dodox/fancy_gym
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

mp wrapper fixes

Browse Source
This commit is contained in:
Fabian
2022-07-06 09:05:35 +02:00
parent eddef33d9a
commit 6704c9d63a
43 changed files with 302 additions and 608 deletions
Download Patch File Download Diff File Expand all files Collapse all files
alr_envs/alr/classic_control/base_reacher/base_reacher.py
+5 -4
View File
@@ -1,9 +1,10 @@
from abc import ABC, abstractmethod
from typing import Union
from typing import Union, Tuple, Optional
import gym
import numpy as np
from gym import spaces
from gym.core import ObsType
from gym.utils import seeding
from alr_envs.alr.classic_control.utils import intersect
@@ -14,8 +15,7 @@ class BaseReacherEnv(gym.Env, ABC):
Base class for all reaching environments.
"""
def __init__(self, n_links: int, random_start: bool = True,
allow_self_collision: bool = False):
def __init__(self, n_links: int, random_start: bool = True, allow_self_collision: bool = False):
super().__init__()
self.link_lengths = np.ones(n_links)
self.n_links = n_links
@@ -70,7 +70,8 @@ class BaseReacherEnv(gym.Env, ABC):
def current_vel(self):
return self._angle_velocity.copy()
def reset(self):
def reset(self, *, seed: Optional[int] = None, return_info: bool = False,
options: Optional[dict] = None, ) -> Union[ObsType, Tuple[ObsType, dict]]:
# Sample only orientation of first link, i.e. the arm is always straight.
if self.random_start:
first_joint = self.np_random.uniform(np.pi / 4, 3 * np.pi / 4)
alr_envs/alr/classic_control/hole_reacher/hole_reacher.py
+5 -2
View File
@@ -1,8 +1,9 @@
from typing import Union
from typing import Union, Optional, Tuple
import gym
import matplotlib.pyplot as plt
import numpy as np
from gym.core import ObsType
from matplotlib import patches
from alr_envs.alr.classic_control.base_reacher.base_reacher_direct import BaseReacherDirectEnv
@@ -51,7 +52,8 @@ class HoleReacherEnv(BaseReacherDirectEnv):
else:
raise ValueError("Unknown reward function {}".format(rew_fct))
def reset(self):
def reset(self, *, seed: Optional[int] = None, return_info: bool = False,
options: Optional[dict] = None, ) -> Union[ObsType, Tuple[ObsType, dict]]:
self._generate_hole()
self._set_patches()
self.reward_function.reset()
@@ -223,6 +225,7 @@ class HoleReacherEnv(BaseReacherDirectEnv):
if __name__ == "__main__":
import time
env = HoleReacherEnv(5)
env.reset()
alr_envs/alr/classic_control/hole_reacher/mp_wrapper.py
+1 -4
View File
@@ -7,6 +7,7 @@ from alr_envs.black_box.raw_interface_wrapper import RawInterfaceWrapper
class MPWrapper(RawInterfaceWrapper):
@property
def context_mask(self):
return np.hstack([
[self.env.random_start] * self.env.n_links, # cos
@@ -25,7 +26,3 @@ class MPWrapper(RawInterfaceWrapper):
@property
def current_vel(self) -> Union[float, int, np.ndarray, Tuple]:
return self.env.current_vel
@property
def dt(self) -> Union[float, int]:
return self.env.dt
alr_envs/alr/classic_control/simple_reacher/mp_wrapper.py
+1 -4
View File
@@ -7,6 +7,7 @@ from alr_envs.black_box.raw_interface_wrapper import RawInterfaceWrapper
class MPWrapper(RawInterfaceWrapper):
@property
def context_mask(self):
return np.hstack([
[self.env.random_start] * self.env.n_links, # cos
@@ -23,7 +24,3 @@ class MPWrapper(RawInterfaceWrapper):
@property
def current_vel(self) -> Union[float, int, np.ndarray, Tuple]:
return self.env.current_vel
@property
def dt(self) -> Union[float, int]:
return self.env.dt
alr_envs/alr/classic_control/simple_reacher/simple_reacher.py
+5 -3
View File
@@ -1,8 +1,9 @@
from typing import Iterable, Union
from typing import Iterable, Union, Optional, Tuple
import matplotlib.pyplot as plt
import numpy as np
from gym import spaces
from gym.core import ObsType
from alr_envs.alr.classic_control.base_reacher.base_reacher_torque import BaseReacherTorqueEnv
@@ -15,7 +16,7 @@ class SimpleReacherEnv(BaseReacherTorqueEnv):
"""
def __init__(self, n_links: int, target: Union[None, Iterable] = None, random_start: bool = True,
allow_self_collision: bool = False,):
allow_self_collision: bool = False, ):
super().__init__(n_links, random_start, allow_self_collision)
# provided initial parameters
@@ -41,7 +42,8 @@ class SimpleReacherEnv(BaseReacherTorqueEnv):
# def start_pos(self):
# return self._start_pos
def reset(self):
def reset(self, *, seed: Optional[int] = None, return_info: bool = False,
options: Optional[dict] = None, ) -> Union[ObsType, Tuple[ObsType, dict]]:
self._generate_goal()
return super().reset()
alr_envs/alr/classic_control/viapoint_reacher/mp_wrapper.py
+1 -4
View File
@@ -7,6 +7,7 @@ from alr_envs.black_box.raw_interface_wrapper import RawInterfaceWrapper
class MPWrapper(RawInterfaceWrapper):
@property
def context_mask(self):
return np.hstack([
[self.env.random_start] * self.env.n_links, # cos
@@ -24,7 +25,3 @@ class MPWrapper(RawInterfaceWrapper):
@property
def current_vel(self) -> Union[float, int, np.ndarray, Tuple]:
return self.env.current_vel
@property
def dt(self) -> Union[float, int]:
return self.env.dt
alr_envs/alr/classic_control/viapoint_reacher/viapoint_reacher.py
+6 -2
View File
@@ -1,8 +1,9 @@
from typing import Iterable, Union
from typing import Iterable, Union, Tuple, Optional
import gym
import matplotlib.pyplot as plt
import numpy as np
from gym.core import ObsType
from gym.utils import seeding
from alr_envs.alr.classic_control.base_reacher.base_reacher_direct import BaseReacherDirectEnv
@@ -40,7 +41,8 @@ class ViaPointReacherEnv(BaseReacherDirectEnv):
# def start_pos(self):
# return self._start_pos
def reset(self):
def reset(self, *, seed: Optional[int] = None, return_info: bool = False,
options: Optional[dict] = None, ) -> Union[ObsType, Tuple[ObsType, dict]]:
self._generate_goal()
return super().reset()
@@ -183,8 +185,10 @@ class ViaPointReacherEnv(BaseReacherDirectEnv):
plt.pause(0.01)
if __name__ == "__main__":
import time
env = ViaPointReacherEnv(5)
env.reset()