before_code stringlengths 30 149k | reviewer_comment stringlengths 16 20.6k | after_code stringlengths 30 163k | diff_context stringlengths 0 22.4k | file_path stringlengths 6 155 | comment_line int64 0 26 | language stringclasses 19
values | quality_score float64 0.07 1 | comment_type stringclasses 9
values | comment_length int64 16 20.6k | before_lines int64 2 4.61k | after_lines int64 2 4.4k | is_negative bool 2
classes | pr_title stringlengths 7 140 | pr_number int64 18 61.2k | repo_name stringclasses 29
values | repo_stars int64 321 56k | repo_language stringclasses 14
values | reviewer_username stringclasses 378
values | author_username stringclasses 739
values | user stringlengths 361 200k | assistant stringlengths 16 20.6k | system stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
link_a = torch.tensor([cube1.base_link.idx], dtype=gs.tc_int, device=gs.device)
link_b = torch.tensor([cube2.base_link.idx], dtype=gs.tc_int, device=gs.device)
scene.sim.rigid_solver.add_weld_constraint(link_a, link_b)
scene.step()
welds = scene.sim.rigid_solver.get_weld_constraints(as_tensor=True... | Just keep them both in the same test. An call the test `test_get_weld_constraints_api` | scene.step()
if i > 350:
assert scene.rigid_solver.collider._collider_state.n_contacts.to_numpy()[0] == 2
assert_allclose(scene.rigid_solver.get_dofs_velocity(), 0, tol=2e-3)
# Push the drones symmetrically and wait for them to collide
drones[0].set_dofs_velocity([0.2], ... | @@ -2352,6 +2352,62 @@ def test_drone_advanced(show_viewer):
assert abs(quat_1[2] - quat_2[2]) < tol
+@pytest.mark.required
+@pytest.mark.parametrize("backend", [gs.cpu])
+def test_get_weld_constraints_basic(show_viewer, tol):
+ scene = gs.Scene(
+ sim_options=gs.options.SimOptions(gravity=(0.0, 0.0,... | tests/test_rigid_physics.py | 26 | Python | 0.357 | suggestion | 86 | 51 | 51 | false | [FEATURE] Add 'get_weld_constraints' API. | 1,370 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | LeonLiu4 | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [FEATURE] Add 'get_weld_constraints' API.
BEFORE CODE: link_a = torch.tensor([cube1.base_link.idx], dtype=gs.tc_int, device=gs.device)
link_b = torch.tensor([cube2.base_link.idx], dtype=gs.tc_int, device=gs.device)... | Just keep them both in the same test. An call the test `test_get_weld_constraints_api` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
@override
def on_mouse_motion(self, x: int, y: int, dx: int, dy: int) -> EVENT_HANDLE_STATE:
super().on_mouse_motion(x, y, dx, dy)
self.prev_mouse_pos = (x, y)
@override
def on_mouse_drag(self, x: int, y: int, dx: int, dy: int, buttons: int, modifiers: int) -> EVENT_HANDLE_STATE:
... | You should rather use `tensor_to_array`. | def on_mouse_motion(self, x: int, y: int, dx: int, dy: int) -> EVENT_HANDLE_STATE:
super().on_mouse_motion(x, y, dx, dy)
self.prev_mouse_pos = (x, y)
@override
def on_mouse_drag(self, x: int, y: int, dx: int, dy: int, buttons: int, modifiers: int) -> EVENT_HANDLE_STATE:
super().on_m... | @@ -60,17 +56,55 @@ def on_mouse_motion(self, x: int, y: int, dx: int, dy: int) -> EVENT_HANDLE_STAT
def on_mouse_drag(self, x: int, y: int, dx: int, dy: int, buttons: int, modifiers: int) -> EVENT_HANDLE_STATE:
super().on_mouse_drag(x, y, dx, dy, buttons, modifiers)
self.prev_mouse_pos = (x, y)
... | genesis/ext/pyrender/interaction/viewer_interaction.py | 26 | Python | 0.286 | suggestion | 40 | 51 | 51 | false | [MISC] Add viewer code to handle moving objects | 1,378 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/viewer_interaction.py
PR Title: [MISC] Add viewer code to handle moving objects
BEFORE CODE:
@override
def on_mouse_motion(self, x: int, y: int, dx: int, dy: int) -> EVENT_HANDLE_STATE:
super().on_mouse_motion(... | You should rather use `tensor_to_array`. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
from typing import TYPE_CHECKING
from typing_extensions import override
import numpy as np
from numpy.typing import NDArray
import torch
import genesis as gs
from genesis.engine.entities.rigid_entity.rigid_entity import RigidEntity
from .aabb import AABB
from .ray import Plane, Ray, RayHit
from .vec3 import Pose, Qu... | Could you delay torch import inside the function using it? Because it will initialise CUDA runtime, which depends on environment variables, so the later it is imported the more leeway it gives to the end-user for modifying these environment variables at runtime. | from typing import TYPE_CHECKING
from typing_extensions import override
import numpy as np
import genesis as gs
from genesis.engine.entities.rigid_entity.rigid_entity import RigidEntity
from .aabb import AABB
from .ray import Plane, Ray, RayHit
from .vec3 import Pose, Quat, Vec3, Color
from .viewer_interaction_base ... | @@ -3,16 +3,15 @@
import numpy as np
from numpy.typing import NDArray
-
-from pyglet.event import EVENT_HANDLE_STATE
+import torch | genesis/ext/pyrender/interaction/viewer_interaction.py | 6 | Python | 0.5 | question | 262 | 31 | 31 | false | [MISC] Add viewer code to handle moving objects | 1,378 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/viewer_interaction.py
PR Title: [MISC] Add viewer code to handle moving objects
BEFORE CODE: from typing import TYPE_CHECKING
from typing_extensions import override
import numpy as np
from numpy.typing import NDArray
import to... | Could you delay torch import inside the function using it? Because it will initialise CUDA runtime, which depends on environment variables, so the later it is imported the more leeway it gives to the end-user for modifying these environment variables at runtime. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
from genesis.engine.scene import Scene
from genesis.ext.pyrender.node import Node
class ViewerInteraction(ViewerInteractionBase):
"""Functionalities to be implemented:
- mouse picking
- mouse dragging
"""
def __init__(self,
camera: 'Node',
scene: 'Scene',
viewpo... | I would recommend `(0.5 * viewport_size[0], 0.5 * viewport_size[1])` |
class ViewerInteraction(ViewerInteractionBase):
"""Functionalities to be implemented:
- mouse picking
- mouse dragging
"""
def __init__(self,
camera: 'Node',
scene: 'Scene',
viewport_size: tuple[int, int],
camera_yfov: float,
log_events: bool = False,
... | @@ -43,13 +34,18 @@ def __init__(self,
camera_fov: float = 60.0,
):
super().__init__(log_events)
- self.camera = camera
- self.scene = scene
- self.viewport_size = viewport_size
- self.camera_yfov = camera_yfov
+ self.camera: 'Node' = camera
+ self.scene:... | genesis/ext/pyrender/interaction/viewer_interaction.py | 26 | Python | 0.357 | suggestion | 68 | 51 | 51 | false | [MISC] Add viewer code to handle moving objects | 1,378 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/viewer_interaction.py
PR Title: [MISC] Add viewer code to handle moving objects
BEFORE CODE: from genesis.engine.scene import Scene
from genesis.ext.pyrender.node import Node
class ViewerInteraction(ViewerInteractionB... | I would recommend `(0.5 * viewport_size[0], 0.5 * viewport_size[1])` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
closest_hit = None
hit_entity: RigidEntity | None = None
ray_hit = self._raycast_against_ground_plane(mouse_ray)
if ray_hit.is_hit:
closest_hit = ray_hit
for entity in self.get_entities():
ray_hit = self.raycast_against_entity... | Why not just store pre-defined colours in a frozen dict of tuples whose keys would be some `IntEnum` (e.g. `COLOR.RED`) ? That would be more pythonic. |
ray_hit = self._raycast_against_ground_plane(mouse_ray)
if ray_hit.is_hit:
closest_hit = ray_hit
for entity in self.get_entities():
ray_hit = self.raycast_against_entity_obb(entity, mouse_ray)
if ray_hit.is_hit:
if... | @@ -91,6 +125,20 @@ def on_draw(self) -> None:
if hit_entity:
self._draw_entity_unrotated_oobb(hit_entity)
+ if self.picked_entity:
+ assert self.mouse_drag_plane is not None
+ assert self.picked_point_in_local is not None
+
+ # dra... | genesis/ext/pyrender/interaction/viewer_interaction.py | 26 | Python | 0.5 | question | 150 | 51 | 51 | false | [MISC] Add viewer code to handle moving objects | 1,378 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/viewer_interaction.py
PR Title: [MISC] Add viewer code to handle moving objects
BEFORE CODE: closest_hit = None
hit_entity: RigidEntity | None = None
ray_hit = self._raycast_against_ground_p... | Why not just store pre-defined colours in a frozen dict of tuples whose keys would be some `IntEnum` (e.g. `COLOR.RED`) ? That would be more pythonic. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
x = x - 0.5 * self.viewport_size[0]
y = y - 0.5 * self.viewport_size[1]
x = 2.0 * x / self.viewport_size[1] * self.tan_half_fov
y = 2.0 * y / self.viewport_size[1] * self.tan_half_fov
else:
# alternative way
projection_matrix = self.camera.... | `Vec3.from_float64` is confusing, this is an array. Why not `from_array` or `from_tensor`? Why `float64` matters here ? You would do the same in both cases? | x = 2.0 * x / self.viewport_size[1] * self.tan_half_fov
y = 2.0 * y / self.viewport_size[1] * self.tan_half_fov
else:
# alternative way
projection_matrix = self.camera.camera.get_projection_matrix(*self.viewport_size)
x = x - 0.5 * self.viewport_size[0... | @@ -117,6 +165,10 @@ def screen_position_to_ray(self, x: float, y: float) -> Ray:
direction = forward + right * x + up * y
return Ray(position, direction)
+ def get_camera_forward(self) -> Vec3:
+ mtx = self.camera.matrix
+ return Vec3.from_float64(-mtx[:3, 2]) | genesis/ext/pyrender/interaction/viewer_interaction.py | 26 | Python | 0.5 | question | 156 | 51 | 51 | false | [MISC] Add viewer code to handle moving objects | 1,378 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/viewer_interaction.py
PR Title: [MISC] Add viewer code to handle moving objects
BEFORE CODE: x = x - 0.5 * self.viewport_size[0]
y = y - 0.5 * self.viewport_size[1]
x = 2.0 * x / self.viewpor... | `Vec3.from_float64` is confusing, this is an array. Why not `from_array` or `from_tensor`? Why `float64` matters here ? You would do the same in both cases? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
from dataclasses import dataclass
from typing import Union
import numpy as np
from numpy.typing import NDArray
from genesis.utils.misc import tensor_to_array
# If not needing runtime checks, we can just use annotated types:
# Vec3 = Annotated[npt.NDArray[np.float32], (3,)]
# Aabb = Annotated[npt.NDArray[np.float32],... | You don't need to do this I think, you may just call `import torch` or `import genesis as gs`, it will be crazy fast if already imported before.
```python
In [19]: %timeit import torch
48.2 ns ± 0.452 ns per loop (mean ± std. dev. of 7 runs, 10,000,000 loops each)
```
I doubt you are chasing 50ns for now :p
| from dataclasses import dataclass
from typing import Union
import numpy as np
from numpy.typing import NDArray
from genesis.utils.misc import tensor_to_array
# If not needing runtime checks, we can just use annotated types:
# Vec3 = Annotated[npt.NDArray[np.float32], (3,)]
# Aabb = Annotated[npt.NDArray[np.float32],... | @@ -4,10 +4,23 @@
import numpy as np
from numpy.typing import NDArray
+from genesis.utils.misc import tensor_to_array
+
# If not needing runtime checks, we can just use annotated types:
# Vec3 = Annotated[npt.NDArray[np.float32], (3,)]
# Aabb = Annotated[npt.NDArray[np.float32], (2, 3)]
+_is_torch_imported: bo... | genesis/ext/pyrender/interaction/vec3.py | 14 | Python | 0.786 | suggestion | 320 | 39 | 39 | false | [MISC] Add viewer code to handle moving objects | 1,378 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/vec3.py
PR Title: [MISC] Add viewer code to handle moving objects
BEFORE CODE: from dataclasses import dataclass
from typing import Union
import numpy as np
from numpy.typing import NDArray
from genesis.utils.misc import tens... | You don't need to do this I think, you may just call `import torch` or `import genesis as gs`, it will be crazy fast if already imported before.
```python
In [19]: %timeit import torch
48.2 ns ± 0.452 ns per loop (mean ± std. dev. of 7 runs, 10,000,000 loops each)
```
I doubt you are chasing 50ns for now :p
| You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
return NotImplemented
def copy(self) -> 'Quat':
return Quat(self.v.copy())
def __repr__(self) -> str:
return f"Quat({self.v[0]}, {self.v[1]}, {self.v[2]}, {self.v[3]})"
@property
def w(self) -> float:
return self.v[0]
@property
def x(self) -> float:
... | I don't like the idea of making this a proper instead of a classic method, because the result is a copy and not a reference, which is confusing for a property I think. | def __repr__(self) -> str:
return f"Quat({self.v[0]}, {self.v[1]}, {self.v[2]}, {self.v[3]})"
def as_tensor(self) -> 'torch.Tensor':
_ensure_torch_imported()
return torch.tensor(self.v, dtype=gs.tc_float)
@property
def w(self) -> float:
return self.v[0]
@property
... | @@ -158,16 +166,27 @@ def y(self) -> float:
def z(self) -> float:
return self.v[3]
+ @property
+ def as_tensor(self) -> 'torch.Tensor': | genesis/ext/pyrender/interaction/vec3.py | 26 | Python | 0.429 | suggestion | 167 | 51 | 51 | false | [MISC] Add viewer code to handle moving objects | 1,378 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/vec3.py
PR Title: [MISC] Add viewer code to handle moving objects
BEFORE CODE: return NotImplemented
def copy(self) -> 'Quat':
return Quat(self.v.copy())
def __repr__(self) -> str:
return f... | I don't like the idea of making this a proper instead of a classic method, because the result is a copy and not a reference, which is confusing for a property I think. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
Morph defined by a plane shape.
