diff --git a/MANIFEST.in b/MANIFEST.in
index 24a63fa..9808edb 100644
--- a/MANIFEST.in
+++ b/MANIFEST.in
@@ -1,4 +1,5 @@
 include requirements.txt
+include requirements-vis.txt
 include requirements-full.txt
 include README.md
 include LICENSE
diff --git a/README.md b/README.md
index 7e0e954..9867ba2 100644
--- a/README.md
+++ b/README.md
@@ -15,10 +15,23 @@ For the latest updates, see: [**alfworld.github.io**](https://alfworld.github.io
 
 ## Quickstart
 
+Create a virtual environment (recommended)
+
+    conda create -n alfworld python=3.9
+    conda activate alfworld
+
+> [!WARNING]
+> If you are using MacOS with an arm-based system, it is recommended to use
+>
+    CONDA_SUBDIR=osx-64 conda create -n alfworld python=3.9
+    conda activate alfworld
+
 Install with pip (python3.9+):
 
     pip install alfworld[full]
 
+> **Note:** Without the `full` extra, it will only install the text version of ALFWorld. To enable visual modalities, use `pip install alfworld[vis]`.
+
 Download PDDL & Game files and pre-trained MaskRCNN detector:
 ```bash
 export ALFWORLD_DATA=<storage_path>
@@ -104,12 +117,10 @@ Play around with [TextWorld and THOR demos](scripts/).
 ## Prerequisites
 
 - Python 3.9+
-- PyTorch 1.2.0 (later versions might be ok)
-- Torchvision 0.4.0 (later versions might be ok)
-- AI2THOR 2.1.0
 
-See [requirements.txt](requirements.txt) for the prerequisites to run ALFWorld.
-See [requirements-full.txt](requirements-full.txt) for the prerequisites to run experiments.
+See [requirements.txt](requirements.txt) for the prerequisites to run ALFWorld text-only version.
+See [requirements-vis.txt](requirements.txt) for the prerequisites to run ALFWorld with both text and visual modalities.
+See [requirements-full.txt](requirements-full.txt) for the full prerequisites to run experiments.
 
 ## Hardware
 
@@ -122,6 +133,9 @@ Tested on:
 
 ## Docker Setup
 
+> [!WARNING]
+> This docker setup has been tested for an older version of ALFWorld.
+
 Pull [vzhong](https://github.com/vzhong)'s image: https://hub.docker.com/r/vzhong/alfworld
 
 **OR**
@@ -212,10 +226,6 @@ You might have to modify `X_DISPLAY` in [gen/constants.py](alfworld/gen/constant
 
 Also, checkout this guide: [Setting up THOR on Google Cloud](https://medium.com/@etendue2013/how-to-run-ai2-thor-simulation-fast-with-google-cloud-platform-gcp-c9fcde213a4a)
 
-## Change Log
-
-18/12/2020:
-- PIP package version available. The repo was refactored.
 
 ## Citations
 
diff --git a/alfworld/agents/agent/base_agent.py b/alfworld/agents/agent/base_agent.py
index f8ecd11..2992b30 100644
--- a/alfworld/agents/agent/base_agent.py
+++ b/alfworld/agents/agent/base_agent.py
@@ -1,7 +1,12 @@
 import logging
 import numpy as np
-import torch
-from transformers import DistilBertModel, DistilBertTokenizer
+
+try:
+    import torch
+    from transformers import DistilBertModel, DistilBertTokenizer
+except ImportError:
+    raise ImportError("torch or transformers not found. Please install them via `pip install alfworld[full]`.")
