HanningHanning's picture
Simplify data layout to task and scene
3f7021a verified
Raw
History Blame Contribute Delete
8.54 kB
from ._base_task import Base_Task
from .utils import *
import sapien
import math
import numpy as np
class adjust_bottle(Base_Task):
def setup_demo(self, **kwags):
self._forced_qpose_tag = kwags.get("forced_qpose_tag")
self._forced_model_id = kwags.get("forced_model_id")
super()._init_task_env_(**kwags)
def load_actors(self):
self.qpose_tag = np.random.randint(0, 2)
qposes = [[0.707, 0.0, 0.0, -0.707], [0.707, 0.0, 0.0, 0.707]]
xlims = [[-0.12, -0.08], [0.08, 0.12]]
self.model_id = np.random.choice([13, 16])
# Optional dataset-balancing controls. Draw the normal random values
# first so the remaining scene RNG stream is unchanged; default task
# behavior therefore stays byte-for-byte compatible.
if self._forced_qpose_tag is not None:
forced_qpose_tag = int(self._forced_qpose_tag)
if forced_qpose_tag not in (0, 1):
raise ValueError("forced_qpose_tag must be 0 (left) or 1 (right)")
self.qpose_tag = forced_qpose_tag
if self._forced_model_id is not None:
forced_model_id = int(self._forced_model_id)
if forced_model_id not in (13, 16):
raise ValueError("forced_model_id must be 13 or 16")
self.model_id = forced_model_id
self.bottle = rand_create_actor(
self,
xlim=xlims[self.qpose_tag],
ylim=[-0.13, -0.08],
zlim=[0.752],
rotate_rand=True,
qpos=qposes[self.qpose_tag],
modelname="001_bottle",
convex=True,
rotate_lim=(0, 0, 0.4),
model_id=self.model_id,
)
if self.use_dynamic:
self.bottle.set_mass(0.01)
for component in self.bottle.actor.get_components():
if isinstance(component, sapien.physx.PhysxRigidDynamicComponent):
component.set_linear_damping(10.0)
component.set_angular_damping(10.0)
self.delay(4)
self.add_prohibit_area(self.bottle, padding=0.15)
self.left_target_pose = [-0.25, -0.12, 0.95, 0, 1, 0, 0]
self.right_target_pose = [0.25, -0.12, 0.95, 0, 1, 0, 0]
def play_once(self):
if self.use_dynamic:
return self._play_once_dynamic()
else:
return self._play_once_static()
def _play_once_static(self):
arm_tag = ArmTag("right" if self.qpose_tag == 1 else "left")
target_pose = (self.right_target_pose if self.qpose_tag == 1 else self.left_target_pose)
# Grasp -> Lift
self.move(self.grasp_actor(self.bottle, arm_tag=arm_tag, pre_grasp_dis=0.1))
self.move(self.move_by_displacement(arm_tag=arm_tag, z=0.1, move_axis="arm"))
# Place
self.move(
self.place_actor(
self.bottle,
target_pose=target_pose,
arm_tag=arm_tag,
functional_point_id=0,
pre_dis=0.0,
is_open=False,
))
self.info["info"] = {
"{A}": f"001_bottle/base{self.model_id}",
"{a}": str(arm_tag),
}
return self.info
def _play_once_dynamic(self):
arm_tag = ArmTag("right" if self.qpose_tag == 1 else "left")
target_pose = (self.right_target_pose if self.qpose_tag == 1 else self.left_target_pose)
current_bottle_pose = self.bottle.get_pose()
end_position = np.array([current_bottle_pose.p[0], current_bottle_pose.p[1], current_bottle_pose.p[2]])
self._intercept_position = end_position.copy()
def robot_action_sequence_sync(need_plan_mode):
grasp_result = self.grasp_actor(self.bottle, arm_tag=arm_tag, pre_grasp_dis=0.1)
if not grasp_result or grasp_result[1] is None or len(grasp_result[1]) == 0:
return
self.move(grasp_result)
table_bounds = (-0.35, 0.35, -0.25, 0.15)
success, start_pos = self.execute_dynamic_workflow(