Note
----
Plane is a primitive with infinite size. Note that the `pos` is the center of the plane, but essetially only defines a point where the plane passes through.
Parameters
----------
pos : tuple, shape (3,), optional
The center position of the ... | Why 2 flags ? Only define one or the other. Moreover, I think it should be one-sided by default. | Morph defined by a plane shape.
Note
----
Plane is a primitive with infinite size. Note that the `pos` is the center of the plane, but essetially only defines a point where the plane passes through.
Parameters
----------
pos : tuple, shape (3,), optional
The center position of the ... | @@ -331,10 +331,14 @@ class Plane(Primitive):
Whether the entity needs to be visualized. Set it to False if you need a invisible object only for collision purposes. Defaults to True. `visualization` and `collision` cannot both be False. **This is only used for RigidEntity.**
collision : bool, optional
... | genesis/options/morphs.py | 26 | Python | 0.286 | question | 96 | 51 | 51 | false | [FEATURE] Enabling rendering plane geometry one-sided. | 1,396 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | LeonLiu4 | Review this pull request change.
Language: Python
File: genesis/options/morphs.py
PR Title: [FEATURE] Enabling rendering plane geometry one-sided.
BEFORE CODE: Morph defined by a plane shape.
Note
----
Plane is a primitive with infinite size. Note that the `pos` is the center of the plane, but esse... | Why 2 flags ? Only define one or the other. Moreover, I think it should be one-sided by default. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
The normal normal of the plane in its local frame. Defaults to (0, 0, 1).
visualization : bool, optional
Whether the entity needs to be visualized. Set it to False if you need a invisible object only for collision purposes. Defaults to True. `visualization` and `collision` cannot both be False. **Th... | This is useless. It will be done automatically. | The normal normal of the plane in its local frame. Defaults to (0, 0, 1).
visualization : bool, optional
Whether the entity needs to be visualized. Set it to False if you need a invisible object only for collision purposes. Defaults to True. `visualization` and `collision` cannot both be False. **Th... | @@ -348,6 +351,8 @@ def __init__(self, **data):
if self.requires_jac_and_IK:
gs.raise_exception("`requires_jac_and_IK` must be False for `Plane`.")
+ self.double_sided = self.double_sided | genesis/options/morphs.py | 26 | Python | 0.143 | suggestion | 47 | 51 | 51 | false | [FEATURE] Enabling rendering plane geometry one-sided. | 1,396 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | LeonLiu4 | Review this pull request change.
Language: Python
File: genesis/options/morphs.py
PR Title: [FEATURE] Enabling rendering plane geometry one-sided.
BEFORE CODE: The normal normal of the plane in its local frame. Defaults to (0, 0, 1).
visualization : bool, optional
Whether the entity needs to be ... | This is useless. It will be done automatically. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
radius: float = 0.5
class Plane(Primitive):
"""
Morph defined by a plane shape.
Note
----
Plane is a primitive with infinite size. Note that the `pos` is the center of the plane, but essetially only defines a point where the plane passes through.
Parameters
----------
pos : tuple... | Terrain morph should also have this option. | radius: float = 0.5
class Plane(Primitive):
"""
Morph defined by a plane shape.
Note
----
Plane is a primitive with infinite size. Note that the `pos` is the center of the plane, but essetially only defines a point where the plane passes through.
Parameters
----------
pos : tuple... | @@ -331,10 +331,13 @@ class Plane(Primitive):
Whether the entity needs to be visualized. Set it to False if you need a invisible object only for collision purposes. Defaults to True. `visualization` and `collision` cannot both be False. **This is only used for RigidEntity.**
collision : bool, optional
... | genesis/options/morphs.py | 26 | Python | 0.143 | suggestion | 43 | 51 | 51 | false | [FEATURE] Enabling rendering plane geometry one-sided. | 1,396 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | LeonLiu4 | Review this pull request change.
Language: Python
File: genesis/options/morphs.py
PR Title: [FEATURE] Enabling rendering plane geometry one-sided.
BEFORE CODE: radius: float = 0.5
class Plane(Primitive):
"""
Morph defined by a plane shape.
Note
----
Plane is a primitive with infinite size... | Terrain morph should also have this option. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
dtype=np.float32,
)
def on_rigid(self):
if self.sim.rigid_solver.is_active():
# TODO: support dynamic switching in GUI later
for rigid_entity in self.sim.rigid_solver.entities:
if rigid_entity.surface.vis_mode == "visual":
geom... | I don't think this change is a good idea. |
pose = gu.trans_quat_to_T(tool_entity.init_pos, tool_entity.init_quat)
double_sided = tool_entity.surface.double_sided
self.add_static_node(
tool_entity, pyrender.Mesh.from_trimesh(mesh, double_sided=double_sided), pose=pose
... | @@ -296,9 +296,7 @@ def on_rigid(self):
mesh=mesh,
poses=geom_T,
smooth=geom.surface.smooth if "collision" not in rigid_entity.surface.vis_mode else False,
- double_sided=(
- ... | genesis/vis/rasterizer_context.py | 26 | Python | 0.143 | suggestion | 41 | 51 | 51 | false | [FEATURE] Enabling rendering plane geometry one-sided. | 1,396 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | LeonLiu4 | Review this pull request change.
Language: Python
File: genesis/vis/rasterizer_context.py
PR Title: [FEATURE] Enabling rendering plane geometry one-sided.
BEFORE CODE: dtype=np.float32,
)
def on_rigid(self):
if self.sim.rigid_solver.is_active():
# TODO: support dynamic s... | I don't think this change is a good idea. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
visual = trimesh.visual.ColorVisuals()
visual._data["vertex_colors"] = np.tile((np.asarray(color) * 255).astype(np.uint8), (len(vertices), 1))
mesh = trimesh.Trimesh(vertices=vertices, faces=faces, visual=visual, process=False)
mesh._cache.id_set()
mesh._cache.cache["face_normals"] = face_normals
... | Move back `uv` where it was initially defined. |
visual = trimesh.visual.ColorVisuals()
visual._data["vertex_colors"] = np.tile((np.asarray(color) * 255).astype(np.uint8), (len(vertices), 1))
mesh = trimesh.Trimesh(vertices=vertices, faces=faces, visual=visual, process=False)
mesh._cache.id_set()
mesh._cache.cache["face_normals"] = face_normals
... | @@ -860,24 +860,36 @@ def create_box(extents=None, color=(1.0, 1.0, 1.0, 1.0), bounds=None, wireframe=
def create_plane(size=1e3, color=None, normal=(0.0, 0.0, 1.0)):
thickness = 1e-2 # for safety
- mesh = trimesh.creation.box(extents=[size, size, thickness])
+ half = size * 0.5
+ verts = np.array(
+ ... | genesis/utils/mesh.py | 26 | Python | 0.286 | suggestion | 46 | 51 | 51 | false | [FEATURE] Enabling rendering plane geometry one-sided. | 1,396 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | LeonLiu4 | Review this pull request change.
Language: Python
File: genesis/utils/mesh.py
PR Title: [FEATURE] Enabling rendering plane geometry one-sided.
BEFORE CODE:
visual = trimesh.visual.ColorVisuals()
visual._data["vertex_colors"] = np.tile((np.asarray(color) * 255).astype(np.uint8), (len(vertices), 1))
mesh... | Move back `uv` where it was initially defined. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
j = i + l * d
delta_node = self.delta(i, j, i_b)
s = ti.u32(0)
t = (l + 1) // 2
while t > 0:
if self.delta(i, i + (s + t) * d, i_b) > delta_node:
s += t
t = ti.select(t > 1, (t + 1) // 2, 0)
gamm... | I wonder if the performance will be slower than the original u64. | j = i + l * d
delta_node = self.delta(i, j, i_b)
s = ti.u32(0)
t = (l + 1) // 2
while t > 0:
if self.delta(i, i + (s + t) * d, i_b) > delta_node:
s += t
t = ti.select(t > 1, (t + 1) // 2, 0)
gamm... | @@ -318,10 +318,13 @@ def delta(self, i, j, i_b):
result = -1
if j >= 0 and j < self.n_aabbs:
result = 64
- x = self.morton_codes[i_b, ti.i32(i)] ^ self.morton_codes[i_b, ti.i32(j)]
- for b in range(64):
- if x & (ti.u64(1) << (63 - b)):
- ... | genesis/engine/bvh.py | 26 | Python | 0.357 | performance | 65 | 51 | 51 | false | [BUG FIX] Fix support of Apple Metal. | 1,404 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | duburcqa | Review this pull request change.
Language: Python
File: genesis/engine/bvh.py
PR Title: [BUG FIX] Fix support of Apple Metal.
BEFORE CODE: j = i + l * d
delta_node = self.delta(i, j, i_b)
s = ti.u32(0)
t = (l + 1) // 2
while t > 0:
if self.... | I wonder if the performance will be slower than the original u64. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self.held_geom = None
def apply_force(self, control_point: Vec3, delta_time: float):
_ensure_torch_imported()
# works ok:
# delta: Vec3 = control_point - self.prev_control_point
# pos = Vec3.from_tensor(self.held_geom.entity.get_pos())
# pos = pos + delta
... | Do not hardcode parameters in the middle of functions. Define global constant (FULL_UPPERCASE) right after imports. | self.prev_control_point = control_point
def detach(self):
self.held_geom = None
def apply_force(self, control_point: Vec3, delta_time: float):
_ensure_torch_imported()
# works ok:
# delta: Vec3 = control_point - self.prev_control_point
# pos = Vec3.from... | @@ -0,0 +1,82 @@
+
+from genesis.engine.entities.rigid_entity.rigid_entity import RigidEntity
+from genesis.engine.entities.rigid_entity.rigid_geom import RigidGeom
+
+from .ray import Plane, Ray, RayHit
+from .vec3 import Pose, Quat, Vec3, Color
+
+from genesis.engine.entities.rigid_entity.rigid_link import RigidLink
... | genesis/ext/pyrender/interaction/mouse_spring.py | 26 | Python | 0.357 | suggestion | 115 | 51 | 51 | false | [MISC] Stub code for physical object dragging | 1,411 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/mouse_spring.py
PR Title: [MISC] Stub code for physical object dragging
BEFORE CODE: self.held_geom = None
def apply_force(self, control_point: Vec3, delta_time: float):
_ensure_torch_imported()
... | Do not hardcode parameters in the middle of functions. Define global constant (FULL_UPPERCASE) right after imports. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
pose: Pose = Pose.from_geom(self.held_geom)
self.held_point_in_local = pose.inverse_transform_point(control_point)
self.prev_control_point = control_point
def detach(self):
self.held_geom = None
def apply_force(self, control_point: Vec3, delta_time: float):
_ensure_torc... | I don't know why, but `0.0 < link.get_mass()` is confusing to me. I would use `link.get_mass() > 0.0` | def attach(self, picked_entity: RigidEntity, control_point: Vec3):
# for now, we just pick the first geometry
self.held_geom = picked_entity.geoms[0]
pose: Pose = Pose.from_geom(self.held_geom)
self.held_point_in_local = pose.inverse_transform_point(control_point)
self.prev_c... | @@ -0,0 +1,82 @@
+
+from genesis.engine.entities.rigid_entity.rigid_entity import RigidEntity
+from genesis.engine.entities.rigid_entity.rigid_geom import RigidGeom
+
+from .ray import Plane, Ray, RayHit
+from .vec3 import Pose, Quat, Vec3, Color
+
+from genesis.engine.entities.rigid_entity.rigid_link import RigidLink
... | genesis/ext/pyrender/interaction/mouse_spring.py | 26 | Python | 0.571 | suggestion | 101 | 51 | 51 | false | [MISC] Stub code for physical object dragging | 1,411 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/mouse_spring.py
PR Title: [MISC] Stub code for physical object dragging
BEFORE CODE: pose: Pose = Pose.from_geom(self.held_geom)
self.held_point_in_local = pose.inverse_transform_point(control_point)
sel... | I don't know why, but `0.0 < link.get_mass()` is confusing to me. I would use `link.get_mass() > 0.0` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
dir: Vec3 = Vec3.zero()
dir.v[i] = 1.0
pos_err: float = dir.dot(pos_err_v)
vel_err: float = dir.dot(vel_err_v)
err: float = tau * pos_err * inv_dt + damp * vel_err
vm: float = inv_mass
imp: float = err * vm
lin_vel += imp *... | Skip line at the end of the line. |
from genesis.engine.entities.rigid_entity.rigid_entity import RigidEntity
from genesis.engine.entities.rigid_entity.rigid_geom import RigidGeom
from .ray import Plane, Ray, RayHit
from .vec3 import Pose, Quat, Vec3, Color
from genesis.engine.entities.rigid_entity.rigid_link import RigidLink
MOUSE_SPRING_POSITION_CO... | @@ -0,0 +1,82 @@
+
+from genesis.engine.entities.rigid_entity.rigid_entity import RigidEntity
+from genesis.engine.entities.rigid_entity.rigid_geom import RigidGeom
+
+from .ray import Plane, Ray, RayHit
+from .vec3 import Pose, Quat, Vec3, Color
+
+from genesis.engine.entities.rigid_entity.rigid_link import RigidLink
... | genesis/ext/pyrender/interaction/mouse_spring.py | 26 | Python | 0.071 | suggestion | 33 | 26 | 81 | false | [MISC] Stub code for physical object dragging | 1,411 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/mouse_spring.py
PR Title: [MISC] Stub code for physical object dragging
BEFORE CODE: dir: Vec3 = Vec3.zero()
dir.v[i] = 1.0
pos_err: float = dir.dot(pos_err_v)
vel_err: float = di... | Skip line at the end of the line. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
@pytest.mark.required
@pytest.mark.parametrize("model_name", ["box_plan"])
@pytest.mark.parametrize("gs_solver", [gs.constraint_solver.CG, gs.constraint_solver.Newton])
@pytest.mark.parametrize("gs_integrator", [gs.integrator.implicitfast, gs.integrator.Euler])
@pytest.mark.parametrize("backend", [gs.cpu])
def test_bo... | What was the original tolerance?
