def main():
args = configure_args()
if is_local:
dir_prefix = ""
else:
dir_prefix="hy-iqn-lfd-full-exp.runfiles/hythe/"
logging.info(f"Executing job: {args.jobname}")
logging.info(f"Experiment server at: {os.getcwd()}")
params = ParameterServer(filename=os.path.join(dir_prefix, params_file),
log_if_default=True)
params = configure_params(params, seed=args.jobname)
experiment_id = params["Experiment"]["random_seed"]
params_filename = os.path.join(params["Experiment"]["dir"], "params_{}.json".format(experiment_id))
behavior = BehaviorDiscreteMacroActionsML(params)
evaluator = GoalReached(params)
observer = NearestAgentsObserver(params)
viewer = MPViewer(params=params,
x_range=[-35, 35],
y_range=[-35, 35],
follow_agent_id=True)
# extract params and save experiment parameters
params["ML"]["BaseAgent"]["SummaryPath"] = os.path.join(params["Experiment"]["dir"], "agent/summaries")
params["ML"]["BaseAgent"]["CheckpointPath"] = os.path.join(params["Experiment"]["dir"], "agent/checkpoints")
params.Save(filename=params_filename)
logging.info('-' * 60)
logging.info("Writing params to :{}".format(params_filename))
logging.info('-' * 60)
# database creation
dbs1 = DatabaseSerializer(test_scenarios=1, test_world_steps=2,
num_serialize_scenarios=num_demo_scenarios)
dbs1.process(os.path.join(dir_prefix, "configuration/database"),
filter_sets="**/**/interaction_merging_light_dense_1D.json")
local_release_filename = dbs1.release(version="lfd_offline")
db1 = BenchmarkDatabase(database_root=local_release_filename)
scenario_generator1, _, _ = db1.get_scenario_generator(0)
env = HyDiscreteHighway(params=params,
scenario_generation=scenario_generator1,
behavior=behavior,
evaluator=evaluator,
observer=observer,
viewer=viewer,
render=False)
scenario, _ = scenario_generator1.get_next_scenario()
world = scenario.GetWorldState()
observer.Reset(world)
assert env.action_space._n == 8, "Action Space is incorrect!"
run(args, params, env, db=db1)
params.Save(params_filename)
logging.info('-' * 60)
logging.info("Writing params to :{}".format(params_filename))
logging.info('-' * 60)
def test_parameters(self):
# initialize Params
p = ParameterServer()
# set new parameter
self.assertTrue(p["LetsTest"]["hierarchy", "bla", True])
# check whether right value is recovered
tester = p["Car"]["Length", "Car Length", 6]
tester2 = p["Car"]["Length", "Car Length", 8]
self.assertEqual(tester, 6)
self.assertEqual(
tester2,
6) # value should not change, since key already exists in dict
# check whether access without description and default value is possible
self.assertEqual(p["Car"]["Length"], 6)
# check whether setting values works
p["Age"] = 24
self.assertEqual(p["Age"], 24)
p["Localization"]["Number of Particles"] = 2000
self.assertEqual(p["Localization"]["Number of Particles"], 2000)
# C++ Test in /bark/commons/Params/params_test.h
# write in parameters in C++ and check whether they can be accessed in python afterwards
#ParamsTest(p)
#self.assertEqual(p["param_cpp"], 16.5)
# add child in python
child = p.AddChild("ch")
self.assertTrue(child["ChildTest"]["hierarchy", "bla", True])
# write parameters to json file
p.Save("written_a_param_test.json")
def write_scenario_parameter(self, config_idx, folder):
benchmark_config = super().get_benchmark_result().get_benchmark_config(config_idx)
if benchmark_config is not None:
params = benchmark_config.scenario.json_params
p = ParameterServer()
p.ConvertToParam(params)
p.Save(os.path.join(folder, "scenario_parameters.json"))
def generate_uct_hypothesis_behavior():
ml_params = ParameterServer(filename="configuration/params/iqn_params_demo_full.json", log_if_default=True)
behavior_ml_params = ml_params["ML"]["BehaviorMPMacroActions"]
mcts_params = ParameterServer(filename="configuration/params/default_uct_params.json", log_if_default=True)
mcts_params["BehaviorUctBase"]["EgoBehavior"] = behavior_ml_params
behavior = BehaviorUCTHypothesis(mcts_params, [])
mcts_params.Save(filename="./default_uct_params.json")
return behavior, mcts_params
def generate_uct_hypothesis_behavior():
if is_local:
dir_prefix = ""
else:
dir_prefix="hy-iqn-lfd-full-beliefs-exp.runfiles/hythe/"
ml_params = ParameterServer(filename=os.path.join(dir_prefix, params_file), log_if_default=True)
behavior_ml_params = ml_params["ML"]["BehaviorMPMacroActions"]
mcts_params = ParameterServer(filename=os.path.join(dir_prefix, "configuration/params/default_uct_params.json"), log_if_default=True)