target_actor=self.bottle,
end_position=end_position,
robot_action_sequence=robot_action_sequence_sync,
table_bounds=table_bounds,
)
if not success:
raise RuntimeError("Failed to generate dynamic trajectory")
for component in self.bottle.actor.get_components():
if isinstance(component, sapien.physx.PhysxRigidDynamicComponent):
try:
component.set_linear_velocity(np.zeros(3))
component.set_angular_velocity(np.zeros(3))
component.set_linear_damping(10.0)
component.set_angular_damping(10.0)
except Exception:
pass
self.move(self.move_by_displacement(arm_tag=arm_tag, z=0.1, move_axis="arm"))
self.move(
self.place_actor(
self.bottle,
target_pose=target_pose,
arm_tag=arm_tag,
functional_point_id=0,
pre_dis=0.0,
is_open=False,
))
self.info["info"] = {
"{A}": f"001_bottle/base{self.model_id}",
"{a}": str(arm_tag),
}
return self.info
def get_dynamic_motion_config(self) -> dict:
if not self.use_dynamic:
return None
current_bottle_pose = self.bottle.get_pose()
end_position = np.array([current_bottle_pose.p[0], current_bottle_pose.p[1], current_bottle_pose.p[2]])
return {
'target_actor': self.bottle,
'end_position': end_position,
'table_bounds': (-0.35, 0.35, -0.15, 0.25),
}
def check_success(self):
target_hight = 0.9
bottle_pose = self.bottle.get_functional_point(0)
basic_check= ((self.qpose_tag == 0 and bottle_pose[0] < -0.15) or
(self.qpose_tag == 1 and bottle_pose[0] > 0.15)) and bottle_pose[2] > target_hight
if (not self.use_dynamic) or getattr(self, 'eval_mode', False):
return basic_check
intercept_pos = getattr(self, '_intercept_position', None)
if intercept_pos is None:
return basic_check
current_center = self.bottle.get_pose().p
displacement = np.linalg.norm(current_center - intercept_pos)
displacement_threshold = 0.1
displacement_check = displacement > displacement_threshold
left_ee_pose = self.robot.get_left_ee_pose()
right_ee_pose = self.robot.get_right_ee_pose()
left_ee_pos = np.array(left_ee_pose[:3])
right_ee_pos = np.array(right_ee_pose[:3])
dist_to_left = np.linalg.norm(current_center[:2] - left_ee_pos[:2])
dist_to_right = np.linalg.norm(current_center[:2] - right_ee_pos[:2])
proximity_threshold = 0.15
proximity_check = (dist_to_left < proximity_threshold) or (dist_to_right < proximity_threshold)
return basic_check and displacement_check and proximity_check
def check_stable(self):
if not self.use_dynamic:
return super().check_stable()
from .utils import cal_quat_dis
actors_list, actors_pose_list = [], []
dynamic_name = self.bottle.get_name()
for actor in self.scene.get_all_actors():
if actor.get_name() != dynamic_name:
actors_list.append(actor)
def get_sim(p1, p2):
return np.abs(cal_quat_dis(p1.q, p2.q) * 180)
is_stable, unstable_list = True, []
def check(times):
nonlocal is_stable
for _ in range(times):
self._update_kinematic_tasks()
self.scene.step()
for idx, actor in enumerate(actors_list):
actors_pose_list[idx].append(actor.get_pose())
for idx, actor in enumerate(actors_list):
final_pose = actors_pose_list[idx][-1]
for pose in actors_pose_list[idx][-200:]:
if get_sim(final_pose, pose) > 3.0:
is_stable = False
unstable_list.append(actor.get_name())
break
for _ in range(2000):
self._update_kinematic_tasks()
self.scene.step()
for actor in actors_list:
actors_pose_list.append([actor.get_pose()])
check(500)
return is_stable, unstable_list