(I tried to print it out, but don't know how to get prints to show when run with pytest) |
@pytest.mark.required
@pytest.mark.parametrize("model_name", ["box_plan"])
@pytest.mark.parametrize("gs_solver", [gs.constraint_solver.CG, gs.constraint_solver.Newton])
@pytest.mark.parametrize("gs_integrator", [gs.integrator.implicitfast, gs.integrator.Euler])
@pytest.mark.parametrize("backend", [gs.cpu])
def test_bo... | @@ -329,6 +329,9 @@ def test_box_plane_dynamics(gs_sim, mj_sim, tol):
@pytest.mark.parametrize("gjk_collision", [True, False])
@pytest.mark.parametrize("backend", [gs.cpu])
def test_simple_kinematic_chain(gs_sim, mj_sim, tol):
+ # after PR #1423, I need to relax the tolerance. It should not affect the test result... | tests/test_rigid_physics.py | 26 | Python | 0.357 | question | 125 | 51 | 51 | false | [BUG FIX] Move forward_kinematics to the end of a step. | 1,423 | Genesis-Embodied-AI/Genesis | 26,702 | Python | gasnica | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [BUG FIX] Move forward_kinematics to the end of a step.
BEFORE CODE:
@pytest.mark.required
@pytest.mark.parametrize("model_name", ["box_plan"])
@pytest.mark.parametrize("gs_solver", [gs.constraint_solver.CG, gs.constraint... | What was the original tolerance?
(I tried to print it out, but don't know how to get prints to show when run with pytest) | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self._rasterizer = None
if self._raytracer is not None:
self._raytracer.destroy()
self._raytracer = None
if self._context is not None:
self._context.destroy()
del self._context
self._context = None
self._renderer = None
... | Is this correct?
viewer_lock is either a real viewer lock or a DummyViewerLock. It will never be None unless it is not built. | self._rasterizer = None
if self._raytracer is not None:
self._raytracer.destroy()
self._raytracer = None
if self._context is not None:
self._context.destroy()
del self._context
self._context = None
self._renderer = None
... | @@ -130,15 +130,16 @@ def reset(self):
self._context.reset()
- # Need to update viewer once here, because otherwise camera will update scene if render is called right after
- # build, which will lead to segfault.
- if self._viewer is not None:
- if self._viewer.is_alive():
-... | genesis/vis/visualizer.py | 26 | Python | 0.357 | question | 127 | 51 | 51 | false | [BUG FIX] Fix broken interactive viewer. | 1,426 | Genesis-Embodied-AI/Genesis | 26,702 | Python | yuhongyi | duburcqa | Review this pull request change.
Language: Python
File: genesis/vis/visualizer.py
PR Title: [BUG FIX] Fix broken interactive viewer.
BEFORE CODE: self._rasterizer = None
if self._raytracer is not None:
self._raytracer.destroy()
self._raytracer = None
if self._cont... | Is this correct?
viewer_lock is either a real viewer lock or a DummyViewerLock. It will never be None unless it is not built. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
import argparse
import genesis as gs
def main():
parser = argparse.ArgumentParser()
parser.add_argument("-v", "--vis", action="store_true", default=False)
args = parser.parse_args()
gs.init()
scene = gs.Scene(
viewer_options=gs.options.ViewerOptions(
camera_pos=(0.0, -2, 1.5... | Move this comment on the very top of this file. | """
NOTE: contype and conaffinity are 32-bit integer bitmasks used for contact filtering of contact pairs.
When the contype of one geom and the conaffinity of the other geom share a common bit set to 1, two geoms can collide.
Plane: contype=0xFFFF, conaffinity=0xFFFF (1111 1111 1111 1111)
Red Cube: contype=1, co... | @@ -0,0 +1,90 @@
+import argparse
+
+import genesis as gs
+
+
+def main():
+ parser = argparse.ArgumentParser()
+ parser.add_argument("-v", "--vis", action="store_true", default=False)
+ args = parser.parse_args()
+
+ gs.init()
+
+ scene = gs.Scene(
+ viewer_options=gs.options.ViewerOptions(
+ ... | examples/collision/contype.py | 25 | Python | 0.143 | refactor | 47 | 50 | 50 | false | [FEATURE] Add support of 'contype'/'conaffinity' for primitive and mesh entities. | 1,438 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Kashu7100 | Review this pull request change.
Language: Python
File: examples/collision/contype.py
PR Title: [FEATURE] Add support of 'contype'/'conaffinity' for primitive and mesh entities.
BEFORE CODE: import argparse
import genesis as gs
def main():
parser = argparse.ArgumentParser()
parser.add_argument("-v", "--v... | Move this comment on the very top of this file. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
show_viewer=args.vis,
)
scene.add_entity(gs.morphs.Plane())
# NOTE: contype and conaffinity are 32-bit integer bitmasks used for contact filtering of contact pairs.
# When the contype of one geom and the conaffinity of the other geom share a common bit set to 1, two geoms can collide.
# Pl... | Do not skip line between each `add_entity` |
gs.init()
scene = gs.Scene(
viewer_options=gs.options.ViewerOptions(
camera_pos=(0.0, -2, 1.5),
camera_lookat=(0.0, 0.0, 0.5),
camera_fov=40,
max_FPS=200,
),
show_viewer=args.vis,
)
scene.add_entity(gs.morphs.Plane())
scene.... | @@ -0,0 +1,90 @@
+import argparse
+
+import genesis as gs
+
+
+def main():
+ parser = argparse.ArgumentParser()
+ parser.add_argument("-v", "--vis", action="store_true", default=False)
+ args = parser.parse_args()
+
+ gs.init()
+
+ scene = gs.Scene(
+ viewer_options=gs.options.ViewerOptions(
+ ... | examples/collision/contype.py | 26 | Python | 0.214 | suggestion | 42 | 51 | 51 | false | [FEATURE] Add support of 'contype'/'conaffinity' for primitive and mesh entities. | 1,438 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Kashu7100 | Review this pull request change.
Language: Python
File: examples/collision/contype.py
PR Title: [FEATURE] Add support of 'contype'/'conaffinity' for primitive and mesh entities.
BEFORE CODE: show_viewer=args.vis,
)
scene.add_entity(gs.morphs.Plane())
# NOTE: contype and conaffinity are 32-bit ... | Do not skip line between each `add_entity` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
dt: Optional[float] = None
gravity: Optional[tuple] = None
enable_collision: bool = True
enable_joint_limit: bool = True
enable_self_collision: bool = True
enable_adjacent_collision: bool = False
disable_constraint: bool = False
max_collision_pairs: int = 300
integrator: gs.integrat... | can we change it to 0.01 to be safer? | """
dt: Optional[float] = None
gravity: Optional[tuple] = None
enable_collision: bool = True
enable_joint_limit: bool = True
enable_self_collision: bool = True
enable_adjacent_collision: bool = False
disable_constraint: bool = False
max_collision_pairs: int = 300
integrator: gs.... | @@ -292,7 +291,7 @@ class RigidOptions(Options):
ls_tolerance: float = 1e-2
sparse_solve: bool = False
contact_resolve_time: Optional[float] = None
- constraint_timeconst: Optional[float] = None
+ constraint_timeconst_min: float = 0.004 | genesis/options/solvers.py | 26 | Python | 0.143 | question | 37 | 51 | 51 | false | [MISC] Enable GJK collision detection algorithm by default. | 1,439 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | duburcqa | Review this pull request change.
Language: Python
File: genesis/options/solvers.py
PR Title: [MISC] Enable GJK collision detection algorithm by default.
BEFORE CODE:
dt: Optional[float] = None
gravity: Optional[tuple] = None
enable_collision: bool = True
enable_joint_limit: bool = True
enable_s... | can we change it to 0.01 to be safer? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
# Calculate enter point and normal
enter = tnear # if 0 <= tnear else tfar
normal = Vec3.zero()
normal.v[min_idx] = -np.sign(ray.direction.v[min_idx])
hit_pos = ray.origin + ray.direction * enter
return RayHit(enter, hit_pos, normal)
def __repr__(self) -> str:
... | Does not make sense. You should probably have some `BaseBoundingBox` abstraction. | if tfar < tnear or tnear < 0: # tfar < 0
return RayHit.no_hit()
# Calculate enter point and normal
enter = tnear # if 0 <= tnear else tfar
normal = Vec3.zero()
normal.v[min_idx] = -np.sign(ray.direction.v[min_idx])
hit_pos = ray.origin + ray.direction * en... | @@ -79,3 +72,33 @@ def from_center_and_size(cls, center: Vec3, size: Vec3) -> 'AABB':
max = center + 0.5 * size
bounds = np.stack((min.v, max.v))
return cls(bounds)
+
+
+class OBB(AABB): | genesis/ext/pyrender/interaction/aabb.py | 26 | Python | 0.429 | refactor | 81 | 51 | 46 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/aabb.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE:
# Calculate enter point and normal
enter = tnear # if 0 <= tnear else tfar
normal = Vec3.zero()
normal.v[m... | Does not make sense. You should probably have some `BaseBoundingBox` abstraction. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
def __repr__(self) -> str:
return f"AABB: Min({self.min.x}, {self.min.y}, {self.min.z}) Max({self.max.x}, {self.max.y}, {self.max.z})"
@classmethod
def from_min_max(cls, min: Vec3, max: Vec3) -> 'AABB':
bounds = np.stack((min.v, max.v))
return cls(bounds)
@classmethod
def ... | Do not compute inverse explicit. This is slow. You should rather implement `inv_transform_point`, `inv_transform_direction` | hit_pos = ray.origin + ray.direction * enter
return RayHit(enter, hit_pos, normal)
def __repr__(self) -> str:
return f"AABB: Min({self.min.x}, {self.min.y}, {self.min.z}) Max({self.max.x}, {self.max.y}, {self.max.z})"
@classmethod
def from_min_max(cls, min: Vec3, max: Vec3) -> 'AAB... | @@ -79,3 +72,33 @@ def from_center_and_size(cls, center: Vec3, size: Vec3) -> 'AABB':
max = center + 0.5 * size
bounds = np.stack((min.v, max.v))
return cls(bounds)
+
+
+class OBB(AABB):
+ pose: Pose
+
+ def __init__(self, v: NDArray[np.float32], pose: Pose):
+ super().__init__(v... | genesis/ext/pyrender/interaction/aabb.py | 26 | Python | 0.571 | performance | 123 | 46 | 38 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/aabb.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE:
def __repr__(self) -> str:
return f"AABB: Min({self.min.x}, {self.min.y}, {self.min.z}) Max({self.max.x}, {self.max.y}, {self.m... | Do not compute inverse explicit. This is slow. You should rather implement `inv_transform_point`, `inv_transform_direction` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
pose: Pose
def __init__(self, v: NDArray[np.float32], pose: Pose):
super().__init__(v)
self.pose = pose
def raycast(self, ray: Ray) -> RayHit:
inv_pose = self.pose.get_inverse()
origin2 = inv_pose.transform_point(ray.origin)
direction2 = inv_pose.transform_direction... | Always specify on which access you stack for readability. | from dataclasses import dataclass
import numpy as np
from numpy.typing import NDArray
from .ray import Ray, RayHit, EPSILON
from .vec3 import Pose, Vec3
class AABB:
v: NDArray[np.float32]
def __init__(self, v: NDArray[np.float32]):
assert v.shape == (2, 3,), f"Aabb must be initialized with a (2,3)-e... | @@ -79,3 +72,33 @@ def from_center_and_size(cls, center: Vec3, size: Vec3) -> 'AABB':
max = center + 0.5 * size
bounds = np.stack((min.v, max.v))
return cls(bounds)
+
+
+class OBB(AABB):
+ pose: Pose
+
+ def __init__(self, v: NDArray[np.float32], pose: Pose):
+ super().__init__(v... | genesis/ext/pyrender/interaction/aabb.py | 26 | Python | 0.214 | suggestion | 57 | 28 | 97 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/aabb.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE: pose: Pose
def __init__(self, v: NDArray[np.float32], pose: Pose):
super().__init__(v)
self.pose = pose
def ra... | Always specify on which access you stack for readability. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
MOUSE_SPRING_VELOCITY_CORRECTION_FACTOR = 1.0
def _ensure_torch_imported() -> None:
global torch
import torch
class MouseSpring:
def __init__(self):
self.held_link: RigidLink | None = None
self.held_point_in_local: Vec3 | None = None
self.prev_control_point: Vec3 | None = None
... | Could you confirm if it is fixed and clean these notes (which does NOT mean removing them entirely!) | MOUSE_SPRING_VELOCITY_CORRECTION_FACTOR = 1.0
def _ensure_torch_imported() -> None:
global torch
import torch
class MouseSpring:
def __init__(self) -> None:
self.held_link: RigidLink | None = None
self.held_point_in_local: Vec3 | None = None
self.prev_control_point: Vec3 | None = N... | @@ -16,65 +16,84 @@ def _ensure_torch_imported() -> None:
class MouseSpring:
def __init__(self):
- self.held_geom: RigidGeom | None = None
+ self.held_link: RigidLink | None = None
self.held_point_in_local: Vec3 | None = None
self.prev_control_point: Vec3 | None = None
- def... | genesis/ext/pyrender/interaction/mouse_spring.py | 26 | Python | 0.429 | suggestion | 100 | 51 | 51 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/mouse_spring.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE: MOUSE_SPRING_VELOCITY_CORRECTION_FACTOR = 1.0
def _ensure_torch_imported() -> None:
global torch
import torch
class MouseS... | Could you confirm if it is fixed and clean these notes (which does NOT mean removing them entirely!) | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self.held_point_in_local: Vec3 | None = None
self.prev_control_point: Vec3 | None = None
def attach(self, picked_link: RigidLink, control_point: Vec3):
# for now, we just pick the first geometry
self.held_link = picked_link
pose: Pose = Pose.from_link(self.held_link)
... | This makes no sense if the type of link is `RigidLink`. Either the assert should be removed or typing should be replaced by `RigidLink | None` | self.held_point_in_local: Vec3 | None = None
self.prev_control_point: Vec3 | None = None
def attach(self, picked_link: RigidLink, control_point: Vec3) -> None:
# for now, we just pick the first geometry
self.held_link = picked_link
pose: Pose = Pose.from_link(self.held_link)... | @@ -16,65 +16,84 @@ def _ensure_torch_imported() -> None:
class MouseSpring:
def __init__(self):
- self.held_geom: RigidGeom | None = None
+ self.held_link: RigidLink | None = None
self.held_point_in_local: Vec3 | None = None
self.prev_control_point: Vec3 | None = None
- def... | genesis/ext/pyrender/interaction/mouse_spring.py | 26 | Python | 0.714 | suggestion | 142 | 51 | 51 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/mouse_spring.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE: self.held_point_in_local: Vec3 | None = None
self.prev_control_point: Vec3 | None = None
def attach(self, picke... | This makes no sense if the type of link is `RigidLink`. Either the assert should be removed or typing should be replaced by `RigidLink | None` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self.held_link = picked_link
pose: Pose = Pose.from_link(self.held_link)
self.held_point_in_local = pose.inverse_transform_point(control_point)
self.prev_control_point = control_point
def detach(self):
self.held_link = None
def apply_force(self, control_point: Vec3, del... | If the simulation is built, `inertial_pos` and `inertial_quat` should not be none. We should update these property getters to read values from taichi field if the simulation is already built. See `sol_params` implementation: https://github.com/Genesis-Embodied-AI/Genesis/blob/767d536b8689cba5f382fe184df8bd270f889ea2/ge... | self.held_link = picked_link
pose: Pose = Pose.from_link(self.held_link)
self.held_point_in_local = pose.inverse_transform_point(control_point)
self.prev_control_point = control_point
def detach(self) -> None:
self.held_link = None
def apply_force(self, control_point: V... | @@ -16,65 +16,84 @@ def _ensure_torch_imported() -> None:
class MouseSpring:
def __init__(self):
- self.held_geom: RigidGeom | None = None
+ self.held_link: RigidLink | None = None
self.held_point_in_local: Vec3 | None = None
self.prev_control_point: Vec3 | None = None
- def... | genesis/ext/pyrender/interaction/mouse_spring.py | 26 | Python | 0.643 | suggestion | 374 | 51 | 51 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/mouse_spring.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE: self.held_link = picked_link
pose: Pose = Pose.from_link(self.held_link)
self.held_point_in_local = pose.inv... | If the simulation is built, `inertial_pos` and `inertial_quat` should not be none. We should update these property getters to read values from taichi field if the simulation is already built. See `sol_params` implementation: https://github.com/Genesis-Embodied-AI/Genesis/blob/767d536b8689cba5f382fe184df8bd270f889ea2/ge... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self.held_link = None
def apply_force(self, control_point: Vec3, delta_time: float):
_ensure_torch_imported()