mcts_params["BehaviorUctBase"]["EgoBehavior"] = behavior_ml_params
behavior = BehaviorUCTHypothesis(mcts_params, [])
mcts_params.Save(filename="./default_uct_params.json")
return behavior, mcts_params
def test_write_params_agent(self):
params = ParameterServer()
behavior = BehaviorConstantAcceleration(params)
execution = ExecutionModelInterpolate(params)
dynamic = SingleTrackModel(params)
shape = Polygon2d([1.25, 1, 0], [
Point2d(0, 0),
Point2d(0, 2),
Point2d(4, 2),
Point2d(4, 0),
Point2d(0, 0)
])
init_state = np.zeros(4)
agent = Agent(init_state, behavior, dynamic, execution, shape,
params.AddChild("agent"))
params.Save("written_agents_param_test.json")
def main():
args = configure_args()
if is_local:
dir_prefix = ""
else:
dir_prefix = "hy-x-iqn-beliefs.runfiles/hythe/"
print("Experiment server at :", os.getcwd())
params = ParameterServer(filename=os.path.join(
dir_prefix, "configuration/params/iqn_params.json"))
params = configure_params(params, seed=args.jobname)
experiment_id = params["Experiment"]["random_seed"]
params_filename = os.path.join(params["Experiment"]["dir"],
"params_{}.json".format(experiment_id))
params_behavior_filename = os.path.join(
params["Experiment"]["dir"],
"behavior_params_{}.json".format(experiment_id))
# check if exp exists and handle preemption
exp_exists = check_if_exp_exists(params)
if exp_exists:
print("Loading existing experiment from: {}".format(
args.jobname, (params["Experiment"]["dir"])))
if os.path.isfile(params_filename):
params = ParameterServer(filename=params_filename,
log_if_default=True)
if os.path.isfile(params_behavior_filename):
params_behavior = ParameterServer(
filename=params_behavior_filename, log_if_default=True)
else:
Path(params["Experiment"]["dir"]).mkdir(parents=True, exist_ok=True)
params["ML"]["BaseAgent"]["SummaryPath"] = os.path.join(
params["Experiment"]["dir"], "agent/summaries")
params["ML"]["BaseAgent"]["CheckpointPath"] = os.path.join(
params["Experiment"]["dir"], "agent/checkpoints")
params_behavior = ParameterServer(filename=os.path.join(
dir_prefix, "configuration/params/1D_desired_gap_no_prior.json"),
log_if_default=True)
params.Save(filename=params_filename)
params_behavior.Save(filename=params_behavior_filename)
# configure belief observer
splits = 2
behavior_space = configure_behavior_space(params_behavior)
hypothesis_set, hypothesis_params = behavior_space.create_hypothesis_set_fixed_split(
split=splits)
observer = BeliefObserver(params, hypothesis_set, splits=splits)
behavior = BehaviorDiscreteMacroActionsML(params)
evaluator = GoalReached(params)
viewer = MPViewer(params=params,
x_range=[-35, 35],
y_range=[-35, 35],
follow_agent_id=True)
# database creation
dbs = DatabaseSerializer(test_scenarios=2,
test_world_steps=2,
num_serialize_scenarios=num_scenarios)
dbs.process(os.path.join(dir_prefix, "configuration/database"),
filter_sets="**/**/interaction_merging_mid_dense_1D_new.json")
local_release_filename = dbs.release(version="test")
db = BenchmarkDatabase(database_root=local_release_filename)
scenario_generator, _, _ = db.get_scenario_generator(scenario_set_id=0)
env = HyDiscreteHighway(params=params,
scenario_generation=scenario_generator,
behavior=behavior,
evaluator=evaluator,
observer=observer,
viewer=viewer,
render=is_local)
print('Observation/state space size', env.observation_space)
run(params, env, exp_exists)
params.Save(filename=params_filename)
logging.info('-' * 60)
logging.info("Writing params to :{}".format(params_filename))
logging.info('-' * 60)
params_behavior.Save(filename=params_behavior_filename)
logging.info('-' * 60)
logging.info(
"Writing behavior params to :{}".format(params_behavior_filename))
logging.info('-' * 60)
return
def main():
args = configure_args()
if is_local:
dir_prefix = ""
else:
dir_prefix = "hy-iqnfd-beliefs-exp.runfiles/hythe/"
print("Executing job :", args.jobname)
print("Experiment server at :", os.getcwd())
params = ParameterServer(filename=os.path.join(
dir_prefix, "configuration/params/iqn_params_demo_full_local.json"),
log_if_default=True)
params = configure_params(params, seed=args.jobname)
experiment_id = params["Experiment"]["random_seed"]
params_filename = os.path.join(params["Experiment"]["dir"],
"params_{}.json".format(experiment_id))
behavior = BehaviorDiscreteMacroActionsML(params)
evaluator = GoalReached(params)