+
 logging.getLogger("transformers.tokenization_utils").setLevel(logging.ERROR)
 
 import alfworld.agents.modules.memory as memory
diff --git a/alfworld/agents/agent/text_dagger_agent.py b/alfworld/agents/agent/text_dagger_agent.py
index c921d5e..9240a45 100644
--- a/alfworld/agents/agent/text_dagger_agent.py
+++ b/alfworld/agents/agent/text_dagger_agent.py
@@ -4,7 +4,10 @@
 from queue import PriorityQueue
 
 import numpy as np
-import torch
+try:
+    import torch
+except ImportError:
+    raise ImportError("torch not found. Please install them via `pip install alfworld[full]`.")
 logging.getLogger("transformers.tokenization_utils").setLevel(logging.ERROR)
 
 from alfworld.agents.agent import BaseAgent
@@ -249,7 +252,6 @@ def command_generation_greedy_generation(self, observation_strings, task_desc_st
                     break
             res = [self.tokenizer.decode(item) for item in input_target_list]
             res = [item.replace("[CLS]", "").replace("[SEP]", "").strip() for item in res]
-            res = [item.replace(" in / on ", " in/on " ) for item in res]
             return res, current_dynamics
 
     def command_generation_beam_search_generation(self, observation_strings, task_desc_strings, previous_dynamics):
@@ -348,7 +350,6 @@ def command_generation_beam_search_generation(self, observation_strings, task_de
                     utte_string = self.tokenizer.decode(utte)
                     utterances.append(utte_string)
                 utterances = [item.replace("[CLS]", "").replace("[SEP]", "").strip() for item in utterances]
-                utterances = [item.replace(" in / on ", " in/on " ) for item in utterances]
                 res.append(utterances)
             return res, current_dynamics
 
diff --git a/alfworld/agents/agent/text_dqn_agent.py b/alfworld/agents/agent/text_dqn_agent.py
index cfa99de..d6941f5 100644
--- a/alfworld/agents/agent/text_dqn_agent.py
+++ b/alfworld/agents/agent/text_dqn_agent.py
@@ -4,8 +4,12 @@
 from queue import PriorityQueue
 
 import numpy as np
-import torch
-import torch.nn.functional as F
+
+try:
+    import torch
+    import torch.nn.functional as F
+except ImportError:
+    raise ImportError("torch not found. Please install them via `pip install alfworld[full]`.")
 logging.getLogger("transformers.tokenization_utils").setLevel(logging.ERROR)
 
 from alfworld.agents.agent import BaseAgent
@@ -261,7 +265,6 @@ def command_generation_by_beam_search(self, observation_strings, task_desc_strin
                     utterances.append(utte_string)
 
                 utterances = [item.replace("[CLS]", "").replace("[SEP]", "").strip() for item in utterances]
-                utterances = [item.replace(" in / on ", " in/on " ) for item in utterances]
                 chosen_actions.append(utterances)
             return chosen_actions, current_dynamics, obs_mask, aggregated_obs_representation
 
@@ -302,7 +305,6 @@ def command_generation_act_greedy(self, observation_strings, task_desc_strings,
                     break
             chosen_actions = [self.tokenizer.decode(item) for item in input_target_list]
             chosen_actions = [item.replace("[CLS]", "").replace("[SEP]", "").strip() for item in chosen_actions]
-            chosen_actions = [item.replace(" in / on ", " in/on " ) for item in chosen_actions]
             chosen_indices = [item[1:] for item in input_target_list]
             for i in range(len(chosen_indices)):
                 if chosen_indices[i][-1] == self.word2id["[SEP]"]:
@@ -400,7 +402,6 @@ def command_generation_act_random(self, observation_strings, task_desc_strings,
                     indicies.append(utte)
 
                 utterances = [item.replace("[CLS]", "").replace("[SEP]", "").strip() for item in utterances]
-                utterances = [item.replace(" in / on ", " in/on " ) for item in utterances]
                 indicies = [item[1:] for item in indicies]
                 for i in range(len(indicies)):
                     if indicies[i][-1] == self.word2id["[SEP]"]:
diff --git a/alfworld/agents/agent/vision_dagger_agent.py b/alfworld/agents/agent/vision_dagger_agent.py
index 6502452..0052bd0 100644
--- a/alfworld/agents/agent/vision_dagger_agent.py
+++ b/alfworld/agents/agent/vision_dagger_agent.py
@@ -1,10 +1,11 @@
-import os
-import sys
 import copy
 
 import numpy as np
-import torch
-import torch.nn.functional as F
+try:
+    import torch
+    import torch.nn.functional as F
+except ImportError:
+    raise ImportError("torch not found. Please install them via `pip install alfworld[full]`.")