# note when threaded: apply_force is called before attach!
# note2: that was before we added a lock to ViewerInteraction; this migth be fixed now
if not self.held_l... | Same. No `get_inverse`, this is slow due to memory allocation.
What about overloading operation "/" to multiply by inverse, à-la-Matlab. | self.held_link = None
def apply_force(self, control_point: Vec3, delta_time: float) -> None:
_ensure_torch_imported()
# note when threaded: apply_force is called before attach!
# note2: that was before we added a lock to ViewerInteraction; this migth be fixed now
if not sel... | @@ -16,65 +16,84 @@ def _ensure_torch_imported() -> None:
class MouseSpring:
def __init__(self):
- self.held_geom: RigidGeom | None = None
+ self.held_link: RigidLink | None = None
self.held_point_in_local: Vec3 | None = None
self.prev_control_point: Vec3 | None = None
- def... | genesis/ext/pyrender/interaction/mouse_spring.py | 26 | Python | 0.643 | performance | 139 | 51 | 51 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/mouse_spring.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE: self.held_link = None
def apply_force(self, control_point: Vec3, delta_time: float):
_ensure_torch_imported()
... | Same. No `get_inverse`, this is slow due to memory allocation.
What about overloading operation "/" to multiply by inverse, à-la-Matlab. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
# note when threaded: apply_force is called before attach!
# note2: that was before we added a lock to ViewerInteraction; this migth be fixed now
if not self.held_link:
return
self.prev_control_point = control_point
# do simple force on COM only:
li... | `inv_spherical_inertia = 0.0` make sense ? |
# note when threaded: apply_force is called before attach!
# note2: that was before we added a lock to ViewerInteraction; this migth be fixed now
if not self.held_link:
return
self.prev_control_point = control_point
# do simple force on COM only:
li... | @@ -16,65 +16,84 @@ def _ensure_torch_imported() -> None:
class MouseSpring:
def __init__(self):
- self.held_geom: RigidGeom | None = None
+ self.held_link: RigidLink | None = None
self.held_point_in_local: Vec3 | None = None
self.prev_control_point: Vec3 | None = None
- def... | genesis/ext/pyrender/interaction/mouse_spring.py | 26 | Python | 0.214 | question | 42 | 51 | 51 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/mouse_spring.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE:
# note when threaded: apply_force is called before attach!
# note2: that was before we added a lock to ViewerIntera... | `inv_spherical_inertia = 0.0` make sense ? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
# do simple force on COM only:
link: RigidLink = self.held_link
assert link
lin_vel: Vec3 = Vec3.from_tensor(link.get_vel())
ang_vel: Vec3 = Vec3.from_tensor(link.get_ang())
link_pose: Pose = Pose.from_link(link)
held_point_in_world: Vec3 = link_pose.transform_po... | I suggest to remove typing when it is irrelevant. We already know that `Vec3.zero()` has type `Vec3`. |
# do simple force on COM only:
link: RigidLink = self.held_link
lin_vel: Vec3 = Vec3.from_tensor(link.get_vel())
ang_vel: Vec3 = Vec3.from_tensor(link.get_ang())
link_pose: Pose = Pose.from_link(link)
held_point_in_world: Vec3 = link_pose.transform_point(self.held_point_... | @@ -16,65 +16,84 @@ def _ensure_torch_imported() -> None:
class MouseSpring:
def __init__(self):
- self.held_geom: RigidGeom | None = None
+ self.held_link: RigidLink | None = None
self.held_point_in_local: Vec3 | None = None
self.prev_control_point: Vec3 | None = None
- def... | genesis/ext/pyrender/interaction/mouse_spring.py | 26 | Python | 0.714 | suggestion | 101 | 51 | 51 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/mouse_spring.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE:
# do simple force on COM only:
link: RigidLink = self.held_link
assert link
lin_vel: Vec3 = Vec3.fr... | I suggest to remove typing when it is irrelevant. We already know that `Vec3.zero()` has type `Vec3`. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
@override
def on_resize(self, width: int, height: int) -> EVENT_HANDLE_STATE:
super().on_resize(width, height)
self.viewport_size = (width, height)
self.tan_half_fov = np.tan(0.5 * self.camera_yfov)
@override
def update_on_sim_step(self) -> None:
with self.lock:
... | `pos += delta_3d_pos` is more efficient. | @override
def on_resize(self, width: int, height: int) -> EVENT_HANDLE_STATE:
super().on_resize(width, height)
self.viewport_size = (width, height)
self.tan_half_fov = np.tan(0.5 * self.camera_yfov)
@override
def update_on_sim_step(self) -> None:
with self.lock:
... | @@ -96,64 +106,58 @@ def on_resize(self, width: int, height: int) -> EVENT_HANDLE_STATE:
@override
def update_on_sim_step(self) -> None:
- if self.picked_entity:
- mouse_ray: Ray = self.screen_position_to_ray(*self.prev_mouse_pos)
- ray_hit: RayHit = self.mouse_drag_plane.raycas... | genesis/ext/pyrender/interaction/viewer_interaction.py | 26 | Python | 0.214 | performance | 40 | 51 | 51 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/viewer_interaction.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE: @override
def on_resize(self, width: int, height: int) -> EVENT_HANDLE_STATE:
super().on_resize(width, height)... | `pos += delta_3d_pos` is more efficient. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self.picked_link = None
self.picked_point_in_local = None
self.mouse_drag_plane = None
self.prev_mouse_3d_pos = None
self.mouse_spring.detach()
@override
def on_resize(self, width: int, height: int) -> EVENT_HANDLE_STATE:
... | same. But you can keep them if you like them x) | self.picked_link = None
self.picked_point_in_local = None
self.mouse_drag_plane = None
self.prev_mouse_3d_pos = None
self.mouse_spring.detach()
@override
def on_resize(self, width: int, height: int) -> EVENT_HANDLE_STATE:
... | @@ -96,64 +106,58 @@ def on_resize(self, width: int, height: int) -> EVENT_HANDLE_STATE:
@override
def update_on_sim_step(self) -> None:
- if self.picked_entity:
- mouse_ray: Ray = self.screen_position_to_ray(*self.prev_mouse_pos)
- ray_hit: RayHit = self.mouse_drag_plane.raycas... | genesis/ext/pyrender/interaction/viewer_interaction.py | 26 | Python | 0.071 | suggestion | 47 | 51 | 51 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/viewer_interaction.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE: self.picked_link = None
self.picked_point_in_local = None
self.mouse_drag_plan... | same. But you can keep them if you like them x) | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
def get_camera_ray(self) -> Ray:
mtx = self.camera.matrix
position = Vec3.from_array(mtx[:3, 3])
forward = Vec3.from_array(-mtx[:3, 2])
return Ray(position, forward)
def _raycast_against_ground_plane(self, ray: Ray) -> RayHit:
ground_plane = Plane(Vec3.from_xyz(0, 0, 1)... | `if link.get_mass() > 0.0:` is more natural for me. |
def get_camera_ray(self) -> Ray:
mtx = self.camera.matrix
position = Vec3.from_array(mtx[:3, 3])
forward = Vec3.from_array(-mtx[:3, 2])
return Ray(position, forward)
def _raycast_against_ground_plane(self, ray: Ray) -> RayHit:
ground_plane = Plane(Vec3.from_xyz(0, 0, 1)... | @@ -197,53 +201,62 @@ def _raycast_against_ground_plane(self, ray: Ray) -> RayHit:
def raycast_against_entity_obb(self, entity: RigidEntity, ray: Ray) -> RayHit:
if isinstance(entity.morph, gs.morphs.Box):
- box: gs.morphs.Box = entity.morph
- size = Vec3.from_xyz(*box.size)
- ... | genesis/ext/pyrender/interaction/viewer_interaction.py | 26 | Python | 0.357 | suggestion | 51 | 51 | 51 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/viewer_interaction.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE:
def get_camera_ray(self) -> Ray:
mtx = self.camera.matrix
position = Vec3.from_array(mtx[:3, 3])
... | `if link.get_mass() > 0.0:` is more natural for me. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
return Vec3.from_array(-mtx[:3, 2])
def get_camera_ray(self) -> Ray:
mtx = self.camera.matrix
position = Vec3.from_array(mtx[:3, 3])
forward = Vec3.from_array(-mtx[:3, 2])
return Ray(position, forward)
def _raycast_against_ground_plane(self, ray: Ray) -> RayHit:
... | What does this comment means? Whether it can move ? If so, you need to change for its parent index. If it is "-1", then it is fixed wrt world (95% sure). Another option is to check the value of `is_fixed`, which is arguably much clearer. | return Vec3.from_array(-mtx[:3, 2])
def get_camera_ray(self) -> Ray:
mtx = self.camera.matrix
position = Vec3.from_array(mtx[:3, 3])
forward = Vec3.from_array(-mtx[:3, 2])
return Ray(position, forward)
def _raycast_against_ground_plane(self, ray: Ray) -> RayHit:
... | @@ -197,53 +201,62 @@ def _raycast_against_ground_plane(self, ray: Ray) -> RayHit:
def raycast_against_entity_obb(self, entity: RigidEntity, ray: Ray) -> RayHit:
if isinstance(entity.morph, gs.morphs.Box):
- box: gs.morphs.Box = entity.morph
- size = Vec3.from_xyz(*box.size)
- ... | genesis/ext/pyrender/interaction/viewer_interaction.py | 26 | Python | 0.643 | question | 237 | 51 | 51 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/viewer_interaction.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE: return Vec3.from_array(-mtx[:3, 2])
def get_camera_ray(self) -> Ray:
mtx = self.camera.matrix
pos... | What does this comment means? Whether it can move ? If so, you need to change for its parent index. If it is "-1", then it is fixed wrt world (95% sure). Another option is to check the value of `is_fixed`, which is arguably much clearer. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
for geom in link.geoms:
obb: OBB = self._get_geom_placeholder_obb(geom)
ray_hit = obb.raycast(ray)
if ray_hit.distance < closest_hit.distance:
ray_hit.geom = geom
closest_h... | Could you add `def full(fill_value):` ? That would be more convenient and efficient. | obb: OBB = self._get_geom_placeholder_obb(geom)
ray_hit = obb.raycast(ray)
if ray_hit.distance < closest_hit.distance:
ray_hit.geom = geom
closest_hit = ray_hit
return closest_hit
... | @@ -197,53 +201,62 @@ def _raycast_against_ground_plane(self, ray: Ray) -> RayHit:
def raycast_against_entity_obb(self, entity: RigidEntity, ray: Ray) -> RayHit:
if isinstance(entity.morph, gs.morphs.Box):
- box: gs.morphs.Box = entity.morph
- size = Vec3.from_xyz(*box.size)
- ... | genesis/ext/pyrender/interaction/viewer_interaction.py | 26 | Python | 0.429 | performance | 84 | 47 | 45 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/viewer_interaction.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE: for geom in link.geoms:
obb: OBB = self._get_geom_placeholder_obb(geom)
... | Could you add `def full(fill_value):` ? That would be more convenient and efficient. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
box: gs.morphs.Box = box_entity.morph
pose = Pose.from_link(box_entity.links[0])
size = Vec3.from_xyz(*box.size)
return OBB.from_center_and_size(size, pose)
def _get_geom_placeholder_obb(self, geom: 'RigidGeom') -> OBB:
pose = Pose.from_geom(geom)
size = 0.125 * Vec3... | Use keyword arguments when passing constants to methods. This makes the code more self-explanatory. | pose = Pose.from_link(box_entity.links[0])
half_extents = 0.5 * Vec3.from_xyz(*box.size)
return OBB(pose, half_extents)
def _get_geom_placeholder_obb(self, geom: 'RigidGeom') -> OBB:
pose = Pose.from_geom(geom)
half_extents = Vec3.full(0.5 * 0.125)
return OBB(pose, h... | @@ -197,53 +201,62 @@ def _raycast_against_ground_plane(self, ray: Ray) -> RayHit:
def raycast_against_entity_obb(self, entity: RigidEntity, ray: Ray) -> RayHit:
if isinstance(entity.morph, gs.morphs.Box):
- box: gs.morphs.Box = entity.morph
- size = Vec3.from_xyz(*box.size)
- ... | genesis/ext/pyrender/interaction/viewer_interaction.py | 26 | Python | 0.286 | suggestion | 99 | 29 | 27 | false | [MISC] Picking boxes and Panda links with a MouseSpring | 1,443 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | gasnica | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/interaction/viewer_interaction.py
PR Title: [MISC] Picking boxes and Panda links with a MouseSpring
BEFORE CODE: box: gs.morphs.Box = box_entity.morph
pose = Pose.from_link(box_entity.links[0])
size = Vec3.from_xyz(... | Use keyword arguments when passing constants to methods. This makes the code more self-explanatory. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
@staticmethod
def _get_trimesh_props(mesh, smooth=False, material=None):
"""Gets the vertex colors, texture coordinates, and material properties
from a :class:`~trimesh.base.Trimesh`.