params_behavior_filename = os.path.join(
params["Experiment"]["dir"],
"behavior_params_{}.json".format(experiment_id))
params_behavior = ParameterServer(filename=os.path.join(
dir_prefix, "configuration/params/1D_desired_gap_no_prior.json"),
log_if_default=True)
params_behavior.Save(filename=params_behavior_filename)
# configure belief observer
splits = 2
behavior_space = BehaviorSpace(params_behavior)
hypothesis_set, hypothesis_params = behavior_space.create_hypothesis_set_fixed_split(
split=splits)
observer = BeliefObserver(params, hypothesis_set, splits=splits)
viewer = MPViewer(params=params,
x_range=[-35, 35],
y_range=[-35, 35],
follow_agent_id=True)
# extract params and save experiment parameters
params["ML"]["BaseAgent"]["SummaryPath"] = os.path.join(
params["Experiment"]["dir"], "agent/summaries")
params["ML"]["BaseAgent"]["CheckpointPath"] = os.path.join(
params["Experiment"]["dir"], "agent/checkpoints")
params.Save(filename=params_filename)
logging.info('-' * 60)
logging.info("Writing params to :{}".format(params_filename))
logging.info('-' * 60)
# database creation
dbs = DatabaseSerializer(test_scenarios=1,
test_world_steps=2,
num_serialize_scenarios=num_scenarios)
dbs.process(os.path.join(dir_prefix, "configuration/database"),
filter_sets="**/**/interaction_merging_light_dense_1D.json")
local_release_filename = dbs.release(version="test")
db = BenchmarkDatabase(database_root=local_release_filename)
scenario_generator, _, _ = db.get_scenario_generator(0)
env = HyDiscreteHighway(params=params,
scenario_generation=scenario_generator,
behavior=behavior,
evaluator=evaluator,
observer=observer,
viewer=viewer,
render=False)
scenario, _ = scenario_generator.get_next_scenario()
world = scenario.GetWorldState()
observer.Reset(world)
assert env.action_space._n == 8, "Action Space is incorrect!"
run(args, params, env, db=db)
params.Save(params_filename)
logging.info('-' * 60)
logging.info("Writing params to :{}".format(params_filename))
logging.info('-' * 60)
params_behavior.Save(filename=params_behavior_filename)
logging.info('-' * 60)
logging.info(
"Writing behavior params to :{}".format(params_behavior_filename))
logging.info('-' * 60)
def main():
args = configure_args()
if is_local:
dir_prefix = ""
else:
dir_prefix = "hy-iqn-exp.runfiles/hythe/"
print("Executing job :", args.jobname)
print("Experiment server at :", os.getcwd())
params = ParameterServer(filename=os.path.join(
dir_prefix, "configuration/params/iqn_params.json"),
log_if_default=True)
params = configure_params(params, seed=args.jobname)
experiment_id = params["Experiment"]["random_seed"]
params_filename = os.path.join(params["Experiment"]["dir"],
"params_{}.json".format(experiment_id))
# check if exp exists and handle preemption
exp_exists = check_if_exp_exists(params)
if exp_exists:
print("Loading existing experiment from: {}".format(
args.jobname, (params["Experiment"]["dir"])))
if os.path.isfile(params_filename):
params = ParameterServer(filename=params_filename,
log_if_default=True)
else:
Path(params["Experiment"]["dir"]).mkdir(parents=True, exist_ok=True)
behavior = BehaviorDiscreteMacroActionsML(params)
evaluator = GoalReached(params)
observer = NearestAgentsObserver(params)
viewer = MPViewer(params=params,
x_range=[-35, 35],
y_range=[-35, 35],
follow_agent_id=True)
# extract params and save experiment parameters
params["ML"]["BaseAgent"]["SummaryPath"] = os.path.join(
params["Experiment"]["dir"], "agent/summaries")
params["ML"]["BaseAgent"]["CheckpointPath"] = os.path.join(
params["Experiment"]["dir"], "agent/checkpoints")
params.Save(filename=params_filename)
logging.info('-' * 60)
logging.info("Writing params to :{}".format(params_filename))
logging.info('-' * 60)
# database creation
dbs = DatabaseSerializer(test_scenarios=2,
test_world_steps=2,
num_serialize_scenarios=num_scenarios)
dbs.process(os.path.join(dir_prefix, "configuration/database"),
filter_sets="**/**/interaction_merging_light_dense_1D.json")
local_release_filename = dbs.release(version="test")
db = BenchmarkDatabase(database_root=local_release_filename)
scenario_generator, _, _ = db.get_scenario_generator(0)
env = HyDiscreteHighway(params=params,
scenario_generation=scenario_generator,
behavior=behavior,
evaluator=evaluator,
observer=observer,
viewer=viewer,
render=False)
assert env.action_space._n == 8, "Action Space is incorrect!"