 
 import alfworld.agents
 import alfworld.agents.modules.memory as memory
diff --git a/alfworld/agents/controller/mrcnn.py b/alfworld/agents/controller/mrcnn.py
index 3ce75c0..34066fa 100644
--- a/alfworld/agents/controller/mrcnn.py
+++ b/alfworld/agents/controller/mrcnn.py
@@ -20,7 +20,10 @@
 from alfworld.agents.controller.base import BaseAgent
 from alfworld.agents.utils.misc import extract_admissible_commands_with_heuristics
 
-import torchvision.transforms as T
+try:
+    import torchvision.transforms as T
+except ImportError:
+    raise ImportError("torchvision not found. Please install them via `pip install alfworld[full]`.")
 
 class MaskRCNNAgent(BaseAgent):
 
diff --git a/alfworld/agents/environment/__init__.py b/alfworld/agents/environment/__init__.py
index 09fe688..e69de29 100644
--- a/alfworld/agents/environment/__init__.py
+++ b/alfworld/agents/environment/__init__.py
@@ -1,3 +0,0 @@
-from alfworld.agents.environment.alfred_tw_env import AlfredTWEnv
-from alfworld.agents.environment.alfred_thor_env import AlfredThorEnv
-from alfworld.agents.environment.alfred_hybrid import AlfredHybrid
\ No newline at end of file
diff --git a/alfworld/agents/environment/alfred_hybrid.py b/alfworld/agents/environment/alfred_hybrid.py
index 0598368..c745557 100644
--- a/alfworld/agents/environment/alfred_hybrid.py
+++ b/alfworld/agents/environment/alfred_hybrid.py
@@ -1,7 +1,7 @@
 import random
-import importlib
 
-import alfworld.agents.environment
+from alfworld.agents.environment.alfred_thor_env import AlfredThorEnv
+from alfworld.agents.environment.alfred_tw_env import AlfredTWEnv
 
 
 class AlfredHybrid(object):
@@ -28,12 +28,12 @@ def choose_env(self):
             return self.tw
 
     def init_env(self, batch_size):
-        AlfredTWEnv = getattr(alfworld.agents.environment, "AlfredTWEnv")(self.config, train_eval=self.train_eval)
-        AlfredThorEnv = getattr(alfworld.agents.environment, "AlfredThorEnv")(self.config, train_eval=self.train_eval)
+        alfred_tw_env = AlfredTWEnv(self.config, train_eval=self.train_eval)
+        alfred_thor_env = AlfredThorEnv(self.config, train_eval=self.train_eval)
 
         self.batch_size = batch_size
-        self.tw = AlfredTWEnv.init_env(batch_size)
-        self.thor = AlfredThorEnv.init_env(batch_size)
+        self.tw = alfred_tw_env.init_env(batch_size)
+        self.thor = alfred_thor_env.init_env(batch_size)
         return self
 
     def seed(self, num):
@@ -56,4 +56,4 @@ def reset(self):
         env = self.choose_env()
         obs, infos = env.reset()
         self.num_resets += self.batch_size
-        return obs, infos
\ No newline at end of file
+        return obs, infos
diff --git a/alfworld/agents/environment/alfred_thor_env.py b/alfworld/agents/environment/alfred_thor_env.py
index 65ee9df..01177c0 100644
--- a/alfworld/agents/environment/alfred_thor_env.py
+++ b/alfworld/agents/environment/alfred_thor_env.py
@@ -1,17 +1,15 @@
 import os
 import json
-import glob
 import numpy as np
 
 import traceback
 import threading
 from queue import Queue
 from threading import Thread
-import sys
 import random
 
 import alfworld.agents
-from alfworld.agents.utils.misc import Demangler, get_templated_task_desc, add_task_to_grammar
+from alfworld.agents.utils.misc import get_templated_task_desc
 from alfworld.env.thor_env import ThorEnv
 from alfworld.agents.expert import HandCodedThorAgent, HandCodedAgentTimeout
 from alfworld.agents.detector.mrcnn import load_pretrained_model
diff --git a/alfworld/agents/expert/handcoded_expert.py b/alfworld/agents/expert/handcoded_expert.py
index 6ac6219..9fa8c3d 100644
--- a/alfworld/agents/expert/handcoded_expert.py
+++ b/alfworld/agents/expert/handcoded_expert.py
@@ -236,7 +236,7 @@ def act(self, game_state, last_action):
