Documentation Update (#261)

Co-authored-by: Mark Towers <mark.m.towers@gmail.com>

.github/workflows/gh-pages.yml

@@ -16,13 +16,13 @@ jobs:
       - uses: actions/setup-python@v4
 
       - name: Install dependencies
-        run: pip install -r docs/requirements.txt && git clone https://github.com/Farama-Foundation/MiniGrid.git && pip install ./MiniGrid
+        run: pip install -r docs/requirements.txt
 
       - name: Register Envs
-        run: cd MiniGrid && pip install -e .
+        run: pip install -e .
 
       - name: Build Envs Docs
-        run: python docs/scripts/gen_mds.py && python docs/scripts/gen_index.py
+        run: python docs/scripts/gen_mds.py
 
       - name: Build
         run: sphinx-build -b dirhtml -v docs _build

.gitignore

@@ -16,5 +16,5 @@ _site
 __pycache__
 .vscode/
 /docs/environments/*.md
-/docs/content/*.md
-/docs/api/*.md
+!docs/environments/index.md
+!docs/environments/babyAI_index.md

README.md

@@ -30,7 +30,6 @@ Please use this bibtex if you want to cite this repository in your publications:
   howpublished = {\url{https://github.com/Farama-Foundation/MiniGrid}},
 }
 ```
-<br>
 
 List of publications & submissions using Minigrid or BabyAI (please open a pull request to add missing entries):
 - [History Compression via Language Models in Reinforcement Learning.](https://proceedings.mlr.press/v162/paischer22a.html) (Johannes Kepler University Linz, PMLR 2022)
@@ -75,7 +74,6 @@ List of publications & submissions using Minigrid or BabyAI (please open a pull
 - [BabyAI: First Steps Towards Grounded Language Learning With a Human In the Loop](https://arxiv.org/abs/1810.08272) (Mila, ICLR, Oct 2018)
 
 This environment has been built as part of work done at [Mila](https://mila.quebec). The Dynamic obstacles environment has been added as part of work done at [IAS in TU Darmstadt](https://www.ias.informatik.tu-darmstadt.de/) and the University of Genoa for mobile robot navigation with dynamic obstacles.
-<br>
 
 ## Installation
 
@@ -92,7 +90,6 @@ git clone https://github.com/Farama-Foundation/MiniGrid
 cd MiniGrid
 pip install -e .
 ```
-<br>
 
 ## Basic Usage
 
@@ -119,7 +116,6 @@ A sample training command is:
 cd torch-rl
 python3 -m scripts.train --env MiniGrid-Empty-8x8-v0 --algo ppo
 ```
-<br>
 
 ## Wrappers
 
@@ -149,7 +145,6 @@ env = RGBImgPartialObsWrapper(env) # Get pixel observations
 env = ImgObsWrapper(env) # Get rid of the 'mission' field
 obs, _ = env.reset() # This now produces an RGB tensor only
 ```
-<br>
 
 ## Design
 
@@ -184,7 +179,6 @@ You can define your own reward function by creating a class derived
 from `MiniGridEnv`. Extending the environment with new object types or new actions
 should be very easy. If you wish to do this, you should take a look at the
 [minigrid/minigrid.py](minigrid/minigrid.py) source file.
-<br>
 
 ## Included Environments
 
@@ -192,7 +186,6 @@ The environments listed below are implemented in the [minigrid/envs](/minigrid/e
 Each environment provides one or more configurations registered with OpenAI gym. Each environment
 is also programmatically tunable in terms of size/complexity, which is useful for curriculum learning
 or to fine-tune difficulty.
-<br>
 
 ### Empty environment
 

docs/.gitignore

@@ -1,9 +0,0 @@
-.DS_Store
-_site
-.jekyll-cache
-__pycache__
-_build/
-.vscode/
-
-environments/
-!environments/index.md

