Merge pull request #9 from MIT-TESSE/0.1.6-SNAPSHOT

0.1.6 snapshot

baselines/goseek-ppo.ipynb

@@ -73,7 +73,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "filename = Path(\"../../goseek-challenge/simulator/goseek-v0.1.0.x86_64\")\n",
+    "filename = Path(\"../../goseek-challenge/simulator/goseek-v0.1.4.x86_64\")\n",
     "assert filename.exists(), f\"Must set a valid path!\""
    ]
   },
@@ -429,4 +429,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 2
-}
+}

setup.py

@@ -23,7 +23,7 @@
 
 setup(
     name="tesse_gym",
-    version="0.1.5",
+    version="0.1.6",
     description="TESSE OpenAI Gym python interface",
     packages=find_packages("src"),
     # tell setuptools that all packages will be under the 'src' directory

src/tesse_gym/core/tesse_gym.py

@@ -209,7 +209,8 @@ def step(self, action: int) -> Tuple[np.ndarray, float, bool, Dict[str, Any]]:
         if reward_info["env_changed"] and not self.done:
             # environment changes will not advance game time
             # advance here so the perception server will be up to date
-            self.advance_game_time(1)
+            if not self.ground_truth_mode:
+                self.advance_game_time(1)
             response = self.get_synced_observation()
 
         self._update_pose(response.metadata)
@@ -243,10 +244,7 @@ def reset(
         self.done = False
         self.steps = 0
 
-        if not self.ground_truth_mode:
-            observation = self.get_synced_observation()
-        else:
-            observation = self.observe()
+        observation = self.get_synced_observation()
 
         self._init_pose(observation.metadata)
         return self.form_agent_observation(observation)

src/tesse_gym/tasks/goseek/goseek.py

@@ -58,14 +58,14 @@ def __init__(
         network_config: Optional[NetworkConfig] = NetworkConfig(),
         scene_id: Optional[int] = None,
         episode_length: Optional[int] = 400,
-        step_rate: Optional[int] = -1,
+        step_rate: Optional[int] = 20,
         n_targets: Optional[int] = 30,
         success_dist: Optional[float] = 2,
         restart_on_collision: Optional[bool] = False,
         init_hook: Optional[Callable[[TesseGym], None]] = set_all_camera_params,
         target_found_reward: Optional[int] = 1,
         ground_truth_mode: Optional[bool] = True,
-        n_target_types: Optional[int] = 1,
+        n_target_types: Optional[int] = 5,
     ):
         """ Initialize the TESSE treasure hunt environment.
 
@@ -86,7 +86,8 @@ def __init__(
                 assumes an external perception pipeline is running. In the latter mode, discrete
                 steps will be translated to continuous control commands and observations will be
                 explicitly synced with sim time.
-            n_target_types (int): Number of target types available to spawn.
+            n_target_types (int): Number of target types available to spawn. GOSEEK challenge 
+                has 5 target types by default. 
         """
         super().__init__(
             build_path,
@@ -149,7 +150,8 @@ def reset(
         if not self.ground_truth_mode:
             self.advance_game_time(1)
 
-        return self.form_agent_observation(self.observe())
+        observation = self.get_synced_observation()
+        return self.form_agent_observation(observation)
 
     def apply_action(self, action: int) -> None:
         """ Make agent take the specified action.