add simple params

README.md

@@ -3,6 +3,7 @@
 We have released a challenging [OpenAI Gym](https://gym.openai.com/) compatible environment. The best strategy for this environment is to have both episodic and semantic memory systems. See the paper for more information.
 
 This env is added to the PyPI server:
+
 ```sh
 pip install room-env
 ```
@@ -76,6 +77,44 @@ Consider there is only one agent. Then this is a POMDP, where *S*<sub>*t*</sub>
 
 Currently there is no RL trained for this. We only have some heuristics. Take a look at the paper for more details.
 
+## Example
+
+Start with default parameters:
+
+```python
+env = gym.make("RoomEnv-v0")
+env.reset()
+env.step("foo")
+```
+
+Start with your own parameters:
+
+```python
+env_params = {
+    "weighting_mode": "highest",
+    "probs": {
+        "commonsense": 0.7,
+        "new_location": 0.1,
+        "new_object": 0.1,
+        "switch_person": 0.5,
+    },
+    "limits": {
+        "heads": None,
+        "tails": None,
+        "names": None,
+        "allow_spaces": False,
+    },
+    "max_step": 1000,
+    "disjoint_entities": True,
+    "num_agents": 2,
+}
+env = gym.make("RoomEnv-v0", **env_params)
+env.reset()
+env.step("foo")
+```
+
+Take a look at [this repo](https://github.com/tae898/explicit-memory) for an actual interaction with this environment.
+
 ## TODOs
 
 1. Add unittests.

room_env/data/top-human-names

@@ -0,0 +1,20 @@
+James
+Robert
+John
+Michael
+William
+David
+Richard
+Joseph
+Thomas
+Charles
+Mary
+Patricia
+Jennifer
+Linda
+Elizabeth
+Barbara
+Susan
+Jessica
+Sarah
+Karen

room_env/envs/room.py

@@ -49,8 +49,7 @@ class RoomEnv(gym.Env):
 
     def __init__(
         self,
-        semantic_knowledge_path: str = "../data/semantic-knowledge-small.json",
-        names_path: str = "../data/top-human-names-small",
+        room_size: str = "small",
         weighting_mode: str = "highest",
         probs: dict = {
             "commonsense": 0.7,
@@ -72,8 +71,7 @@ def __init__(
 
         Args
         ----
-        semantic_knowledge_path: path to the ConceptNet semantic knowledge.
-        names_path: path to the list of human names.
+        room_size: small or big
         weighting_mode: Either "weighted" or "highest"
         probs: the probabilities that govern the room environment changes.
         limits: Limitation on the triples.
@@ -84,6 +82,15 @@ def __init__(
 
         """
         super().__init__()
+        self.room_size = room_size
+        if self.room_size.lower() == "small":
+            semantic_knowledge_path = "../data/semantic-knowledge-small.json"
+            names_path = "../data/top-human-names-small"
+
+        elif self.room_size.lower() in ["large", "big"]:
+            semantic_knowledge_path = "../data/semantic-knowledge.json"
+            names_path = "../data/top-human-names"
+
         logging.debug("Creating an Observation-Question-Answer generator object ...")
         self.limits = limits
 

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = room_env
-version = 0.1.2
+version = 0.1.4
 author = Taewoon Kim
 author_email = [email protected]
 description = The Room environment