                 # if holding something irrelavant, then discard it from where it was pickedup
                 if len(self.inventory) > 0 and not self.is_agent_holding_right_object:
                     if self.curr_recep == self.got_inventory_from_recep:
-                        return "put {} in/on {}".format(self.inventory[0], self.got_inventory_from_recep)
+                        return "move {} to {}".format(self.inventory[0], self.got_inventory_from_recep)
                     else:
                         return "go to {}".format(self.got_inventory_from_recep)
 
@@ -276,7 +276,7 @@ def act(self, game_state, last_action):
                 return "open {}".format(self.curr_recep)
             else:
                 obj = self.inventory[0]
-                return "put {} in/on {}".format(obj, self.curr_recep)
+                return "move {} to {}".format(obj, self.curr_recep)
 
         # OPEN
         if sub_action == 'open':
@@ -321,7 +321,7 @@ def act(self, game_state, last_action):
         # if holding something irrelavant, then discard it from where it was pickedup
         if len(self.inventory) > 0 and not self.is_agent_holding_right_object:
             if self.curr_recep == self.got_inventory_from_recep:
-                return "put {} in/on {}".format(self.inventory[0], self.got_inventory_from_recep)
+                return "move {} to {}".format(self.inventory[0], self.got_inventory_from_recep)
             else:
                 return "go to {}".format(self.got_inventory_from_recep)
 
diff --git a/alfworld/agents/expert/handcoded_expert_tw.py b/alfworld/agents/expert/handcoded_expert_tw.py
index d58ac21..da8bdc1 100644
--- a/alfworld/agents/expert/handcoded_expert_tw.py
+++ b/alfworld/agents/expert/handcoded_expert_tw.py
@@ -14,7 +14,7 @@ def get_predicates(self, game_state, obj, parent):
         obs_at_curr_recep = self.obs_at_recep[self.curr_recep] if self.curr_recep in self.obs_at_recep else ""
         is_obj_in_obs = "you see" in obs_at_curr_recep and " {} ".format(obj) in obs_at_curr_recep
         at_right_recep = parent in self.curr_recep
-        can_put_object = "put {} in/on {}".format(obj, parent) in admissible_commands_wo_num_ids
+        can_put_object = "move {} to {}".format(obj, parent) in admissible_commands_wo_num_ids
         can_take_object = any("take {}".format(obj) in ac for ac in admissible_commands_wo_num_ids)
         return at_right_recep, can_put_object, can_take_object, is_obj_in_obs
 
@@ -35,7 +35,7 @@ def get_predicates(self, game_state, obj, parent):
         obs_at_curr_recep = self.obs_at_recep[self.curr_recep] if self.curr_recep in self.obs_at_recep else ""
         is_obj_in_obs = "you see" in obs_at_curr_recep and " {} ".format(obj) in obs_at_curr_recep
         at_right_recep = parent in self.curr_recep
-        can_put_object = "put {} in/on {}".format(obj, parent) in admissible_commands_wo_num_ids
+        can_put_object = "move {} to {}".format(obj, parent) in admissible_commands_wo_num_ids
         can_take_object = any("take {}".format(obj) in ac for ac in admissible_commands_wo_num_ids)
         return at_right_recep, can_put_object, can_take_object, is_obj_in_obs, is_one_object_already_inside_receptacle, trying_to_take_the_same_object
 
@@ -68,7 +68,7 @@ def get_predicates(self, game_state, obj, parent):
         obs_at_curr_recep = self.obs_at_recep[self.curr_recep] if self.curr_recep in self.obs_at_recep else ""
         is_obj_in_obs = "you see" in obs_at_curr_recep and " {} ".format(obj) in obs_at_curr_recep
         at_right_recep = parent in self.curr_recep
-        can_put_object = "put {} in/on {}".format(obj, parent) in admissible_commands_wo_num_ids