run(params, env, exp_exists)
params.Save(params_filename)
logging.info('-' * 60)
logging.info("Writing params to :{}".format(params_filename))
logging.info('-' * 60)
class ExperimentRunner:
"""The ExperimentRunner-Class provides an easy-to-use interface to
train, visualize, evaluate, and manage experiments.
Additionally, it creates an Experiment only from a json that is
hashes before training. Thus, trained results can be matched to executions
and evaluations.
"""
def __init__(self,
json_file=None,
params=None,
mode="visualize",
random_seed=0):
self._logger = logging.getLogger()
self._experiment_json = json_file
if params is not None:
self._params = params
else:
self._params = ParameterServer(filename=json_file)
self._experiment_folder, self._json_name = \
self.GetExperimentsFolder(json_file)
# set random seeds
self._random_seed = random_seed
np.random.seed(random_seed)
tf.random.set_seed(random_seed)
self.SetCkptsAndSummaries()
self._experiment = self.BuildExperiment(json_file, mode)
self.Visitor(mode)
def Visitor(self, mode):
if mode == "train":
self._experiment._params.Save(self._runs_folder + "params.json")
self.Train()
if mode == "visualize":
self.Visualize()
if mode == "evaluate":
self.Evaluate()
if mode == "print":
self.PrintExperiment()
if mode == "save":
self.SaveExperiment(FLAGS.save_path)
def BuildExperiment(self, json_file, mode):
return Experiment(json_file, self._params, mode)
@staticmethod
def GetExperimentsFolder(json_file):
dir_name = Path(json_file).parent
if not os.path.isdir(dir_name):
assert f"{dir_name} does not exist."
base_name = os.path.basename(json_file)
file_name = os.path.splitext(base_name)[0]
return dir_name, file_name
@staticmethod
def GenerateHash(params):
"""
Hash-function to indicate whether the same json is used
as during training.
"""
exp_params = params.ConvertToDict()
return hashlib.sha1(repr(sorted(
exp_params.items())).encode('utf-8')).hexdigest()
def CompareHashes(self):
experiment_hash = self.GenerateHash(self._params)
if os.path.isfile(self._hash_file_path):
file = open(self._hash_file_path, 'r')
old_experiment_hash = file.readline()
file.close()
if experiment_hash != old_experiment_hash:
self._logger.warning(
"\033[31m Trained experiment hash does not match \033[0m")
def SetCkptsAndSummaries(self):
self._runs_folder = \
str(self._experiment_folder) + "/" + self._json_name + "/" + str(self._random_seed) + "/"
ckpt_folder = self._runs_folder + "ckpts/"
summ_folder = self._runs_folder + "summ/"
self._logger.info(f"Run folder of the agent {self._runs_folder}.")
self._hash_file_path = self._runs_folder + "hash.txt"
self._params["ML"]["BehaviorTFAAgents"]["CheckpointPath"] = \
ckpt_folder
self._params["ML"]["TFARunner"]["SummaryPath"] = \
summ_folder
def Train(self):
if not os.path.isfile(self._hash_file_path):
os.makedirs(os.path.dirname(self._hash_file_path), exist_ok=True)
file = open(self._hash_file_path, 'w')
file.write(str(self.GenerateHash(self._experiment.params)))
file.close()
else:
self.CompareHashes()
self._experiment.runner.SetupSummaryWriter()
self._experiment.runner.Train()
def Evaluate(self):
self.CompareHashes()
num_episodes = \
self._params["Experiment"]["NumEvaluationEpisodes"]
evaluation_results = self._experiment.runner.Run(
num_episodes=num_episodes, render=False, trace_colliding_ids=True)
# metrics = self._evaluate(evaluation_results)
# ...
def Visualize(self):
self.CompareHashes()
num_episodes = \
self._params["Experiment"]["NumVisualizationEpisodes"]
self._experiment.runner.Run(num_episodes=num_episodes, render=True)
def PrintExperiment(self):
pprint.pprint(self._experiment.params.ConvertToDict())
def SaveExperiment(self, file_path):
self._params.Save(file_path)