docs/_templates/base.html

@@ -1,42 +0,0 @@
-{% extends "furo/base.html" %}
-
-{%- block regular_scripts -%}
-{{ super() }}
-
-<script>
-    (() => {
-        if (!localStorage.getItem("shownCookieAlert")) {
-            const boxElem = document.createElement("div");
-            boxElem.classList.add("cookie_alert");
-            const containerElem = document.createElement("div");
-            containerElem.classList.add("cookie_container");
-            const textElem = document.createElement("p");
-            textElem.innerHTML = `This page uses <a href="https://analytics.google.com/">
-                                Google Analytics</a> to collect statistics. You can disable it by blocking
-                                the JavaScript coming from www.google-analytics.com.`;
-                                containerElem.appendChild(textElem);
-            const closeBtn = document.createElement("button");
-            closeBtn.innerHTML = `<?xml version="1.0" ?><svg viewBox="0 0 32 32" xmlns="http://www.w3.org/2000/svg"><defs><style>.cls-1{fill:none;stroke:#000;stroke-linecap:round;stroke-linejoin:round;stroke-width:2px;}</style></defs><title/><g id="cross"><line class="cls-1" x1="7" x2="25" y1="7" y2="25"/><line class="cls-1" x1="7" x2="25" y1="25" y2="7"/></g></svg>`
-            closeBtn.onclick = () => {
-                localStorage.setItem("shownCookieAlert", "true");
-                boxElem.style.display = "none";
-            }
-            containerElem.appendChild(closeBtn);
-            boxElem.appendChild(containerElem);
-            document.body.appendChild(boxElem);
-        }
-    })()
-
-</script>
-
-<!-- Google tag (gtag.js) -->
-<script async src="https://www.googletagmanager.com/gtag/js?id=G-JGXSLW7N06"></script>
-<script>
-  window.dataLayer = window.dataLayer || [];
-  function gtag(){dataLayer.push(arguments);}
-  gtag('js', new Date());
-
-  gtag('config', 'G-JGXSLW7N06');
-</script>
-
-{%- endblock regular_scripts -%}

docs/api/wrapper.md

@@ -0,0 +1,39 @@
+---
+title: Wrapper
+lastpage:
+---
+
+## Wrapper
+Wraps an environment to allow a modular transformation of the :meth:`step` and :meth:`reset` methods. This class is the base class for all wrappers. The subclass could override some methods to change the behavior of the original environment without touching the original code. Note: Don't forget to call ``super().__init__(env)`` if the subclass overrides :meth:`__init__`.
+
+MiniGrid is built to support tasks involving natural language and sparse rewards.
+The observations are dictionaries, with an 'image' field, partially observable
+view of the environment, a 'mission' field which is a textual string
+describing the objective the agent should reach to get a reward, and a 'direction'
+field which can be used as an optional compass. Using dictionaries makes it
+easy for you to add additional information to observations
+if you need to, without having to encode everything into a single tensor.
+
+There are a variety of wrappers to change the observation format available in [minigrid/wrappers.py](/minigrid/wrappers.py). 
+If your RL code expects one single tensor for observations, take a look at `FlatObsWrapper`. 
+There is also an `ImgObsWrapper` that gets rid of the 'mission' field in observations, leaving only the image field tensor.
+
+Please note that the default observation format is a partially observable view of the environment using a
+compact and efficient encoding, with 3 input values per visible grid cell, 7x7x3 values total.
+These values are **not pixels**. If you want to obtain an array of RGB pixels as observations instead,
+use the `RGBImgPartialObsWrapper`. You can use it as follows:
+
+```python
+import gymnasium as gym
+from minigrid.wrappers import RGBImgPartialObsWrapper, ImgObsWrapper
+
+env = gym.make('MiniGrid-Empty-8x8-v0')
+env = RGBImgPartialObsWrapper(env) # Get pixel observations
+env = ImgObsWrapper(env) # Get rid of the 'mission' field
+obs, _ = env.reset() # This now produces an RGB tensor only
+```
+
+```{toctree}
+:hidden:
+wrappers
+```

