    loss                 | 141         |\n",
-      "|    n_updates            | 60          |\n",
-      "|    policy_gradient_loss | -0.116      |\n",
-      "|    value_loss           | 235         |\n",
-      "-----------------------------------------\n",
-      "----------------------------------------\n",
-      "| rollout/                |            |\n",
-      "|    ep_len_mean          | 6          |\n",
-      "|    ep_rew_mean          | 3.03       |\n",
-      "| time/                   |            |\n",
-      "|    fps                  | 537        |\n",
-      "|    iterations           | 8          |\n",
-      "|    time_elapsed         | 30         |\n",
-      "|    total_timesteps      | 16384      |\n",
-      "| train/                  |            |\n",
-      "|    approx_kl            | 0.02457074 |\n",
-      "|    clip_fraction        | 0.423      |\n",
-      "|    clip_range           | 0.2        |\n",
-      "|    entropy_loss         | -9.45      |\n",
-      "|    explained_variance   | 0.171      |\n",
-      "|    learning_rate        | 0.0003     |\n",
-      "|    loss                 | 111        |\n",
-      "|    n_updates            | 70         |\n",
-      "|    policy_gradient_loss | -0.112     |\n",
-      "|    value_loss           | 212        |\n",
-      "----------------------------------------\n",
-      "-----------------------------------------\n",
-      "| rollout/                |             |\n",
-      "|    ep_len_mean          | 6           |\n",
-      "|    ep_rew_mean          | 9.54        |\n",
-      "| time/                   |             |\n",
-      "|    fps                  | 532         |\n",
-      "|    iterations           | 9           |\n",
-      "|    time_elapsed         | 34          |\n",
-      "|    total_timesteps      | 18432       |\n",
-      "| train/                  |             |\n",
-      "|    approx_kl            | 0.024578478 |\n",
-      "|    clip_fraction        | 0.417       |\n",
-      "|    clip_range           | 0.2         |\n",
-      "|    entropy_loss         | -9.45       |\n",
-      "|    explained_variance   | 0.178       |\n",
-      "|    learning_rate        | 0.0003      |\n",
-      "|    loss                 | 121         |\n",
-      "|    n_updates            | 80          |\n",
-      "|    policy_gradient_loss | -0.114      |\n",
-      "|    value_loss           | 232         |\n",
-      "-----------------------------------------\n",
-      "-----------------------------------------\n",
-      "| rollout/                |             |\n",
-      "|    ep_len_mean          | 6           |\n",
-      "|    ep_rew_mean          | 3.81        |\n",
-      "| time/                   |             |\n",
-      "|    fps                  | 527         |\n",
-      "|    iterations           | 10          |\n",
-      "|    time_elapsed         | 38          |\n",
-      "|    total_timesteps      | 20480       |\n",
-      "| train/                  |             |\n",
-      "|    approx_kl            | 0.022704324 |\n",
-      "|    clip_fraction        | 0.379       |\n",
-      "|    clip_range           | 0.2         |\n",
-      "|    entropy_loss         | -9.45       |\n",
-      "|    explained_variance   | 0.194       |\n",
-      "|    learning_rate        | 0.0003      |\n",
-      "|    loss                 | 108         |\n",
-      "|    n_updates            | 90          |\n",
-      "|    policy_gradient_loss | -0.112      |\n",
-      "|    value_loss           | 216         |\n",
-      "-----------------------------------------\n"
-     ]
-    },
-    {
-     "data": {
-      "text/plain": [
-       "<stable_baselines3.ppo.ppo.PPO at 0x7f86ef4ddcd0>"
-      ]
-     },
-     "execution_count": 3,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
    "source": [
-    "total_timesteps = 20_000\n",
+    "total_timesteps = 100000\n",
-    "model = PPO(\"MlpPolicy\", env, verbose=1, device='cuda')\n",
+    "model = DQN(\"MlpPolicy\", env, verbose=0)\n",
-    "model.learn(total_timesteps=total_timesteps)"
+    "model.learn(total_timesteps=total_timesteps, progress_bar=True)"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def test(model):\n",
+    "\n",
+    "    end_rewards = []\n",
+    "\n",
+    "    for i in range(1000):\n",
+    "        \n",
+    "        state = env.reset()\n",
+    "\n",
+    "        done = False\n",
+    "\n",
+    "        while not done:\n",
+    "\n",
+    "            action, _states = model.predict(state, deterministic=True)\n",
+    "\n",
+    "            state, reward, done, info = env.step(action)\n",
+    "            \n",
+    "        end_rewards.append(reward == 0)\n",
+    "        \n",
+    "    return np.sum(end_rewards) / len(end_rewards)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -272,69 +87,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
-    "model = PPO.load(\"dqn_wordle\")"
+    "model = DQN.load(\"dqn_wordle\")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 7,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "100%|██████████| 1000/1000 [00:03<00:00, 252.17it/s]"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "[[ 7 18  1 19 16  3  3  3  2  3]\n",
-      " [16  9  5 14  4  3  3  3  3  3]\n",
-      " [16  9  5 14  4  3  3  3  3  3]\n",
-      " [16  9  5 14  4  3  3  3  3  3]\n",
-      " [ 7 18  1 19 16  3  3  3  2  3]\n",
-      " [ 7 18  1 19 16  3  3  3  2  3]] -54 {'correct': False, 'guesses': defaultdict(<class 'int'>, {'grasp': 3, 'piend': 3})}\n",
-      "0\n"
-     ]
-    },
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "\n"
-     ]
-    }
-   ],
-   "source": [
-    "env = gym_wordle.wordle.WordleEnv()\n",
-    "\n",
-    "for i in tqdm(range(1000)):\n",
-    "        \n",
-    "    state, info = env.reset()\n",
-    "\n",
-    "    done = False\n",
-    "\n",
-    "    wins = 0\n",
-    "\n",
-    "    while not done:\n",
-    "\n",
-    "        action, _states = model.predict(state, deterministic=True)\n",
-    "\n",
-    "        state, reward, done, truncated, info = env.step(action)\n",
-    "\n",
-    "    if info[\"correct\"]:\n",
-    "        wins += 1\n",
-    "\n",
-    "print(state, reward, info)\n",
-    "\n",
-    "print(wins)\n"
    ]
   },
   {
@@ -342,7 +99,9 @@
    "execution_count": null,
    "metadata": {},
    "outputs": [],
-   "source": []
+   "source": [
+    "print(test(model))"
+   ]
   }
  ],
  "metadata": {
@@ -361,7 +120,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.8.10"
+   "version": "3.11.5"
   }
  },
  "nbformat": 4,