+        can_put_object = "move {} to {}".format(obj, parent) in admissible_commands_wo_num_ids
         can_take_object = any("take {}".format(obj) in ac for ac in admissible_commands_wo_num_ids)
         can_heat_object = "heat {} with {}".format(obj, "microwave") in admissible_commands_wo_num_ids
         return at_right_recep, can_heat_object, can_put_object, can_take_object, is_obj_in_obs, is_the_object_agent_holding_hot
@@ -87,7 +87,7 @@ def get_predicates(self, game_state, obj, parent):
         obs_at_curr_recep = self.obs_at_recep[self.curr_recep] if self.curr_recep in self.obs_at_recep else ""
         is_obj_in_obs = "you see" in obs_at_curr_recep and " {} ".format(obj) in obs_at_curr_recep
         at_right_recep = parent in self.curr_recep
-        can_put_object = "put {} in/on {}".format(obj, parent) in admissible_commands_wo_num_ids
+        can_put_object = "move {} to {}".format(obj, parent) in admissible_commands_wo_num_ids
         can_cool_object = "cool {} with {}".format(obj, "fridge") in admissible_commands_wo_num_ids
         can_take_object = any("take {}".format(obj) in ac for ac in admissible_commands_wo_num_ids)
         return at_right_recep, can_cool_object, can_put_object, can_take_object, is_obj_in_obs, is_the_object_agent_holding_cool
@@ -106,7 +106,7 @@ def get_predicates(self, game_state, obj, parent):
         obs_at_curr_recep = self.obs_at_recep[self.curr_recep] if self.curr_recep in self.obs_at_recep else ""
         is_obj_in_obs = "you see" in obs_at_curr_recep and " {} ".format(obj) in obs_at_curr_recep
         at_right_recep = parent in self.curr_recep
-        can_put_object = "put {} in/on {}".format(obj, parent) in admissible_commands_wo_num_ids
+        can_put_object = "move {} to {}".format(obj, parent) in admissible_commands_wo_num_ids
         can_clean_object = "clean {} with {}".format(obj, "sinkbasin") in admissible_commands_wo_num_ids
         can_take_object = any("take {}".format(obj) in ac for ac in admissible_commands_wo_num_ids)
         return at_right_recep, can_clean_object, can_put_object, can_take_object, is_obj_in_obs, is_the_object_agent_holding_isclean
@@ -125,4 +125,4 @@ def get_task_policy(self, task_param):
         if task_class_str in globals():
             return globals()[task_class_str]
         else:
-            raise Exception("Invalid Task Type: %s" % task_type)
\ No newline at end of file
+            raise Exception("Invalid Task Type: %s" % task_type)
diff --git a/alfworld/agents/utils/misc.py b/alfworld/agents/utils/misc.py
index 632c21f..3d0297e 100644
--- a/alfworld/agents/utils/misc.py
+++ b/alfworld/agents/utils/misc.py
@@ -154,7 +154,7 @@ def extract_admissible_commands_with_heuristics(intro, frame_desc, feedback,
         "open {recep}",
         "close {recep}",
         "take {obj} from {recep}",
-        "put {obj} in/on {recep}",
+        "move {obj} to {recep}",
         "use {lamp}",
         "heat {obj} with {microwave}",
         "cool {obj} with {fridge}",
@@ -162,6 +162,7 @@ def extract_admissible_commands_with_heuristics(intro, frame_desc, feedback,
         "slice {obj} with {knife}",
         "inventory",
         "look",
+        "help",
         "examine {obj}",
         "examine {recep}"
     ]
@@ -182,7 +183,7 @@ def extract_admissible_commands_with_heuristics(intro, frame_desc, feedback,
                 for obj in objects:
                     if 'desklamp' not in obj and 'floorlamp' not in obj:
                         admissible_commands.append(t.format(recep=at_recep, obj=obj))
-        elif 'put {obj} in/on {recep}' in t:
+        elif 'move {obj} to {recep}' in t:
             if in_inv and at_recep:
                 admissible_commands.append(t.format(recep=at_recep, obj=in_inv))
         elif '{obj}' in t and '{microwave}' in t:
@@ -246,7 +247,7 @@ def extract_admissible_commands(intro, frame_desc):
         "open {recep}",
         "close {recep}",
         "take {obj} from {recep}",
-        "put {obj} in/on {recep}",
+        "move {obj} to {recep}",
         "use {lamp}",
         "heat {obj} with {microwave}",
         "cool {obj} with {fridge}",
@@ -254,6 +255,7 @@ def extract_admissible_commands(intro, frame_desc):
         "slice {obj} with {knife}",
         "inventory",
         "look",
+        "help",
         "examine {obj}",
         "examine {recep}"
     ]
diff --git a/alfworld/data/alfred.pddl b/alfworld/data/alfred.pddl
index b087c0e..9c0d812 100644
--- a/alfworld/data/alfred.pddl
+++ b/alfworld/data/alfred.pddl
@@ -495,4 +495,14 @@
  )
 
 
+(:action help
+    :parameters (?a - agent)
+    :precondition
+        ()
+    :effect
+        (and
+            (checked ?a)
+        )
+)
+
 )
diff --git a/alfworld/data/alfred.twl2 b/alfworld/data/alfred.twl2
index d391142..47a7c9f 100644
--- a/alfworld/data/alfred.twl2
+++ b/alfworld/data/alfred.twl2
@@ -20,7 +20,7 @@ grammar :: """
 
         "GotoLocation.feedback": [
             {
-                "rhs": "You arrive at {lend.name}. #examineReceptacle.feedback#"
+                "rhs": "You arrive at {r.name}. #examineReceptacle.feedback#"
             }
         ],
 
@@ -62,7 +62,7 @@ grammar :: """
 
         "PutObject.feedback": [
             {
-                "rhs": "You put the {o.name} in/on the {r.name}."
+                "rhs": "You move the {o.name} to the {r.name}."
             }
         ],
 
@@ -208,7 +208,7 @@ action PickupFullReceptacleObject {
 }
 
 action PutObject {
-    template :: "put {o} in/on {r}";
+    template :: "move {o} to {r}";
     feedback :: "#PutObject.feedback#";
 }
 
@@ -218,7 +218,7 @@ action PutObjectInReceptacleObject {
 }
 
 action PutEmptyReceptacleObjectinReceptacle {
-    template :: "put {o} in/on {r}";
+    template :: "move {o} to {r}";
     feedback :: "#PutEmptyReceptacleObjectinReceptacle.feedback#";
 }
 
@@ -420,3 +420,18 @@ action look {
         }
     """;
 }
+
+action help {
+    template :: "help";
+    feedback :: "#help.feedback#";
+
+    grammar :: """
+        {
+            "help.feedback": [
+                {
+                    "rhs": "\nAvailable commands:\n  look:                             look around your current location\n  inventory:                        check your current inventory\n  go to (receptacle):               move to a receptacle\n  open (receptacle):                open a receptacle\n  close (receptacle):               close a receptacle\n  take (object) from (receptacle):  take an object from a receptacle\n  move (object) to (receptacle):  place an object in or on a receptacle\n  examine (something):              examine a receptacle or an object\n  use (object):                     use an object\n  heat (object) with (receptacle):  heat an object using a receptacle\n  clean (object) with (receptacle): clean an object using a receptacle\n  cool (object) with (receptacle):  cool an object using a receptacle\n  slice (object) with (object):     slice an object using a sharp object\n"
+                }
+            ]
+        }
+    """;
+}
diff --git a/alfworld/env/thor_env.py b/alfworld/env/thor_env.py
index 3ee602f..4e3ad4e 100644
--- a/alfworld/env/thor_env.py
+++ b/alfworld/env/thor_env.py
@@ -1,10 +1,14 @@
-import cv2
 import copy
 import os
 import glob
 import numpy as np
 from collections import Counter, OrderedDict
-from ai2thor.controller import Controller
+
+try:
+    from ai2thor.controller import Controller
+    import cv2
+except ImportError:
+    raise ImportError("ai2thor or cv2 not found. Please install them via `pip install alfworld[vis]`.")