"""
colors = None
texcoords = None
# If the trimesh visual is undefined, retur... | not sure if this is a correct fix. Basically, we will get an error if the color argument in draw_debug_arrow takes (r,g,b) instead of (r,g,b,a). You can reproduce it by reverting this change and running @duburcqa
```
pytest -s -v tests/test_rigid_physics.py::test_examples_api
```
|
@staticmethod
def _get_trimesh_props(mesh, smooth=False, material=None):
"""Gets the vertex colors, texture coordinates, and material properties
from a :class:`~trimesh.base.Trimesh`.
"""
colors = None
texcoords = None
# If the trimesh visual is undefined, retur... | @@ -266,7 +266,7 @@ def _get_trimesh_props(mesh, smooth=False, material=None):
else:
colors = vc[mesh.faces].reshape((3 * len(mesh.faces), vc.shape[1]))
material = MetallicRoughnessMaterial(
- alphaMode="OPAQUE" if (colors[..., 3] == 255)... | genesis/ext/pyrender/mesh.py | 26 | Python | 1 | bug | 288 | 51 | 51 | false | [BUG FIX] Fix errors in APIs that are used by rigid examples | 1,471 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | SonSang | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/mesh.py
PR Title: [BUG FIX] Fix errors in APIs that are used by rigid examples
BEFORE CODE:
@staticmethod
def _get_trimesh_props(mesh, smooth=False, material=None):
"""Gets the vertex colors, texture coordinates, and mater... | not sure if this is a correct fix. Basically, we will get an error if the color argument in draw_debug_arrow takes (r,g,b) instead of (r,g,b,a). You can reproduce it by reverting this change and running @duburcqa
```
pytest -s -v tests/test_rigid_physics.py::test_examples_api
```
| You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
scale=0.05,
),
surface=gs.surfaces.Default(color=(0.5, 0.5, 1.0, 1)),
)
########################## entities ##########################
robot = scene.add_entity(
morph=gs.morphs.URDF(
scale=1.0,
file="urdf/shadow_hand/shadow_hand.urdf",
),
... | do you know how to fix this non-continuous memory? Although it's not a breaking bug @duburcqa | scale=0.05,
),
surface=gs.surfaces.Default(color=(0.5, 0.5, 1.0, 1)),
)
########################## entities ##########################
robot = scene.add_entity(
morph=gs.morphs.URDF(
scale=1.0,
file="urdf/shadow_hand/shadow_hand.urdf",
),
... | @@ -66,24 +66,27 @@ def main():
target_quat = np.array([1, 0, 0, 0])
index_finger_distal = robot.get_link("index_finger_distal")
middle_finger_distal = robot.get_link("middle_finger_distal")
- forearm = robot.get_link("forearm")
+ wrist = robot.get_link("wrist") # there is no link named "forearm"
... | examples/rigid/ik_shadow_hand.py | 26 | Python | 0.357 | bug | 94 | 47 | 50 | false | [BUG FIX] Fix errors in APIs that are used by rigid examples | 1,471 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | SonSang | Review this pull request change.
Language: Python
File: examples/rigid/ik_shadow_hand.py
PR Title: [BUG FIX] Fix errors in APIs that are used by rigid examples
BEFORE CODE: scale=0.05,
),
surface=gs.surfaces.Default(color=(0.5, 0.5, 1.0, 1)),
)
########################## entities... | do you know how to fix this non-continuous memory? Although it's not a breaking bug @duburcqa | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
horizontal_scale=horizontal_scale,
vertical_scale=vertical_scale,
subterrain_types=[
["flat_terrain", "random_uniform_terrain"],
["pyramid_sloped_terrain", "discrete_obstacles_terrain"],
],
),
)
ball = scene.add_entity(
... | You should rather specify the device argument to torch.tensor | horizontal_scale=horizontal_scale,
vertical_scale=vertical_scale,
subterrain_types=[
["flat_terrain", "random_uniform_terrain"],
["pyramid_sloped_terrain", "discrete_obstacles_terrain"],
],
),
)
ball = scene.add_entity(
... | @@ -64,8 +64,8 @@ def main():
).unsqueeze(-1)
cols = horizontal_scale * torch.range(0, height_field.shape[1] - 1, 1).unsqueeze(0).repeat(
height_field.shape[0], 1
- ).unsqueeze(-1)
- heights = vertical_scale * torch.tensor(height_field).unsqueeze(-1)
+ ).unsqueeze(-1).to(rows.device) | examples/rigid/terrain_subterrain.py | 26 | Python | 0.286 | suggestion | 61 | 37 | 37 | false | [BUG FIX] Fix errors in APIs that are used by rigid examples | 1,471 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | SonSang | Review this pull request change.
Language: Python
File: examples/rigid/terrain_subterrain.py
PR Title: [BUG FIX] Fix errors in APIs that are used by rigid examples
BEFORE CODE: horizontal_scale=horizontal_scale,
vertical_scale=vertical_scale,
subterrain_types=[
["... | You should rather specify the device argument to torch.tensor | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
import math
import os
import sys
import tempfile
import xml.etree.ElementTree as ET
from pathlib import Path
import mujoco
import numpy as np
import pytest
import torch
import trimesh
import genesis as gs
import genesis.utils.geom as gu
from genesis.utils.misc import get_assets_dir, tensor_to_array
from .utils impor... | This unit test is not great, there is too much noise and the name is not clear. It should fit in 10 lines at most. | pos=(0.0, 0.0, 1.0),
contype=2,
conaffinity=2,
)
)
box3 = scene.add_entity(
gs.morphs.Box(
size=(0.5, 0.5, 0.5),
pos=(0.0, 0.0, 1.5),
contype=1,
conaffinity=1,
)
)
scene.build()
for _ in rang... | @@ -2751,3 +2751,101 @@ def test_contype_conaffinity(show_viewer, tol):
assert_allclose(box1.get_pos(), np.array([0.0, 0.0, 0.25]), atol=1e-3)
assert_allclose(box2.get_pos(), np.array([0.0, 0.0, 0.75]), atol=1e-3)
assert_allclose(box3.get_pos(), np.array([0.0, 0.0, 0.75]), atol=1e-3)
+
+
+@pytest.mark.pa... | tests/test_rigid_physics.py | 26 | Python | 0.429 | nitpick | 114 | 2,754 | 51 | false | [BUG FIX] Fix errors in APIs that are used by rigid examples | 1,471 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | SonSang | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [BUG FIX] Fix errors in APIs that are used by rigid examples
BEFORE CODE: import math
import os
import sys
import tempfile
import xml.etree.ElementTree as ET
from pathlib import Path
import mujoco
import numpy as np
impor... | This unit test is not great, there is too much noise and the name is not clear. It should fit in 10 lines at most. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
else:
gs.raise_exception(f"Unsupported MJCF joint type: {mj_type}")
j_info["type"], j_info["n_qs"], j_info["n_dofs"] = gs_type, n_qs, n_dofs
# Parsing joint parameters that are type-agnostic
mj_dof_offset = mj.jnt_dofadr[i_j] if i_j != -1 else 0
mj_qpos_offset = mj.j... | Just delete this instead of commenting out. | else:
gs.raise_exception(f"Unsupported MJCF joint type: {mj_type}")
j_info["type"], j_info["n_qs"], j_info["n_dofs"] = gs_type, n_qs, n_dofs
# Parsing joint parameters that are type-agnostic
mj_dof_offset = mj.jnt_dofadr[i_j] if i_j != -1 else 0
mj_qpos_offset = mj.j... | @@ -234,13 +234,14 @@ def parse_link(mj, i_l, scale):
j_info["dofs_damping"] = mj.dof_damping[mj_dof_offset : (mj_dof_offset + n_dofs)]
j_info["dofs_invweight"] = mj.dof_invweight0[mj_dof_offset : (mj_dof_offset + n_dofs)]
j_info["dofs_armature"] = mj.dof_armature[mj_dof_offset : (mj_dof_offs... | genesis/utils/mjcf.py | 26 | Python | 0.143 | suggestion | 43 | 51 | 51 | false | [Feature] Joint friction / frictionloss | 1,479 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/utils/mjcf.py
PR Title: [Feature] Joint friction / frictionloss
BEFORE CODE: else:
gs.raise_exception(f"Unsupported MJCF joint type: {mj_type}")
j_info["type"], j_info["n_qs"], j_info["n_dofs"] = gs_type, n_qs, n_dofs
... | Just delete this instead of commenting out. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
gs.morphs.Box(
size=(0.5, 0.5, 0.5),
pos=(0.0, 0.0, 1.5),
contype=1,
conaffinity=1,
)
)
scene.build()
for _ in range(100):
scene.step()
assert_allclose(box2.get_pos(), box3.get_pos(), atol=1e-3)
assert_allclose(box1.get_pos(),... | Yes, that would be great to find out what is happening here. | size=(1.0, 1.0, 1.0),
pos=(0.0, 0.0, 2.0),
)
)
scene.build()
for link in box.links:
for vgeom, geom in zip(link.vgeoms, link.geoms):
assert_allclose(geom.get_pos(), vgeom.get_pos(), atol=tol)
assert_allclose(geom.get_quat(), vgeom.get_quat(), ... | @@ -2806,3 +2806,14 @@ def test_contype_conaffinity(show_viewer, tol):
assert_allclose(box1.get_pos(), np.array([0.0, 0.0, 0.25]), atol=1e-3)
assert_allclose(box2.get_pos(), np.array([0.0, 0.0, 0.75]), atol=1e-3)
assert_allclose(box3.get_pos(), np.array([0.0, 0.0, 0.75]), atol=1e-3)
+
+
+@pytest.mark.par... | tests/test_rigid_physics.py | 26 | Python | 0.214 | suggestion | 60 | 28 | 51 | false | [Feature] Joint friction / frictionloss | 1,479 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [Feature] Joint friction / frictionloss
BEFORE CODE: gs.morphs.Box(
size=(0.5, 0.5, 0.5),
pos=(0.0, 0.0, 1.5),
contype=1,
conaffinity=1,
)
)
scene.bui... | Yes, that would be great to find out what is happening here. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
size=(0.5, 0.5, 0.5),
pos=(0.0, 0.0, 1.0),
contype=2,
conaffinity=2,
)
)
box3 = scene.add_entity(
gs.morphs.Box(
size=(0.5, 0.5, 0.5),
pos=(0.0, 0.0, 1.5),
contype=1,
conaffinity=1,
)
)
... | I would prefer a procedurally generated XML model for something this simple. | gravity=(0.0, 0.0, -10.0),
),
show_viewer=show_viewer,
)
box = scene.add_entity(
gs.morphs.Box(
size=(1.0, 1.0, 1.0),
pos=(0.0, 0.0, 2.0),
)
)
scene.build()
for link in box.links:
for vgeom, geom in zip(link.vgeoms, link.g... | @@ -2806,3 +2806,14 @@ def test_contype_conaffinity(show_viewer, tol):
assert_allclose(box1.get_pos(), np.array([0.0, 0.0, 0.25]), atol=1e-3)
assert_allclose(box2.get_pos(), np.array([0.0, 0.0, 0.75]), atol=1e-3)
assert_allclose(box3.get_pos(), np.array([0.0, 0.0, 0.75]), atol=1e-3)
+
+
+@pytest.mark.par... | tests/test_rigid_physics.py | 26 | Python | 0.286 | suggestion | 76 | 35 | 51 | false | [Feature] Joint friction / frictionloss | 1,479 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [Feature] Joint friction / frictionloss
BEFORE CODE: size=(0.5, 0.5, 0.5),
pos=(0.0, 0.0, 1.0),
contype=2,
conaffinity=2,
)
)
box3 = scene.add_entity(
... | I would prefer a procedurally generated XML model for something this simple. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
box3 = scene.add_entity(
gs.morphs.Box(
size=(0.5, 0.5, 0.5),
pos=(0.0, 0.0, 1.5),
contype=1,
conaffinity=1,
)
)
scene.build()
for _ in range(100):
scene.step()
assert_allclose(box2.get_pos(), box3.get_pos(), atol=1e-3)
as... | Some initial velocity would be better. | gs.morphs.Box(
size=(1.0, 1.0, 1.0),
pos=(0.0, 0.0, 2.0),
)
)
scene.build()
for link in box.links:
for vgeom, geom in zip(link.vgeoms, link.geoms):
assert_allclose(geom.get_pos(), vgeom.get_pos(), atol=tol)
assert_allclose(geom.get_qua... | @@ -2806,3 +2806,14 @@ def test_contype_conaffinity(show_viewer, tol):
assert_allclose(box1.get_pos(), np.array([0.0, 0.0, 0.25]), atol=1e-3)
assert_allclose(box2.get_pos(), np.array([0.0, 0.0, 0.75]), atol=1e-3)
assert_allclose(box3.get_pos(), np.array([0.0, 0.0, 0.75]), atol=1e-3)
+
+
+@pytest.mark.par... | tests/test_rigid_physics.py | 26 | Python | 0.143 | suggestion | 38 | 29 | 51 | false | [Feature] Joint friction / frictionloss | 1,479 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [Feature] Joint friction / frictionloss
BEFORE CODE: box3 = scene.add_entity(
gs.morphs.Box(
size=(0.5, 0.5, 0.5),
pos=(0.0, 0.0, 1.5),
contype=1,
conaffinity=1,
... | Some initial velocity would be better. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
)
box3 = scene.add_entity(
gs.morphs.Box(
size=(0.5, 0.5, 0.5),
pos=(0.0, 0.0, 1.5),
contype=1,
conaffinity=1,
)
)
scene.build()
for _ in range(100):
scene.step()
assert_allclose(box2.get_pos(), box3.get_pos(), atol=1e-3)
... | No need to specify qpos if it is zero. | box = scene.add_entity(
gs.morphs.Box(
size=(1.0, 1.0, 1.0),