docs/api/wrappers.md

@@ -0,0 +1,77 @@
+---
+title: Wrappers
+tocdepth: 2
+lastpage:
+---
+
+# Action Bonus
+
+```{eval-rst}
+.. autoclass:: minigrid.wrappers.ActionBonus
+```
+
+# Dict Observation Space
+
+```{eval-rst}
+.. autoclass:: minigrid.wrappers.DictObservationSpaceWrapper
+```
+
+# Direction Obs
+
+```{eval-rst}
+.. autoclass:: minigrid.wrappers.DirectionObsWrapper
+```
+
+# FlatObs
+
+```{eval-rst}
+.. autoclass:: minigrid.wrappers.FlatObsWrapper
+```
+
+# Fully Obs
+
+```{eval-rst}
+.. autoclass:: minigrid.wrappers.FullyObsWrapper
+```
+
+# Observation
+
+```{eval-rst}
+.. autoclass:: minigrid.wrappers.ObservationWrapper
+```
+
+# One Hot Partial Obs
+
+```{eval-rst}
+.. autoclass:: minigrid.wrappers.OneHotPartialObsWrapper
+```
+
+# Reseed
+
+```{eval-rst}
+.. autoclass:: minigrid.wrappers.ReseedWrapper
+```
+
+# RGB Img Obs
+
+```{eval-rst}
+.. autoclass:: minigrid.wrappers.RGBImgObsWrapper
+```
+
+# State Bonus
+
+```{eval-rst}
+.. autoclass:: minigrid.wrappers.StateBonus
+```
+
+# Symbolic Obs
+
+```{eval-rst}
+.. autoclass:: minigrid.wrappers.SymbolicObsWrapper
+```
+
+# View Size
+
+```{eval-rst}
+.. autoclass:: minigrid.wrappers.ViewSizeWrapper
+```

docs/conf.py

@@ -43,7 +43,6 @@ extensions = [
     "sphinx.ext.autodoc",
     "sphinx.ext.githubpages",
     "myst_parser",
-    "notfound.extension",
 ]
 
 source_suffix = {
@@ -90,5 +89,3 @@ html_context["slug"] = "minigrid"
 
 html_static_path = ["_static"]
 html_css_files = []
-
-notfound_urls_prefix = None