 
 import alfworld.gen.constants as constants
 from alfworld.env.tasks import get_task
diff --git a/alfworld/info.py b/alfworld/info.py
index f5be671..83a5066 100644
--- a/alfworld/info.py
+++ b/alfworld/info.py
@@ -1,4 +1,4 @@
-__version__ = '0.3.5'
+__version__ = '0.4.0'
 
 import os
 from os.path import join as pjoin
diff --git a/requirements-vis.txt b/requirements-vis.txt
new file mode 100644
index 0000000..fb91442
--- /dev/null
+++ b/requirements-vis.txt
@@ -0,0 +1,7 @@
+ai2thor==2.1.0
+opencv-python
+pandas
+torch
+torchvision
+tqdm
+werkzeug==2.0.3
\ No newline at end of file
diff --git a/requirements.txt b/requirements.txt
index 8b4a6fb..843e46c 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,8 +1 @@
-ai2thor==2.1.0
-opencv-python
-pandas
-textworld[pddl]>=1.6.1
-torch
-torchvision
-tqdm
-werkzeug==2.0.3
\ No newline at end of file
+textworld[pddl]>=1.6.1
\ No newline at end of file
diff --git a/scripts/README.md b/scripts/README.md
index aa072c8..c389f90 100644
--- a/scripts/README.md
+++ b/scripts/README.md
@@ -45,7 +45,7 @@ You heat the potato 1 using the microwave 1.
 You arrive at loc 8. On the sinkbasin 1, you see a knife 3, a egg 2,
 and a dishsponge 3.
 
-> put potato 1 in/on sinkbasin 1
+> move potato 1 to sinkbasin 1
 You won!
 ```
 
diff --git a/scripts/alfworld-download b/scripts/alfworld-download
index ae9ec3e..f1721ea 100755
--- a/scripts/alfworld-download
+++ b/scripts/alfworld-download
@@ -17,7 +17,7 @@ from alfworld.info import ALFRED_PDDL_PATH, ALFRED_TWL2_PATH
 
 JSON_FILES_URL = "https://github.com/alfworld/alfworld/releases/download/0.2.2/json_2.1.1_json.zip"
 PDDL_FILES_URL = "https://github.com/alfworld/alfworld/releases/download/0.2.2/json_2.1.1_pddl.zip"
-TW_PDDL_FILES_URL = "https://github.com/alfworld/alfworld/releases/download/0.2.2/json_2.1.1_tw-pddl.zip"
+TW_PDDL_FILES_URL = "https://github.com/alfworld/alfworld/releases/download/0.4.0/json_2.1.2_tw-pddl.zip"
 MRCNN_URL = "https://github.com/alfworld/alfworld/releases/download/0.2.2/mrcnn_alfred_objects_sep13_004.pth"
 CHECKPOINTS_URL = "https://github.com/alfworld/alfworld/releases/download/0.2.2/pretrained_checkpoints.zip"
 SEQ2SEQ_DATA_URL = "https://github.com/alfworld/alfworld/releases/download/0.2.2/seq2seq_data.zip"
diff --git a/scripts/alfworld-play-thor b/scripts/alfworld-play-thor
index fc6f512..22caecf 100755
--- a/scripts/alfworld-play-thor
+++ b/scripts/alfworld-play-thor
@@ -134,6 +134,10 @@ if __name__ == "__main__":
 
         # Remove problem which contains movable receptacles.
         problems = [p for p in problems if "movable_recep" not in p]
+
+        if len(problems) == 0:
+            raise ValueError(f"Can't find problem files in {ALFWORLD_DATA}. Did you run alfworld-data?")
+        
         args.problem = os.path.dirname(random.choice(problems))
 
     if "movable_recep" in args.problem:
diff --git a/scripts/alfworld-play-tw b/scripts/alfworld-play-tw
index 8e52bb0..9a014f6 100755
--- a/scripts/alfworld-play-tw
+++ b/scripts/alfworld-play-tw
@@ -34,7 +34,7 @@ def main(args):
     # dump game file
     gamedata = dict(**GAME_LOGIC, pddl_problem=open(pddl_file).read())
     gamefile = os.path.join(os.path.dirname(pddl_file), 'game.tw-pddl')
-    json.dump(gamedata, open(gamefile, "w"))
+    #json.dump(gamedata, open(gamefile, "w"))
 
 
     expert = AlfredExpert(expert_type=AlfredExpertType.PLANNER)
@@ -90,6 +90,10 @@ if __name__ == "__main__":
 
         # Remove problem which contains movable receptacles.
         problems = [p for p in problems if "movable_recep" not in p]
+
+        if len(problems) == 0:
+            raise ValueError(f"Can't find problem files in {ALFWORLD_DATA}. Did you run alfworld-data?")
