Merge branch 'ltl_configue' of github.com:bark-simulator/bark-ml into…

… ltl_configue

.github/workflows/main.yml

@@ -14,8 +14,9 @@ jobs:
         run: virtualenv -p python3 ./bark_ml/python_wrapper/venv  --system-site-packages
       - name: Entering the virtual python environment..
         run: . ./bark_ml/python_wrapper/venv/bin/activate
-      - name: Installing and upgrading pip packages..
+      - name: Installing and upgrading pip..
         run: pip install --upgrade pip
+      - name: Installing and upgrading pip packages..
         run: pip install --upgrade --trusted-host pypi.org -r ./utils/docker/installers/requirements.txt
       - name: Installing Tensorflow packages..
         run: pip install tensorflow-addons==0.11.2 graph-nets==1.1.0 ray==1.0.0
@@ -25,6 +26,8 @@ jobs:
         run: pip install torch==1.9.0+cpu torchvision -f https://download.pytorch.org/whl/torch_stable.html
       - name: Add torch environment variable
         run: export LD_LIBRARY_PATH=./bark_ml/python_wrapper/venv/lib/python3.7/site-packages/torch/lib/
+      - name: reinstall protobuf3.20.*.
+        run: pip install protobuf==3.20.*
       - name: Running tests
         run: bazel test //bark_ml:unit_tests
 

bark_ml/environments/blueprints/highway/highway.py

@@ -8,14 +8,16 @@
 
 import os
 import numpy as np
-from bark.runtime.viewer.buffered_mp_viewer import BufferedMPViewer
+# from bark.runtime.viewer.buffered_mp_viewer import BufferedMPViewer
+from bark.runtime.viewer.matplotlib_viewer import MPViewer
+from bark.runtime.viewer.video_renderer import VideoRenderer
 from bark.runtime.scenario.scenario_generation.config_with_ease import \
   LaneCorridorConfig, ConfigWithEase
 from bark.core.world.goal_definition import GoalDefinitionPolygon
 from bark.core.geometry import Polygon2d, Point2d
 
 from bark_ml.environments.blueprints.blueprint import Blueprint
-from bark_ml.evaluators.evaluator_configs import RewardShapingEvaluator
+from bark_ml.evaluators.general_evaluator import GeneralEvaluator
 from bark_ml.behaviors.cont_behavior import BehaviorContinuousML
 from bark_ml.behaviors.discrete_behavior import BehaviorDiscreteMacroActionsML
 from bark_ml.core.observers import NearestObserver
@@ -64,7 +66,7 @@ def __init__(self,
 
     ego_lane_id = 2
     lane_configs = []
-    for i in range(0, 4):
+    for i in range(1, 4):
       is_controlled = True if (ego_lane_id == i) else None
       s_min = 0
       s_max = 250
@@ -92,18 +94,19 @@ def __init__(self,
         random_seed=random_seed,
         params=params,
         lane_corridor_configs=lane_configs)
+
+    dt = 0.2  
     if viewer:
-      # viewer = MPViewer(params=params,
-      #                   use_world_bounds=True)
-      viewer = BufferedMPViewer(
-        params=params,
-        x_range=[-55, 55],
-        y_range=[-55, 55],
-        follow_agent_id=True)
-    dt = 0.2
+      viewer = MPViewer(params=params,
+                        x_range=[-150, 150],
+                        y_range=[-150, 150],
+                        follow_agent_id=True)
+      if params["Experiment"]["ExportVideos"]:
+        viewer = VideoRenderer(renderer=viewer, world_step_time=dt)
+
     params["ML"]["RewardShapingEvaluator"]["PotentialVelocityFunctor"][
           "DesiredVel", "Desired velocity for the ego agent.", 20]
-    evaluator = RewardShapingEvaluator(params)
+    evaluator = GeneralEvaluator(params)
     observer = NearestObserver(params)
     ml_behavior = ml_behavior
 

bark_ml/environments/blueprints/merging/merging.py

@@ -7,7 +7,8 @@
 # https://opensource.org/licenses/MIT
 
 import os
-from bark.runtime.viewer.buffered_mp_viewer import BufferedMPViewer
+# from bark.runtime.viewer.buffered_mp_viewer import BufferedMPViewer
+from bark.runtime.viewer.matplotlib_viewer import MPViewer
 from bark.runtime.scenario.scenario_generation.config_with_ease import \
   LaneCorridorConfig, ConfigWithEase
 from bark.core.world.opendrive import XodrDrivingDirection
@@ -96,11 +97,10 @@ def __init__(self,
         params=params,
         lane_corridor_configs=[left_lane, right_lane])
     if viewer:
-      viewer = BufferedMPViewer(
-        params=params,
-        x_range=[-25, 25],
-        y_range=[-25, 25],
-        follow_agent_id=True)
+      viewer = MPViewer(params=params,
+                        x_range=[-150, 150],
+                        y_range=[-150, 150],
+                        follow_agent_id=True)
     dt = 0.2
     params["ML"]["RewardShapingEvaluator"]["PotentialVelocityFunctor"][
       "DesiredVel", "Desired velocity for the ego agent.", 20]

bark_ml/evaluators/evaluator_configs.py

@@ -6,7 +6,7 @@
 from bark.core.world.evaluation import \
   EvaluatorGoalReached, EvaluatorCollisionEgoAgent, \
   EvaluatorStepCount, EvaluatorDrivableArea
-
+from bark.core.geometry import Point2d
 from bark_ml.evaluators.general_evaluator import *
 
 class GoalReached(GeneralEvaluator):
@@ -177,6 +177,9 @@ def __init__(self, params):
       labels_list = []
       for label_conf in rule_config["RuleConfig"]["labels"]:
         label_params_dict = label_conf["params"].ConvertToDict()
+        if label_conf["type"] == "EgoBeyondPointLabelFunction" or label_conf["type"] == "AgentBeyondPointLabelFunction":
+          merge_point = label_params_dict["point"]
+          label_params_dict["point"] = Point2d(merge_point[0],merge_point[1])
         labels_list.append(eval("{}(*(label_params_dict.values()))".format(label_conf["type"])))
       print("labels_list:",labels_list)
 

bark_ml/evaluators/general_evaluator.py

@@ -238,12 +238,14 @@ def __call__(self, observed_world, action, eval_results):
       prev_state, cur_state = self.GetPrevAndCurState(observed_world)
       prev_dist = self.DistanceToPolygon(observed_world, prev_state)
       cur_dist = self.DistanceToPolygon(observed_world, cur_state)
+      # print("!!!!!!!!!!!!!current distance to goal is ", cur_dist)
       prev_pot = self.DistancePotential(
         prev_dist, self._params["MaxDist", "", 100.],
         self._params["DistExponent", "", 0.2])
       cur_pot = self.DistancePotential(
         cur_dist, self._params["MaxDist", "", 100.],
         self._params["DistExponent", "", 0.2])
+      print("!!!!!!!!!!!!!current potential is ", 0.99*cur_pot - prev_pot)
       return False, self.WeightedReward(self._params["Gamma", "", 0.99]*cur_pot - prev_pot), {}
     return False, 0, {}