docs/content/basic_usage.md

@@ -19,7 +19,7 @@ The environment being run can be selected with the `--env` option, eg:
 ./minigrid/manual_control.py --env MiniGrid-Empty-8x8-v0
 ```
 
-## Reinforcement Learning
+## Training an Agent
 
 If you want to train an agent with reinforcement learning, I recommend using the code found in the [torch-rl](https://github.com/lcswillems/torch-rl) repository. 
 This code has been tested and is known to work with this environment. The default hyper-parameters are also known to converge.

docs/content/pubs.md

@@ -3,7 +3,7 @@ layout: "contents"
 title: Publications
 firstpage:
 ---
-#List of Publications
+# List of Publications
 
 
 List of publications & submissions using Minigrid or BabyAI (please open a pull request to add missing entries):

docs/environments/babyAI_index.md

@@ -0,0 +1,46 @@
+---
+firstpage:
+lastpage:
+---
+
+## BabyAI Environments
+
+```{toctree}
+:hidden:
+:caption: BabyAI Environments
+action_obj_door
+boss_level
+find_obj
+go_to_imp_unlock
+go_to_local
+go_to_obj
+go_to_obj_door
+go_to_obj_maze
+go_to_red_ball
+go_to_red_ball_grey
+go_to_red_ball_no_dists
+go_to_red_blue_ball
+go_to_seq
+key_in_box
+mini_boss_level
+move_two_across
+one_room
+open
+open_door_loc
+open_doors_order
+open_red_blue_doors_debug
+open_red_door
+pickup
+pickup_above
+pickup_dist_debug
+pickup_loc
+put_next
+put_next_local
+synth
+synth_loc
+synth_seq
+unblock_pickup
+unlock_local_dist
+unlock_pickup_dist
+unlock_to_unlock
+```

docs/environments/index.md

@@ -0,0 +1,39 @@
+---
+firstpage:
+lastpage:
+---
+
+## Included Environments
+
+The environments listed below are implemented in the [minigrid/envs](/minigrid/envs) directory.
+Each environment provides one or more configurations registered with OpenAI gym. Each environment
+is also programmatically tunable in terms of size/complexity, which is useful for curriculum learning
+or to fine-tune difficulty.
+
+```{toctree}
+:hidden:
+:caption: Minigrid Environments
+boss_level_no_unlock
+blocked_unlock_pickup
+simple_crossing
+dist_shift
+door_key
+dynamic
+empty
+fetch
+four_rooms
+go_to_door
+go_to_object
+key_corridor
+lava_gap
+locked_room
+memory
+multi_room
+obstructed_maze
+playground
+put_near
+red_blue_doors
+unlock
+unlock_pickup
+
+```

docs/index.md

@@ -3,10 +3,9 @@ hide-toc: true
 firstpage:
 lastpage:
 ---
+## Minigrid (formerly gym-minigrid) contains simple and easily configurable grid world environments for reinforcement learning
 
-# MiniGrid is a simple and easily configurable grid world environments for reinforcement learning
-
-[![pre-commit](https://img.shields.io/badge/pre--commit-enabled-brightgreen?logo=pre-commit&logoColor=white)](https://pre-commit.com/)
+[![pre-commit](https://img.shields.io/badge/pre--commit-enabled-brightgreen?logo=pre-commit&logoColor=white)](https://pre-commit.com/) 
 [![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
 
 There are other gridworld Gymnasium environments out there, but this one is
@@ -17,11 +16,10 @@ laptop, which means you can run your experiments faster. A known-working RL
 implementation can be found [in this repository](https://github.com/lcswillems/torch-rl).
 
 Requirements:
-
--   Python 3.7 to 3.10
--   OpenAI Gym v0.26
--   NumPy 1.18+
--   Matplotlib (optional, only needed for display) - 3.0+
+- Python 3.7 to 3.10
+- Gymnasium v0.26
+- NumPy 1.18+
+- Matplotlib (optional, only needed for display) - 3.0+
 
 Please use this bibtex if you want to cite this repository in your publications:
 
@@ -57,18 +55,25 @@ pip install -e .
 :hidden:
 :caption: Introduction
 
-content/installation
 content/basic_usage
-api/wrappers
-content/pubs
+content/publications
+```
+
+```{toctree}
+:hidden:
+:caption: Wrappers
+
+api/wrapper
 ```
 
+
 ```{toctree}
 :hidden:
 :caption: Environments
 
 environments/design
 environments/index
+environments/babyAI_index
 ```
 
 ```{toctree}

docs/requirements.txt

@@ -3,5 +3,4 @@ sphinx-notfound-page
 myst-parser
 git+https://github.com/Farama-Foundation/Celshast#egg=furo
 moviepy
-pygame
-pygments==2.11.2
+pygame

docs/scripts/gen_index.py

@@ -1,83 +0,0 @@
-__author__ = "Feng Gu"
