generalized test_policy. Now has exported brain and random. changed min/max in sample

conf/simulator/cartpole-updated-simparam.yaml

@@ -16,20 +16,20 @@ simulator:
   initial_states:
     cart_position:
       inkling_name: initial_cart_position
-      min: 0
-      max: 1
+      min: -0.05
+      max: 0.05
     cart_velocity:
       inkling_name: initial_cart_velocity
-      min: 0
-      max: 1
+      min: -0.05
+      max: 0.05
     pole_angle: 
       inkling_name: initial_pole_angle
-      min: 0
-      max: 1
+      min: -0.05
+      max: 0.05
     pole_angular_velocity:
       inkling_name: initial_angular_velocity
-      min: 0
-      max: 1
+      min: -0.05
+      max: 0.05
   policy: bonsai
   logging: enable
   workspace_setup: True

ddm_predictor.py

@@ -4,6 +4,8 @@ import random
 import time
 from typing import Any, Dict, List
 from omegaconf import ListConfig
+from functools import partial
+from policies import random_policy, brain_policy
 
 import numpy as np
 
@@ -224,11 +226,12 @@ def env_setup():
     return workspace, access_key
 
 
-def test_random_policy(
+def test_policy(
     num_episodes: int = 5,
     num_iterations: int = 5,
     sim: Simulator = None,
     config: Dict[str, float] = None,
+    policy=random_policy,
 ):
     """Test a policy using random actions over a fixed number of episodes
 
@@ -238,9 +241,6 @@ def test_random_policy(
         number of iterations to run, by default 10
     """
 
-    def random_action():
-        return {k: random.random() for k in sim.action_keys}
-
     def _config_clean(in_config: Dict):
 
         new_config = {}
@@ -259,7 +259,7 @@ def test_random_policy(
         sim.episode_start(new_config)
         sim_state = sim.get_state()
         while not terminal:
-            action = random_action()
+            action = policy(sim_state)
             sim.episode_step(action)
             sim_state = sim.get_state()
             logger.info(f"Running iteration #{iteration} for episode #{episode}")
@@ -337,7 +337,15 @@ def main(cfg: DictConfig):
     sim.episode_start()
 
     if policy == "random":
-        test_random_policy(sim=sim, config={**episode_inits, **initial_states})
+        random_policy_from_keys = partial(random_policy, action_keys=sim.action_keys)
+        test_policy(sim=sim, config={**episode_inits, **initial_states}, policy=random_policy_from_keys)
+    elif isinstance(policy, int):
+        # If docker PORT provided, set as exported brain PORT
+        port = policy
+        url = f"http://localhost:{port}"
+        print(f"Connecting to exported brain running at {url}...")
+        trained_brain_policy = partial(brain_policy, exported_brain_url=url)
+        test_policy(sim=sim, config={**episode_inits, **initial_states}, policy=trained_brain_policy)
     elif policy == "bonsai":
         if workspace_setup:
             logger.info(f"Loading workspace information form .env")

policies.py

@@ -0,0 +1,26 @@
+"""
+Fixed policies to test our sim integration with. These are intended to take
+Brain states and return Brain actions.
+"""
+
+import random
+from typing import Dict
+import requests
+
+
+def random_policy(state, action_keys):
+    """
+    Ignore the state, move randomly.
+    """
+    action = {k: random.random() for k in action_keys}
+    return action
+
+
+def brain_policy(
+    state: Dict[str, float], exported_brain_url: str = "http://localhost:5000"
+):
+
+    prediction_endpoint = f"{exported_brain_url}/v1/prediction"
+    response = requests.get(prediction_endpoint, json=state)
+
+    return response.json()