pos=(0.0, 0.0, 2.0),
)
)
scene.build()
for link in box.links:
for vgeom, geom in zip(link.vgeoms, link.geoms):
assert_allclose(geom.get_pos(), vgeom.get_pos(), atol=tol)
... | @@ -2806,3 +2806,14 @@ def test_contype_conaffinity(show_viewer, tol):
assert_allclose(box1.get_pos(), np.array([0.0, 0.0, 0.25]), atol=1e-3)
assert_allclose(box2.get_pos(), np.array([0.0, 0.0, 0.75]), atol=1e-3)
assert_allclose(box3.get_pos(), np.array([0.0, 0.0, 0.75]), atol=1e-3)
+
+
+@pytest.mark.par... | tests/test_rigid_physics.py | 26 | Python | 0.071 | suggestion | 38 | 30 | 51 | false | [Feature] Joint friction / frictionloss | 1,479 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [Feature] Joint friction / frictionloss
BEFORE CODE: )
box3 = scene.add_entity(
gs.morphs.Box(
size=(0.5, 0.5, 0.5),
pos=(0.0, 0.0, 1.5),
contype=1,
conaffini... | No need to specify qpos if it is zero. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
def func_update_constraint(
i_b,
qacc: array_class.V_ANNOTATION,
Ma: array_class.V_ANNOTATION,
cost: array_class.V_ANNOTATION,
dofs_state: array_class.DofsState,
constraint_state: array_class.ConstraintState,
static_rigid_sim_config: ti.template(),
):
n_dofs = constraint_state.qfrc_const... | What about defining 'rf' as you did above ? | cost: array_class.V_ANNOTATION,
dofs_state: array_class.DofsState,
constraint_state: array_class.ConstraintState,
static_rigid_sim_config: ti.template(),
):
n_dofs = constraint_state.qfrc_constraint.shape[0]
constraint_state.prev_cost[i_b] = cost[i_b]
cost[i_b] = gs.ti_float(0.0)
constr... | @@ -1681,9 +1716,23 @@ def func_update_constraint(
if ti.static(static_rigid_sim_config.solver_type == gs.constraint_solver.Newton):
constraint_state.prev_active[i_c, i_b] = constraint_state.active[i_c, i_b]
constraint_state.active[i_c, i_b] = 1
- if i_c >= constraint_state.n_const... | genesis/engine/solvers/rigid/constraint_solver_decomp.py | 26 | Python | 0.071 | question | 43 | 51 | 51 | false | [Feature] Joint friction / frictionloss | 1,479 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_solver_decomp.py
PR Title: [Feature] Joint friction / frictionloss
BEFORE CODE: def func_update_constraint(
i_b,
qacc: array_class.V_ANNOTATION,
Ma: array_class.V_ANNOTATION,
cost: array_class.V_ANNOTATIO... | What about defining 'rf' as you did above ? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
cost: array_class.V_ANNOTATION,
dofs_state: array_class.DofsState,
constraint_state: array_class.ConstraintState,
static_rigid_sim_config: ti.template(),
):
n_dofs = constraint_state.qfrc_constraint.shape[0]
constraint_state.prev_cost[i_b] = cost[i_b]
cost[i_b] = gs.ti_float(0.0)
constr... | Move 'f' outside for efficiency and consistency. | ):
n_dofs = constraint_state.qfrc_constraint.shape[0]
constraint_state.prev_cost[i_b] = cost[i_b]
cost[i_b] = gs.ti_float(0.0)
constraint_state.gauss[i_b] = gs.ti_float(0.0)
for i_c in range(constraint_state.n_constraints[i_b]):
if ti.static(static_rigid_sim_config.solver_type == gs.constr... | @@ -1681,9 +1716,23 @@ def func_update_constraint(
if ti.static(static_rigid_sim_config.solver_type == gs.constraint_solver.Newton):
constraint_state.prev_active[i_c, i_b] = constraint_state.active[i_c, i_b]
constraint_state.active[i_c, i_b] = 1
- if i_c >= constraint_state.n_const... | genesis/engine/solvers/rigid/constraint_solver_decomp.py | 26 | Python | 0.143 | style | 48 | 51 | 51 | false | [Feature] Joint friction / frictionloss | 1,479 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_solver_decomp.py
PR Title: [Feature] Joint friction / frictionloss
BEFORE CODE: cost: array_class.V_ANNOTATION,
dofs_state: array_class.DofsState,
constraint_state: array_class.ConstraintState,
static_rig... | Move 'f' outside for efficiency and consistency. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
tmp_quad_total0 = constraint_state.quad_gauss[_i0 + 0, i_b]
tmp_quad_total1 = constraint_state.quad_gauss[_i0 + 1, i_b]
tmp_quad_total2 = constraint_state.quad_gauss[_i0 + 2, i_b]
for i_c in range(constraint_state.n_constraints[i_b]):
x = constraint_state.Jaref[i_c, i_b] + al... | You are setting value equal to zero. Just do this directly. Moreover, I don't this it is worth making `qf` a vector. Just define `qf_1` and `qf_2`. That would be more efficient. | tmp_quad_total1 = constraint_state.quad_gauss[1, i_b]
tmp_quad_total2 = constraint_state.quad_gauss[2, i_b]
for i_c in range(constraint_state.n_constraints[i_b]):
x = constraint_state.Jaref[i_c, i_b] + alpha * constraint_state.jv[i_c, i_b]
active = 1
ne = constraint_state.n_constrain... | @@ -1295,9 +1307,32 @@ def func_ls_point_fn(
tmp_quad_total1 = constraint_state.quad_gauss[_i0 + 1, i_b]
tmp_quad_total2 = constraint_state.quad_gauss[_i0 + 2, i_b]
for i_c in range(constraint_state.n_constraints[i_b]):
+ x = constraint_state.Jaref[i_c, i_b] + alpha * constraint_st... | genesis/engine/solvers/rigid/constraint_solver_decomp.py | 26 | Python | 0.5 | performance | 177 | 51 | 51 | false | [Feature] Joint friction / frictionloss | 1,479 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_solver_decomp.py
PR Title: [Feature] Joint friction / frictionloss
BEFORE CODE: tmp_quad_total0 = constraint_state.quad_gauss[_i0 + 0, i_b]
tmp_quad_total1 = constraint_state.quad_gauss[_i0 + 1, i_b]
... | You are setting value equal to zero. Just do this directly. Moreover, I don't this it is worth making `qf` a vector. Just define `qf_1` and `qf_2`. That would be more efficient. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self._sim = sim
self._scene = scene
self._dt: float = options.dt
self._substep_dt: float = options.dt / sim.substeps
self._init_gravity = getattr(options, "gravity", None)
self._gravity = None
self._entities: list[Entity] = gs.List()
# force fields
... | `gs.logger.debug` should be enough.
"Gravity is not defined, skipping 'set_gravity'." would be more appropriate. | self._sim = sim
self._scene = scene
self._dt: float = options.dt
self._substep_dt: float = options.dt / sim.substeps
self._init_gravity = getattr(options, "gravity", None)
self._gravity = None
self._entities: list[Entity] = gs.List()
# force fields
... | @@ -42,14 +42,26 @@ def build(self):
@gs.assert_built
def set_gravity(self, gravity, envs_idx=None):
if self._gravity is None:
+ gs.logger.warning("Gravity is not initialized, skipping set_gravity.") | genesis/engine/solvers/base_solver.py | 26 | Python | 0.714 | bug | 115 | 51 | 51 | false | [BUG FIX] Fix gravity setters. | 1,498 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/engine/solvers/base_solver.py
PR Title: [BUG FIX] Fix gravity setters.
BEFORE CODE: self._sim = sim
self._scene = scene
self._dt: float = options.dt
self._substep_dt: float = options.dt / sim.substeps
self._init_... | `gs.logger.debug` should be enough.
"Gravity is not defined, skipping 'set_gravity'." would be more appropriate. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
file="urdf/panda_bullet/hand.urdf",
fixed=True,
),
)
scene.build()
assert scene.rigid_solver.n_equalities == 1
qvel = scene.rigid_solver.dofs_state.vel.to_numpy()
qvel[-1] = 1
scene.rigid_solver.dofs_state.vel.from_numpy(qvel)
for i in range(200):
sce... | Prefer `n_envs=2`, or `n_envs=3`, it should be enough. | file="urdf/panda_bullet/hand.urdf",
fixed=True,
),
)
scene.build()
assert scene.rigid_solver.n_equalities == 1
qvel = scene.rigid_solver.dofs_state.vel.to_numpy()
qvel[-1] = 1
scene.rigid_solver.dofs_state.vel.from_numpy(qvel)
for i in range(200):
sce... | @@ -2177,18 +2177,38 @@ def test_gravity(show_viewer, tol):
)
sphere = scene.add_entity(gs.morphs.Sphere())
- scene.build(n_envs=2)
+ scene.build(n_envs=7) | tests/test_rigid_physics.py | 26 | Python | 0.429 | suggestion | 54 | 51 | 51 | false | [BUG FIX] Fix gravity setters. | 1,498 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [BUG FIX] Fix gravity setters.
BEFORE CODE: file="urdf/panda_bullet/hand.urdf",
fixed=True,
),
)
scene.build()
assert scene.rigid_solver.n_equalities == 1
qvel = scene.rigid... | Prefer `n_envs=2`, or `n_envs=3`, it should be enough. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
scene.sim.set_gravity(torch.tensor([0.0, 0.0, 0.0]), envs_idx=[0, 1])
scene.sim.set_gravity(torch.tensor([0.0, 0.0, 100.0]), envs_idx=3)
scene.sim.set_gravity(torch.tensor([[0.0, 0.0, -10.0], [0.0, 0.0, -1.0]]), envs_idx=[2, 4])
with np.testing.assert_raises(AssertionError):
scene.sim.set_gravi... | You should probably check the acceleration instead. | scene.sim.set_gravity(torch.tensor([0.0, 0.0, 3.0]), envs_idx=2)
with np.testing.assert_raises(AssertionError):
scene.sim.set_gravity(torch.tensor([0.0, -10.0]))
with np.testing.assert_raises(AssertionError):
scene.sim.set_gravity(torch.tensor([[0.0, 0.0, -10.0], [0.0, 0.0, -10.0]]), envs_... | @@ -2177,18 +2177,38 @@ def test_gravity(show_viewer, tol):
)
sphere = scene.add_entity(gs.morphs.Sphere())
- scene.build(n_envs=2)
+ scene.build(n_envs=7)
- scene.sim.set_gravity(torch.tensor([0.0, 0.0, -9.8]), envs_idx=0)
- scene.sim.set_gravity(torch.tensor([0.0, 0.0, 9.8]), envs_idx=1)
+ ... | tests/test_rigid_physics.py | 26 | Python | 0.286 | suggestion | 51 | 51 | 51 | false | [BUG FIX] Fix gravity setters. | 1,498 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [BUG FIX] Fix gravity setters.
BEFORE CODE: scene.sim.set_gravity(torch.tensor([0.0, 0.0, 0.0]), envs_idx=[0, 1])
scene.sim.set_gravity(torch.tensor([0.0, 0.0, 100.0]), envs_idx=3)
scene.sim.set_gravity(torch.t... | You should probably check the acceleration instead. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
scene.sim.set_gravity(torch.tensor([[-2.0, 0.0, 0.0]] * 10))
scene.sim.set_gravity(torch.tensor([0.0, 0.0, 0.0]), envs_idx=[0, 1])
scene.sim.set_gravity(torch.tensor([0.0, 0.0, 100.0]), envs_idx=3)
scene.sim.set_gravity(torch.tensor([[0.0, 0.0, -10.0], [0.0, 0.0, -1.0]]), envs_idx=[2, 4])
with np.t... | Are you sure squeezing is necessary? | scene.sim.set_gravity(torch.tensor([[1.0, 0.0, 0.0], [0.0, 2.0, 0.0]]), envs_idx=[0, 1])
scene.sim.set_gravity(torch.tensor([0.0, 0.0, 3.0]), envs_idx=2)
with np.testing.assert_raises(AssertionError):
scene.sim.set_gravity(torch.tensor([0.0, -10.0]))
with np.testing.assert_raises(AssertionErro... | @@ -2177,18 +2177,37 @@ def test_gravity(show_viewer, tol):
)
sphere = scene.add_entity(gs.morphs.Sphere())
- scene.build(n_envs=2)
+ scene.build(n_envs=7)
- scene.sim.set_gravity(torch.tensor([0.0, 0.0, -9.8]), envs_idx=0)
- scene.sim.set_gravity(torch.tensor([0.0, 0.0, 9.8]), envs_idx=1)
+ ... | tests/test_rigid_physics.py | 26 | Python | 0.071 | question | 36 | 51 | 51 | false | [BUG FIX] Fix gravity setters. | 1,498 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [BUG FIX] Fix gravity setters.
BEFORE CODE: scene.sim.set_gravity(torch.tensor([[-2.0, 0.0, 0.0]] * 10))
scene.sim.set_gravity(torch.tensor([0.0, 0.0, 0.0]), envs_idx=[0, 1])
scene.sim.set_gravity(torch.tensor(... | Are you sure squeezing is necessary? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self._idx: int = sensor_idx
self._manager: SensorManager = sensor_manager
self._cache_idx: int = -1 # cache_idx is set by the SensorManager during the scene build phase
# =============================== implementable methods ===============================
@gs.assert_unbuilt
def b... | 0 by default; something like lidar have `n`, and imu has `0` (special return processing) |
# initialized by SensorManager during build
self._read_delay_steps: int = 0
self._shape_indices: list[tuple[int, int]] = []
self._shared_metadata: dict[str, Any] | None = None
self._cache: "TensorRingBuffer" | None = None
# =============================== implementable meth... | @@ -12,10 +17,62 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ # These class variables are used by SensorManager to determine the cache metadata for the sensor.
+ # Sensor implementations should override these class variable values.