-__email__ = "contact@fenggu.me"
-
-"""
-   isort:skip_file
-"""
-
-import os
-
-readme_path = os.path.join(
-    os.path.dirname(os.path.dirname(os.path.dirname(__file__))),
-    "README.md",
-)
-
-output_path = os.path.join(
-    os.path.dirname(os.path.dirname(__file__)),
-    "index.md",
-)
-
-sections = []
-all_text = """---
-hide-toc: true
-firstpage:
-lastpage:
----\n"""
-
-index_toctree = """
-```{toctree}
-:hidden:
-:caption: Introduction
-
-content/installation
-content/basic_usage
-content/pubs
-```
-
-```{toctree}
-:hidden:
-:caption: Wrappers
-
-api/wrappers
-```
-
-
-```{toctree}
-:hidden:
-:caption: Environments
-
-environments/design
-environments/index
-```
-
-```{toctree}
-:hidden:
-:caption: Development
-
-Github <https://github.com/Farama-Foundation/MiniGrid>
-Donate <https://farama.org/donations>
-Contribute to the Docs <https://github.com/Farama-Foundation/MiniGrid/blob/master/.github/PULL_REQUEST_TEMPLATE.md>
-```
-
-"""
-# gen index.md
-with open(readme_path) as f:
-    readme = f.read()
-
-    """
-    sections = [description, publications, installation, basic usage, wrappers, design, included environments&etc]
-    """
-    sections = readme.split("<br>")
-    all_text += sections[0]
-    all_text += sections[2]
-all_text += index_toctree
-
-with open(output_path, "w") as f:
-    f.write(all_text)
-
-
-"""
-1. gen index.md
-2. gen /environments/index.md
-3. gen /environments/design.md
-"""

docs/scripts/gen_mds.py

@@ -1,5 +1,3 @@
-import minigrid.wrappers
-
 __author__ = "Feng Gu"
 __email__ = "contact@fenggu.me"
 
@@ -7,13 +5,13 @@ __email__ = "contact@fenggu.me"
    isort:skip_file
 """
 
-import inspect
 import os
 import re
 
 from gymnasium.envs.registration import registry
 from tqdm import tqdm
 from utils import trim
+from itertools import chain
 
 readme_path = os.path.join(
     os.path.dirname(os.path.dirname(os.path.dirname(__file__))),
@@ -28,6 +26,7 @@ all_envs = list(registry.values())
 
 filtered_envs_by_type = {}
 env_names = []
+babyai_envs = {}
 
 # Obtain filtered list
 for env_spec in tqdm(all_envs):
@@ -35,11 +34,18 @@ for env_spec in tqdm(all_envs):
     split = env_spec.entry_point.split(".")
     # ignore gymnasium.envs.env_type:Env
     env_module = split[0]
-    if env_module != "minigrid":
+
+    if len(split) > 2 and "babyai" in split[2]:
+        curr_babyai_env = split[2]
+        babyai_env_name = curr_babyai_env.split(":")[1]
+        babyai_envs[babyai_env_name] = env_spec
+    elif env_module == "minigrid":
+        env_name = split[1]
+        filtered_envs_by_type[env_name] = env_spec
+    # if env_module != "minigrid":
+    else:
         continue
 
-    env_name = split[1]
-    filtered_envs_by_type[env_name] = env_spec
 
 filtered_envs = {
     env[0]: env[1]
@@ -49,7 +55,15 @@ filtered_envs = {
     )
 }
 
-for env_name, env_spec in filtered_envs.items():
+filtered_babyai_envs = {
+    env[0]: env[1]
+    for env in sorted(
+        babyai_envs.items(),
+        key=lambda item: item[1].entry_point.split(".")[1],
+    )
+}
+
+for env_name, env_spec in chain(filtered_envs.items(), filtered_babyai_envs.items()):
     made = env_spec.make()
 
     docstring = trim(made.unwrapped.__doc__)
@@ -87,136 +101,3 @@ title: {title_env_name}
     file = open(v_path, "w+", encoding="utf-8")
     file.write(all_text)
     file.close()
-
-
-# gen /environments/index.md
-index_texts = """---
-firstpage:
-lastpage:
----
-
-"""
-env_index_toctree = """
-```{toctree}
-:hidden:
-"""
-sections = []
-
-with open(readme_path) as f:
-    readme = f.read()
-
-    """
-    sections = [description, publications, installation, basic usage, wrappers, design, included environments&etc]
-    """
-    sections = readme.split("<br>")
-    index_texts += sections[6]
-    index_texts += env_index_toctree
-
-    for env_name in env_names:
-        index_texts += env_name + "\n"
-
-    index_texts += """\n```\n"""
-    f.close()
-
-output_path = os.path.join(
-    os.path.dirname(os.path.dirname(__file__)),
-    "environments",
-    "index.md",
-)
-
-# output index.md
-with open(output_path, "w+") as f:
-    f.write(index_texts)
-    f.close()
-
-# gen /environments/design.md
-design_path = os.path.join(
-    os.path.dirname(os.path.dirname(__file__)),
-    "environments",
-    "design.md",
-)
-
-design_texts = """---
-layout: "contents"
-title: Design
-firstpage:
----\n"""
-
-design_texts += sections[5]
-
-with open(design_path, "w+") as f:
-    f.write(design_texts)
-    f.close()
-
-
-# gen /environments/wrappers.md
-
-wrappers_path = os.path.join(
-    os.path.dirname(os.path.dirname(__file__)),
-    "api",
-    "wrappers.md",
-)
-
-wrappers_texts = (
-    """---
-title: Wrappers
-lastpage:
----\n"""
-    + sections[4]
-    + "\n"
-)
-
-for name, obj in inspect.getmembers(minigrid.wrappers):
-    if inspect.isclass(obj) and obj.__doc__ is not None:
-        formatted_doc = " ".join(trim(obj.__doc__).split())
-        wrappers_texts += f"""## {name}
-{formatted_doc}\n\n"""
-
-with open(wrappers_path, "w+") as f:
-    f.write(wrappers_texts)
-    f.close()
-
-
-# gen content/pubs.md
-
-pubs_path = os.path.join(
-    os.path.dirname(os.path.dirname(__file__)),
-    "content",
-    "pubs.md",
-)
-
-pubs_texts = (
-    """---
-layout: "contents"
-title: Publications
-firstpage:
----\n#List of Publications\n"""
-    + sections[1]
-    + "\n"
-)
-
-with open(pubs_path, "w+") as f:
-    f.write(pubs_texts)
-    f.close()
-
-# gen content/basic_usage.md
-
-pubs_path = os.path.join(
-    os.path.dirname(os.path.dirname(__file__)),
-    "content",
-    "basic_usage.md",
-)
-
-pubs_texts = (
-    """---
-layout: "contents"
-title: Basic Usage
-firstpage:
----\n"""
-    + sections[3]
-    + "\n"
-)
-
-with open(pubs_path, "w+") as f:
-    f.write(pubs_texts)
-    f.close()

minigrid/envs/blockedunlockpickup.py

@@ -7,6 +7,8 @@ from minigrid.core.world_object import Ball
 class BlockedUnlockPickupEnv(RoomGrid):
 
     """
+    ![BlockedUnlockPickup](../_static/figures/BlockedUnlockPickup.png)
+
     ### Description
 
     The agent has to pick up a box which is placed in another room, behind a

minigrid/envs/crossing.py

@@ -11,6 +11,14 @@ from minigrid.minigrid_env import MiniGridEnv
 class CrossingEnv(MiniGridEnv):
 
     """
+    ![LavaCrossingS9N1](../_static/figures/LavaCrossingS9N1.png)<br />
+    ![LavaCrossingS9N2](../_static/figures/LavaCrossingS9N2.png)<br />
+    ![LavaCrossingS9N3](../_static/figures/LavaCrossingS9N3.png)<br />
+    ![LavaCrossingS11N5](../_static/figures/LavaCrossingS11N5.png)<br />
+    ![SimpleCrossingS9N1](../_static/figures/SimpleCrossingS9N1.png)<br />
+    ![SimpleCrossingS9N2](../_static/figures/SimpleCrossingS9N2.png)<br />
+    ![SimpleCrossingS9N3](../_static/figures/SimpleCrossingS9N3.png)<br />
+    ![SimpleCrossingS11N5](../_static/figures/SimpleCrossingS11N5.png)<br />
     ### Description
 
     Depending on the `obstacle_type` parameter:

minigrid/envs/distshift.py

@@ -7,10 +7,13 @@ from minigrid.minigrid_env import MiniGridEnv
 class DistShiftEnv(MiniGridEnv):
 
     """
+    ![DistShift1](../_static/figures/DistShift1.png)<br />
+    ![DistShift2](../_static/figures/DistShift2.png)
+
     ### Description
 
-    This environment is based on one of the DeepMind [AI safety gridworlds]
-    (https://github.com/deepmind/ai-safety-gridworlds). The agent starts in the
+    This environment is based on one of the DeepMind [AI safety gridworlds](https://github.com/deepmind/ai-safety-gridworlds).
+    The agent starts in the
     top-left corner and must reach the goal which is in the top-right corner,
     but has to avoid stepping into lava on its way. The aim of this environment
     is to test an agent's ability to generalize. There are two slightly

minigrid/envs/doorkey.py

@@ -7,6 +7,8 @@ from minigrid.minigrid_env import MiniGridEnv
 class DoorKeyEnv(MiniGridEnv):
 
     """
+    ![door-key-curriculum](../_static/figures/door-key-curriculum.gif)<br />
+    ![door-key-env](../_static/figures/door-key-env.png)
     ### Description
 
     This environment has a key that the agent must pick up in order to unlock a

minigrid/envs/dynamicobstacles.py

@@ -10,6 +10,8 @@ from minigrid.minigrid_env import MiniGridEnv
 