+
         args.problem = os.path.dirname(random.choice(problems))
 
     if "movable_recep" in args.problem:
diff --git a/scripts/patch_tw-pddl.py b/scripts/patch_tw-pddl.py
new file mode 100644
index 0000000..ac0fe5a
--- /dev/null
+++ b/scripts/patch_tw-pddl.py
@@ -0,0 +1,71 @@
+# This script is to patch the 2.1.1 game.tw-pddl files to bring up to 2.1.2.
+# This script will add a new "help" action and separate the "PutObject" action
+#  into two separate actions: PutObjectInContainer and PutObjectOnSupporter.
+# This script will also add the corresponding grammar for the new actions.
+# This script also patch the grammar to fix a typo in the go-to feedback.
+# The script will create a backup of the original file before patching.
+
+import os
+import json
+from glob import glob
+from os.path import join as pjoin
+from tqdm import tqdm
+
+from alfworld.info import ALFWORLD_DATA
+import os
+import json
+
+import tqdm
+
+
+HELP_ACTION_PDDL = """\
+(:action help
+    :parameters (?a - agent)
+    :precondition
+        ()
+    :effect
+        (and
+            (checked ?a)
+        )
+)
+"""
+
+HELP_GRAMMAR = """\
+
+action help {
+    template :: "help";
+    feedback :: "\nAvailable commands:\n  look:                             look around your current location\n  inventory:                        check your current inventory\n  go to (receptacle):               move to a receptacle\n  open (receptacle):                open a receptacle\n  close (receptacle):               close a receptacle\n  take (object) from (receptacle):  take an object from a receptacle\n  move (object) to (receptacle):  place an object in or on a receptacle\n  examine (something):              examine a receptacle or an object\n  use (object):                     use an object\n  heat (object) with (receptacle):  heat an object using a receptacle\n  clean (object) with (receptacle): clean an object using a receptacle\n  cool (object) with (receptacle):  cool an object using a receptacle\n  slice (object) with (object):     slice an object using a sharp object\n";
+}
+"""
+
+def patch_twpddl(filename):
+    with open(filename, "r") as f:
+        data = json.load(f)
+
+    # Make backup if doesn't exist
+    if not os.path.exists(filename + ".bak"):
+        with open(filename + ".bak", "w") as f:
+            json.dump(data, f)
+
+    # Always start from backup.
+    with open(filename + ".bak") as f:
+        data = json.load(f)
+
+    # Patch domain pddl.
+    before, after = data["pddl_domain"].rsplit(")", 1)
+    data["pddl_domain"] = f"{before}{HELP_ACTION_PDDL}){after}"
+
+    # Patch grammar.
+    data["grammar"] = data["grammar"].replace("You arrive at {lend.name}.", "You arrive at {r.name}.")
+    data["grammar"] = data["grammar"].replace("put {o} in/on {r}", "move {o} to {r}")
+    data["grammar"] = data["grammar"].replace("You put the {o.name} in/on the {r.name}.", "You move the {o.name} to the {r.name}.")
+    data["grammar"] += HELP_GRAMMAR
+
+    with open(filename, "w") as f:
+        json.dump(data, f)
+
+
+filenames = glob(pjoin(ALFWORLD_DATA, "json_2.1.1/**/**/**/*.tw-pddl"))
+
+for filename in tqdm(filenames):
+   patch_twpddl(filename)
diff --git a/setup.py b/setup.py
index 23223bc..7ece51e 100644
--- a/setup.py
+++ b/setup.py
@@ -26,5 +26,6 @@
     install_requires=open('requirements.txt').readlines(),
     extras_require={
         'full': open('requirements-full.txt').readlines(),
+        'vis': open('requirements-vis.txt').readlines(),
     }
 )