+ CACHE_DTYPE: torc... | genesis/sensors/base_sensor.py | 26 | Python | 0.357 | suggestion | 88 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | Milotrince | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: self._idx: int = sensor_idx
self._manager: SensorManager = sensor_manager
self._cache_idx: int = -1 # cache_idx is set by the SensorManager d... | 0 by default; something like lidar have `n`, and imu has `0` (special return processing) | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
"""
def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
self._options: "SensorOptions" = sensor_options
self._idx: int = sensor_idx
self._manager: "SensorManager" = sensor_manager
self._cache_idx: int = -1 # cache_idx is set by... | I don' think it is the right way to do it. The cache buffer length should be the same for all sensors, and the sensor manager should use `read_delay_step = max(sensor._read_delay_steps for sensor in self._sensors)` to determine how many data mush be stored in cache. It will make everything much easier to implement and ... |
def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
self._options: "SensorOptions" = sensor_options
self._idx: int = sensor_idx
self._manager: "SensorManager" = sensor_manager
# initialized by SensorManager during build
sel... | @@ -12,10 +21,127 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.786 | performance | 377 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: """
def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
self._options: "SensorOptions" = sensor_op... | I don' think it is the right way to do it. The cache buffer length should be the same for all sensors, and the sensor manager should use `read_delay_step = max(sensor._read_delay_steps for sensor in self._sensors)` to determine how many data mush be stored in cache. It will make everything much easier to implement and ... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
def build(self):
"""
This method is called by the SensorManager during the scene build phase to initialize the sensor.
This is where any shared metadata should be initialized.
"""
delay_steps_float = self._options.read_delay / self._manager._sim.dt
self._read_delay_st... | `gt` is not a great acronym because it usually means "greater than". What about 'true' or 'raw' ? Or just 'ground_truth' |
def build(self):
"""
This method is called by the SensorManager during the scene build phase to initialize the sensor.
This is where any shared metadata should be initialized.
"""
raise NotImplementedError("Sensors must implement `build()`.")
def _get_return_format(self... | @@ -12,10 +21,127 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.571 | suggestion | 120 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: def build(self):
"""
This method is called by the SensorManager during the scene build phase to initialize the sensor.
This is where any s... | `gt` is not a great acronym because it usually means "greater than". What about 'true' or 'raw' ? Or just 'ground_truth' | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
if not np.isclose(delay_steps_float, self._read_delay_steps):
gs.logger.warn(
f"Read delay should be a multiple of the simulation time step. Got {self._options.read_delay} and "
f"{self._manager._sim.dt}. Actual read delay will be {1/self._read_delay_steps}."
... | No need to assert build for private methods I think but as you want. |
def _get_return_format(self) -> dict[str, tuple[int, ...]] | tuple[int, ...]:
"""
Data format of the read() return value.
Returns
-------
return_format : dict | tuple
- If tuple, the final shape of the read() return value.
e.g. (2, 3) means read(... | @@ -12,10 +21,127 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.214 | suggestion | 68 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: if not np.isclose(delay_steps_float, self._read_delay_steps):
gs.logger.warn(
f"Read delay should be a multiple of the simulation ... | No need to assert build for private methods I think but as you want. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
def _get_return_format(self) -> dict[str, tuple[int, int]] | tuple[int, int] | None:
"""
Data format of the read() return value.
dict: a dictionary with string keys and tensor values will be returned.
The tuple (start_idx, end_idx) will be used to get a slice of the values from... | You should describe what the cache length means, eventually with examples. | return_format : dict | tuple
- If tuple, the final shape of the read() return value.
e.g. (2, 3) means read() will return a tensor of shape (2, 3).
- If dict a dictionary with string keys and tensor values will be returned.
e.g. {"pos": (3,), "quat": (4,)}... | @@ -12,10 +21,127 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.286 | performance | 74 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: def _get_return_format(self) -> dict[str, tuple[int, int]] | tuple[int, int] | None:
"""
Data format of the read() return value.
dict: a d... | You should describe what the cache length means, eventually with examples. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
self._options: "SensorOptions" = sensor_options
self._idx: int = sensor_idx
self._manager: "SensorManager" = sensor_manager
self._cache_idx: int = -1 # cache_idx is set by the Sens... | You should provide some concrete examples.
Besides, the formatting of your docstring is not correct. | self._options: "SensorOptions" = sensor_options
self._idx: int = sensor_idx
self._manager: "SensorManager" = sensor_manager
# initialized by SensorManager during build
self._read_delay_steps: int = 0
self._shape_indices: list[tuple[int, int]] = []
self._shared_me... | @@ -12,10 +21,127 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.429 | style | 103 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
self._options: "SensorOptions" = sensor_options
... | You should provide some concrete examples.
Besides, the formatting of your docstring is not correct. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
This is where any shared metadata should be initialized.
"""
delay_steps_float = self._options.read_delay / self._manager._sim.dt
self._read_delay_steps = round(delay_steps_float)
if not np.isclose(delay_steps_float, self._read_delay_steps):
gs.logger.warn(
... | Skip line between short documentation and "details". | This method is called by the SensorManager during the scene build phase to initialize the sensor.
This is where any shared metadata should be initialized.
"""
raise NotImplementedError("Sensors must implement `build()`.")
def _get_return_format(self) -> dict[str, tuple[int, ...]] | ... | @@ -12,10 +21,127 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.214 | suggestion | 52 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: This is where any shared metadata should be initialized.
"""
delay_steps_float = self._options.read_delay / self._manager._sim.dt
self... | Skip line between short documentation and "details". | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
raise NotImplementedError("Sensors must implement `return_format()`.")
@gs.assert_built
def _update_shared_gt_cache(self):
"""
Update the shared sensor ground truth cache for all sensors of this class.
Shared information is stored in SensorManager
"""
raise NotIm... | Having 'get_cache_size', 'get_cache_length', and 'get_return_format' is too much. Let's discuss this together. | def _get_cache_length(self) -> int:
"""
The length of the cache for this sensor instance, e.g. number of points for a Lidar point cloud.
"""
raise NotImplementedError("Sensors must implement `cache_length()`.")
@classmethod
def _update_shared_ground_truth_cache(
cls,... | @@ -12,10 +21,127 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.357 | performance | 110 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: raise NotImplementedError("Sensors must implement `return_format()`.")
@gs.assert_built
def _update_shared_gt_cache(self):
"""
Update... | Having 'get_cache_size', 'get_cache_length', and 'get_return_format' is too much. Let's discuss this together. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
from typing import Any, Type, TYPE_CHECKING
import torch
from genesis.utils.ring_buffer import TensorRingBuffer
if TYPE_CHECKING:
from genesis.options.sensors import SensorOptions
from .base_sensor import Sensor
class SensorManager:
SENSOR_TYPES_MAP: dict[Type["SensorOptions"], Type["Sensor"]] = {}
... | For cache could be global, and sensor-type based. it is the big advantage of storing data flattened. But you need one cache per datatype still. | from typing import TYPE_CHECKING, Any, Type
import numpy as np
import torch
import genesis as gs
from genesis.utils.ring_buffer import TensorRingBuffer
if TYPE_CHECKING:
from genesis.options.sensors import SensorOptions
from .base_sensor import Sensor
class SensorManager:
SENSOR_TYPES_MAP: dict[Type["... | @@ -0,0 +1,63 @@
+from typing import Any, Type, TYPE_CHECKING
+
+import torch
+
+from genesis.utils.ring_buffer import TensorRingBuffer
+
+if TYPE_CHECKING:
+ from genesis.options.sensors import SensorOptions
+ from .base_sensor import Sensor
+
+
+class SensorManager:
+ SENSOR_TYPES_MAP: dict[Type["SensorOptio... | genesis/sensors/sensor_manager.py | 19 | Python | 0.571 | performance | 143 | 44 | 44 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/sensor_manager.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: from typing import Any, Type, TYPE_CHECKING
import torch
from genesis.utils.ring_buffer import TensorRingBuffer
if TYPE_CHECKING:
from genesis.options.senso... | For cache could be global, and sensor-type based. it is the big advantage of storing data flattened. But you need one cache per datatype still. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
"""
Base class for all types of sensors.
A sensor must have a read() method that returns the sensor data.
"""
def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
self._options: "SensorOptions" = sensor_options
self._idx: int = senso... | You should probably set the absolute tolerance yourself. | """
Base class for all types of sensors.
"""
def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
self._options: "SensorOptions" = sensor_options
self._idx: int = sensor_idx
self._manager: "SensorManager" = sensor_manager
... | @@ -12,10 +21,138 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.286 | suggestion | 56 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: """
Base class for all types of sensors.
A sensor must have a read() method that returns the sensor data.
"""
def __init__(self, sensor_options: ... | You should probably set the absolute tolerance yourself. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
"""
def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
self._options: "SensorOptions" = sensor_options
self._idx: int = sensor_idx
self._manager: "SensorManager" = sensor_manager
# initialized during build
self._cache_... | `self._read_delay_steps * self._manager._sim.dt` instead of `1/self._read_delay_steps` ? |
def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
self._options: "SensorOptions" = sensor_options
self._idx: int = sensor_idx
self._manager: "SensorManager" = sensor_manager
# initialized by SensorManager during build
sel... | @@ -12,10 +21,138 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.429 | question | 88 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: """
def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
self._options: "SensorOptions" = sensor_op... | `self._read_delay_steps * self._manager._sim.dt` instead of `1/self._read_delay_steps` ? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self._idx: int = sensor_idx
self._manager: "SensorManager" = sensor_manager
# initialized during build
self._cache_idx: int = -1 # set by the SensorManager during build
self._read_delay_steps: int = 0
self._cache_size: int = 0
# =============================== impl... | Usually the number of elements of an array is called `size`, not `length` | self._manager: "SensorManager" = sensor_manager
# initialized by SensorManager during build
self._read_delay_steps: int = 0
self._shape_indices: list[tuple[int, int]] = []
self._shared_metadata: dict[str, Any] | None = None
self._cache: "TensorRingBuffer" | None = None
... | @@ -12,10 +21,138 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.357 | suggestion | 73 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: self._idx: int = sensor_idx
self._manager: "SensorManager" = sensor_manager
# initialized during build
self._cache_idx: int = -1 # s... | Usually the number of elements of an array is called `size`, not `length` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
def _update_shared_cache(self):
"""
Update the shared sensor cache for all sensors of this class using metadata in SensorManager.
"""
raise NotImplementedError("Sensors must implement `update_shared_cache()`.")
def _get_cache_length(self) -> int:
"""
The length o... | This is weird. What is `self._cache` ? It is not declared anywhere. | @classmethod
def _get_cache_dtype(cls) -> torch.dtype:
"""
The dtype of the cache for this sensor.
"""
raise NotImplementedError("Sensors must implement `get_cache_dtype()`.")
# =============================== shared methods ===============================
@gs.assert_bu... | @@ -12,10 +21,138 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.357 | performance | 67 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: def _update_shared_cache(self):
"""
Update the shared sensor cache for all sensors of this class using metadata in SensorManager.
"""
... | This is weird. What is `self._cache` ? It is not declared anywhere. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
def _get_cache_dtype(self) -> torch.dtype:
"""
The dtype of the cache for this sensor.
"""
raise NotImplementedError("Sensors must implement `get_cache_dtype()`.")
# =============================== shared methods ===============================
@gs.assert_built
def rea... | Could you preallocate the full vector of environment as it is done in `Scene` ? For maximum efficiency is should be a torch tensor on device if not a scalar. | """
Read the sensor data (with noise applied if applicable).
"""
return self._get_formatted_data(self._cache.get(self._read_delay_steps), envs_idx)
@gs.assert_built
def read_ground_truth(self, envs_idx: List[int] | None = None):
"""
Read the ground truth sensor d... | @@ -12,10 +21,138 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.571 | performance | 157 | 51 | 40 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE:
def _get_cache_dtype(self) -> torch.dtype:
"""
The dtype of the cache for this sensor.
"""
raise NotImplementedError("Sensors mus... | Could you preallocate the full vector of environment as it is done in `Scene` ? For maximum efficiency is should be a torch tensor on device if not a scalar. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
envs_idx = 0 if self._manager._sim.n_envs == 0 else np.arange(tensor.shape[0])
return_format = self._get_return_format()
if isinstance(return_format, tuple):
return (
tensor[envs_idx, self._cache_idx : self._cache_idx + self._cache_size]
.clone()... | Remove this. You already know that the format is valid at the point. it was checked at build time. | from typing import TYPE_CHECKING, Any, List, Type
import numpy as np
import taichi as ti
import torch
import genesis as gs
from genesis.repr_base import RBC
if TYPE_CHECKING:
from genesis.options.sensors import SensorOptions
from genesis.utils.ring_buffer import TensorRingBuffer
from .sensor_manager imp... | @@ -12,10 +21,138 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.286 | refactor | 98 | 43 | 131 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: envs_idx = 0 if self._manager._sim.n_envs == 0 else np.arange(tensor.shape[0])
return_format = self._get_return_format()
if isinstance(r... | Remove this. You already know that the format is valid at the point. it was checked at build time. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
"""
raise NotImplementedError("Sensors must implement `cache_length()`.")
def _get_cache_dtype(self) -> torch.dtype:
"""
The dtype of the cache for this sensor.
"""
raise NotImplementedError("Sensors must implement `get_cache_dtype()`.")
# ======================... | This is basically `_get_unflattened` ? More common terminology. | @gs.assert_built
def read(self, envs_idx: List[int] | None = None):
"""
Read the sensor data (with noise applied if applicable).
"""
return self._get_formatted_data(self._cache.get(self._read_delay_steps), envs_idx)
@gs.assert_built
def read_ground_truth(self, envs_idx: ... | @@ -12,10 +21,138 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.357 | suggestion | 63 | 51 | 42 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: """
raise NotImplementedError("Sensors must implement `cache_length()`.")
def _get_cache_dtype(self) -> torch.dtype:
"""
The dtyp... | This is basically `_get_unflattened` ? More common terminology. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
# =============================== shared methods ===============================
@gs.assert_built
def read(self, envs_idx: List[int] | None = None):
"""
Read the sensor data (with noise applied if applicable).
"""
return self._get_return_data_from_tensor(self._cache.get(self... | Just assume that the input data was cloned before calling this function. (and add this detail in the documentation) | """
Read the ground truth sensor data (without noise).
"""
return self._get_formatted_data(self._manager.get_cloned_from_ground_truth_cache(self), envs_idx)
def _get_formatted_data(
self, tensor: torch.Tensor, envs_idx: list[int] | None
) -> torch.Tensor | dict[str, torc... | @@ -12,10 +21,138 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.357 | suggestion | 115 | 51 | 33 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: # =============================== shared methods ===============================
@gs.assert_built
def read(self, envs_idx: List[int] | None = None):
... | Just assume that the input data was cloned before calling this function. (and add this detail in the documentation) | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
@gs.assert_built
def read(self, envs_idx: List[int] | None = None):
"""
Read the sensor data (with noise applied if applicable).