 class DynamicObstaclesEnv(MiniGridEnv):
     """
+    ![dynamic_obstacles](../_static/figures/dynamic_obstacles.gif)
+
     ### Description
 
     This environment is an empty room with moving obstacles.

minigrid/envs/empty.py

@@ -6,6 +6,8 @@ from minigrid.minigrid_env import MiniGridEnv
 
 class EmptyEnv(MiniGridEnv):
     """
+    ![empty-env](../_static/figures/empty-env.png)
+
     ### Description
 
     This environment is an empty room, and the goal of the agent is to reach the

minigrid/envs/fetch.py

@@ -8,6 +8,8 @@ from minigrid.minigrid_env import MiniGridEnv
 class FetchEnv(MiniGridEnv):
 
     """
+    ![fetch-env](../_static/figures/fetch-env.png)
+
     ### Description
 
     This environment has multiple objects of assorted types and colors. The

minigrid/envs/fourrooms.py

@@ -7,6 +7,8 @@ from minigrid.minigrid_env import MiniGridEnv
 class FourRoomsEnv(MiniGridEnv):
 
     """
+    ![four-rooms-env](../_static/figures/four-rooms-env.png)
+
     ### Description
 
     Classic four room reinforcement learning environment. The agent must

minigrid/envs/gotodoor.py

@@ -7,6 +7,8 @@ from minigrid.minigrid_env import MiniGridEnv
 
 class GoToDoorEnv(MiniGridEnv):
     """
+    ![gotodoor-6x6](../_static/figures/gotodoor-6x6.png)<br />
+    [Video](../_static/figures/gotodoor-6x6.mp4)
     ### Description
 
     This environment is a room with four doors, one on each wall. The agent

minigrid/envs/keycorridor.py

@@ -6,6 +6,13 @@ from minigrid.core.roomgrid import RoomGrid
 class KeyCorridorEnv(RoomGrid):
 
     """
+    ![KeyCorridorS3R1](../_static/figures/KeyCorridorS3R1.png)<br />
+    ![KeyCorridorS3R2](../_static/figures/KeyCorridorS3R2.png)<br />
+    ![KeyCorridorS3R3](../_static/figures/KeyCorridorS3R3.png)<br />
+    ![KeyCorridorS4R3](../_static/figures/KeyCorridorS4R3.png)<br />
+    ![KeyCorridorS5R3](../_static/figures/KeyCorridorS5R3.png)<br />
+    ![KeyCorridorS6R3](../_static/figures/KeyCorridorS6R3.png)
+
     ### Description
 
     This environment is similar to the locked room environment, but there are

minigrid/envs/lavagap.py

@@ -9,6 +9,8 @@ from minigrid.minigrid_env import MiniGridEnv
 class LavaGapEnv(MiniGridEnv):
 
     """
+    ![LavaGapS6](../_static/figures/LavaGapS6.png)
+
     ### Description
 
     The agent has to reach the green goal square at the opposite corner of the

minigrid/envs/multiroom.py

@@ -16,6 +16,8 @@ class MultiRoom:
 class MultiRoomEnv(MiniGridEnv):
 
     """
+    ![multi-room](../_static/figures/multi-room.gif)
+
     ### Description
 
     This environment has a series of connected rooms with doors that must be

minigrid/envs/obstructedmaze.py

@@ -7,6 +7,16 @@ from minigrid.core.world_object import Ball, Box, Key
 class ObstructedMazeEnv(RoomGrid):
 
     """
+    ![ObstructedMaze-1Dl](../_static/figures/ObstructedMaze-1Dl.png)<br />
+    ![ObstructedMaze-1Dlh](../_static/figures/ObstructedMaze-1Dlh.png)<br />
+    ![ObstructedMaze-1Dlhb](../_static/figures/ObstructedMaze-1Dlhb.png)<br />
+    ![ObstructedMaze-1Q](../_static/figures/ObstructedMaze-1Q.png)<br />
+    ![ObstructedMaze-2Dl](../_static/figures/ObstructedMaze-2Dl.png)<br />
+    ![ObstructedMaze-2Dlh](../_static/figures/ObstructedMaze-2Dlh.png)<br />
+    ![ObstructedMaze-2Dlhb](../_static/figures/ObstructedMaze-2Dlhb.png)<br />
+    ![ObstructedMaze-2Q](../_static/figures/ObstructedMaze-2Q.png)<br />
+    ![ObstructedMaze-4Q](../_static/figures/ObstructedMaze-4Q.png)<br />
+
     ### Description
 
     The agent has to pick up a box which is placed in a corner of a 3x3 maze.

minigrid/envs/unlock.py

@@ -5,6 +5,8 @@ from minigrid.core.roomgrid import RoomGrid
 class UnlockEnv(RoomGrid):
 
     """
+    ![Unlock](../_static/figures/Unlock.png)
+
     ### Description
 
     The agent has to open a locked door. This environment can be solved without

minigrid/envs/unlockpickup.py

@@ -6,6 +6,8 @@ from minigrid.core.roomgrid import RoomGrid
 class UnlockPickupEnv(RoomGrid):
 
     """
+    ![UnlockPickup](../_static/figures/UnlockPickup.png)
+
     ### Description
 
     The agent has to pick up a box which is placed in another room, behind a