"""
return self._get_return_data_from_tensor(self._cache.get(self._read_delay_steps), envs_idx)
@gs.assert_built
def read_ground_truth(self, e... | You should not squeeze unconditionally. | """
return self._get_formatted_data(self._manager.get_cloned_from_ground_truth_cache(self), envs_idx)
def _get_formatted_data(
self, tensor: torch.Tensor, envs_idx: list[int] | None
) -> torch.Tensor | dict[str, torch.Tensor]:
# Note: This method does not clone the data tensor, ... | @@ -12,10 +21,138 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.143 | suggestion | 39 | 51 | 31 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: @gs.assert_built
def read(self, envs_idx: List[int] | None = None):
"""
Read the sensor data (with noise applied if applicable).
"""
... | You should not squeeze unconditionally. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
"""
raise NotImplementedError("Sensors must implement `get_cache_dtype()`.")
# =============================== shared methods ===============================
@gs.assert_built
def read(self, envs_idx: List[int] | None = None):
"""
Read the sensor data (with noise applied if ... | Avoid code duplication by doing:
```python
return_shapes = return_format.values() if isinstance(return_format, dict) else (return_format,)
return_values = []
for shape in return_shapes:
value = ...
return_values.append(value)
if isinstance(return_format, dict):
return dict(zip(return_format.keys()... |
@gs.assert_built
def read_ground_truth(self, envs_idx: List[int] | None = None):
"""
Read the ground truth sensor data (without noise).
"""
return self._get_formatted_data(self._manager.get_cloned_from_ground_truth_cache(self), envs_idx)
def _get_formatted_data(
sel... | @@ -12,10 +21,138 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.857 | style | 367 | 51 | 36 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: """
raise NotImplementedError("Sensors must implement `get_cache_dtype()`.")
# =============================== shared methods ===================... | Avoid code duplication by doing:
```python
return_shapes = return_format.values() if isinstance(return_format, dict) else (return_format,)
return_values = []
for shape in return_shapes:
value = ...
return_values.append(value)
if isinstance(return_format, dict):
return dict(zip(return_format.keys()... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
"""
Read the ground truth sensor data (without noise).
"""
return self._get_return_data_from_tensor(self._ground_truth_cache, envs_idx)
def _get_return_data_from_tensor(self, tensor: torch.Tensor, envs_idx: List[int] | None) -> torch.Tensor:
if envs_idx is None:
... | You should precompute your exact start-stop indices for each value instead of doing it at runtime. | from typing import TYPE_CHECKING, Any, List, Type
import numpy as np
import taichi as ti
import torch
import genesis as gs
from genesis.repr_base import RBC
if TYPE_CHECKING:
from genesis.options.sensors import SensorOptions
from genesis.utils.ring_buffer import TensorRingBuffer
from .sensor_manager imp... | @@ -12,10 +21,138 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.286 | suggestion | 98 | 50 | 131 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: """
Read the ground truth sensor data (without noise).
"""
return self._get_return_data_from_tensor(self._ground_truth_cache, envs_idx... | You should precompute your exact start-stop indices for each value instead of doing it at runtime. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
def read_ground_truth(self, envs_idx: List[int] | None = None):
"""
Read the ground truth sensor data (without noise).
"""
return self._get_return_data_from_tensor(self._ground_truth_cache, envs_idx)
def _get_return_data_from_tensor(self, tensor: torch.Tensor, envs_idx: List[int... | Start by extracting `tensor[envs_idx, self._cache_idx : self._cache_idx + self._cache_size]` once, instead of doing it repeatedly. | from typing import TYPE_CHECKING, Any, List, Type
import numpy as np
import taichi as ti
import torch
import genesis as gs
from genesis.repr_base import RBC
if TYPE_CHECKING:
from genesis.options.sensors import SensorOptions
from genesis.utils.ring_buffer import TensorRingBuffer
from .sensor_manager imp... | @@ -12,10 +21,138 @@ class Sensor(RBC):
A sensor must have a read() method that returns the sensor data.
"""
+ def __init__(self, sensor_options: "SensorOptions", sensor_idx: int, sensor_manager: "SensorManager"):
+ self._options: "SensorOptions" = sensor_options
+ self._idx: int = sensor_i... | genesis/sensors/base_sensor.py | 26 | Python | 0.571 | performance | 130 | 51 | 131 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/base_sensor.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: def read_ground_truth(self, envs_idx: List[int] | None = None):
"""
Read the ground truth sensor data (without noise).
"""
return ... | Start by extracting `tensor[envs_idx, self._cache_idx : self._cache_idx + self._cache_size]` once, instead of doing it repeatedly. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
from .base_sensor import Sensor
class SensorManager:
SENSOR_TYPES_MAP: dict[Type["SensorOptions"], Type["Sensor"]] = {}
def __init__(self, sim):
self._sim = sim
self._sensors: list["Sensor"] = []
self._sensors_metadata: dict[Type["Sensor"], dict[str, Any]] = {}
self._grou... | Define variable `sensor_type = type(sensor)` instead of using `sensor.__class__` everywhere. | if TYPE_CHECKING:
from genesis.options.sensors import SensorOptions
from .base_sensor import Sensor
class SensorManager:
SENSOR_TYPES_MAP: dict[Type["SensorOptions"], Type["Sensor"]] = {}
def __init__(self, sim):
self._sim = sim
self._sensors_by_type: dict[Type["Sensor"], list["Senso... | @@ -0,0 +1,69 @@
+from typing import TYPE_CHECKING, Any, Type
+
+import torch
+
+from genesis.utils.ring_buffer import TensorRingBuffer
+
+if TYPE_CHECKING:
+ from genesis.options.sensors import SensorOptions
+
+ from .base_sensor import Sensor
+
+
+class SensorManager:
+ SENSOR_TYPES_MAP: dict[Type["SensorOpt... | genesis/sensors/sensor_manager.py | 26 | Python | 0.571 | suggestion | 92 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/sensor_manager.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE:
from .base_sensor import Sensor
class SensorManager:
SENSOR_TYPES_MAP: dict[Type["SensorOptions"], Type["Sensor"]] = {}
def __init__(self, sim):
... | Define variable `sensor_type = type(sensor)` instead of using `sensor.__class__` everywhere. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
class SensorManager:
SENSOR_TYPES_MAP: dict[Type["SensorOptions"], Type["Sensor"]] = {}
def __init__(self, sim):
self._sim = sim
self._sensors: list["Sensor"] = []
self._sensors_metadata: dict[Type["Sensor"], dict[str, Any]] = {}
self._ground_truth_cache: dict[Type[torch.dtype]... | Just use `setdefault`:
```python
self._sensors_metadata.setdefault(sensor_type, {})
self._first_sensor_per_type.setdefault(sensor_type, sensor)
``` | from .base_sensor import Sensor
class SensorManager:
SENSOR_TYPES_MAP: dict[Type["SensorOptions"], Type["Sensor"]] = {}
def __init__(self, sim):
self._sim = sim
self._sensors_by_type: dict[Type["Sensor"], list["Sensor"]] = {}
self._sensors_metadata: dict[Type["Sensor"], dict[str, ... | @@ -0,0 +1,69 @@
+from typing import TYPE_CHECKING, Any, Type
+
+import torch
+
+from genesis.utils.ring_buffer import TensorRingBuffer
+
+if TYPE_CHECKING:
+ from genesis.options.sensors import SensorOptions
+
+ from .base_sensor import Sensor
+
+
+class SensorManager:
+ SENSOR_TYPES_MAP: dict[Type["SensorOpt... | genesis/sensors/sensor_manager.py | 26 | Python | 0.929 | suggestion | 151 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/sensor_manager.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE:
class SensorManager:
SENSOR_TYPES_MAP: dict[Type["SensorOptions"], Type["Sensor"]] = {}
def __init__(self, sim):
self._sim = sim
self._se... | Just use `setdefault`:
```python
self._sensors_metadata.setdefault(sensor_type, {})
self._first_sensor_per_type.setdefault(sensor_type, sensor)
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self._sim = sim
self._sensors: list["Sensor"] = []
self._sensors_metadata: dict[Type["Sensor"], dict[str, Any]] = {}
self._ground_truth_cache: dict[Type[torch.dtype], torch.Tensor] = {}
self._cache: dict[Type[torch.dtype], TensorRingBuffer] = {}
self._first_sensor_per_typ... | Same.
```python
cache_size_per_dtype.set_default(sensor._get_cache_dtype(), 0)
``` |
def __init__(self, sim):
self._sim = sim
self._sensors_by_type: dict[Type["Sensor"], list["Sensor"]] = {}
self._sensors_metadata: dict[Type["Sensor"], dict[str, Any]] = {}
self._ground_truth_cache: dict[Type[torch.dtype], torch.Tensor] = {}
self._cache: dict[Type[torch.dtype... | @@ -0,0 +1,69 @@
+from typing import TYPE_CHECKING, Any, Type
+
+import torch
+
+from genesis.utils.ring_buffer import TensorRingBuffer
+
+if TYPE_CHECKING:
+ from genesis.options.sensors import SensorOptions
+
+ from .base_sensor import Sensor
+
+
+class SensorManager:
+ SENSOR_TYPES_MAP: dict[Type["SensorOpt... | genesis/sensors/sensor_manager.py | 26 | Python | 0.714 | performance | 85 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/sensor_manager.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: self._sim = sim
self._sensors: list["Sensor"] = []
self._sensors_metadata: dict[Type["Sensor"], dict[str, Any]] = {}
self._ground_t... | Same.
```python
cache_size_per_dtype.set_default(sensor._get_cache_dtype(), 0)
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self._sensors_metadata: dict[Type["Sensor"], dict[str, Any]] = {}
self._ground_truth_cache: dict[Type[torch.dtype], torch.Tensor] = {}
self._cache: dict[Type[torch.dtype], TensorRingBuffer] = {}
self._first_sensor_per_type: dict[Type["Sensor"], "Sensor"] = {}
def create_sensor(self,... | Create a variable for `dtype`:
```python
dtype = sensor._get_cache_dtype()
``` | self._sim = sim
self._sensors_by_type: dict[Type["Sensor"], list["Sensor"]] = {}
self._sensors_metadata: dict[Type["Sensor"], dict[str, Any]] = {}
self._ground_truth_cache: dict[Type[torch.dtype], torch.Tensor] = {}
self._cache: dict[Type[torch.dtype], TensorRingBuffer] = {}
... | @@ -0,0 +1,69 @@
+from typing import TYPE_CHECKING, Any, Type
+
+import torch
+
+from genesis.utils.ring_buffer import TensorRingBuffer
+
+if TYPE_CHECKING:
+ from genesis.options.sensors import SensorOptions
+
+ from .base_sensor import Sensor
+
+
+class SensorManager:
+ SENSOR_TYPES_MAP: dict[Type["SensorOpt... | genesis/sensors/sensor_manager.py | 26 | Python | 0.714 | performance | 81 | 51 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/sensor_manager.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: self._sensors_metadata: dict[Type["Sensor"], dict[str, Any]] = {}
self._ground_truth_cache: dict[Type[torch.dtype], torch.Tensor] = {}
self... | Create a variable for `dtype`:
```python
dtype = sensor._get_cache_dtype()
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
if sensor.__class__ not in self._first_sensor_per_type:
self._first_sensor_per_type[sensor.__class__] = sensor
sensor.build()
if sensor._get_cache_dtype() not in cache_size_per_dtype:
cache_size_per_dtype[sensor._get_cache_dtype()] = 0
s... | `self._sensors_by_type` is not define. | def build(self):
max_cache_buf_len = 0
cache_size_per_dtype = {}
for sensor_cls, sensors in self._sensors_by_type.items():
self._sensors_metadata[sensor_cls] = {}
dtype = sensor_cls._get_cache_dtype()
cache_size_per_dtype.setdefault(dtype, 0)
c... | @@ -0,0 +1,69 @@
+from typing import TYPE_CHECKING, Any, Type
+
+import torch
+
+from genesis.utils.ring_buffer import TensorRingBuffer
+
+if TYPE_CHECKING:
+ from genesis.options.sensors import SensorOptions
+
+ from .base_sensor import Sensor
+
+
+class SensorManager:
+ SENSOR_TYPES_MAP: dict[Type["SensorOpt... | genesis/sensors/sensor_manager.py | 26 | Python | 0.357 | suggestion | 38 | 35 | 51 | false | [FEATURE] Add dedicated sensor manager. | 1,518 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/sensor_manager.py
PR Title: [FEATURE] Add dedicated sensor manager.
BEFORE CODE: if sensor.__class__ not in self._first_sensor_per_type:
self._first_sensor_per_type[sensor.__class__] = sensor
sensor.buil... | `self._sensors_by_type` is not define. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
texture = texture.astype(np.uint8)
else:
raise TypeError("Invalid type {} for texture".format(type(texture)))
# Format array by picking out correct texture channels or padding
if texture.ndim == 2:
texture = texture[:, :, np.newaxis]
if target_channel... | Just do `texture = texture[:, :, (0, 0, 0)]`.
Plus add a comment to explain what it does (as you did in the description of the PR). | texture = texture.astype(np.uint8)
else:
raise TypeError("Invalid type {} for texture".format(type(texture)))
# Format array by picking out correct texture channels or padding
if texture.ndim == 2:
texture = texture[:, :, np.newaxis]
if target_channel... | @@ -87,15 +87,23 @@ def format_texture_source(texture, target_channels="RGB"):
if texture.shape[2] == 1:
texture = np.repeat(texture, 3, axis=2)
elif texture.shape[2] == 2:
- raise ValueError("Cannot reformat 2-channel texture into RGB")
+ l = tex... | genesis/ext/pyrender/utils.py | 26 | Python | 0.5 | suggestion | 134 | 51 | 51 | false | [FEATURE] Support 2-channel (LA) textures in Rasterizer. | 1,519 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | LeonLiu4 | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/utils.py
PR Title: [FEATURE] Support 2-channel (LA) textures in Rasterizer.
BEFORE CODE: texture = texture.astype(np.uint8)
else:
raise TypeError("Invalid type {} for texture".format(type(texture)))
... | Just do `texture = texture[:, :, (0, 0, 0)]`.
Plus add a comment to explain what it does (as you did in the description of the PR). | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
texture = texture[:, :, 0]
texture = texture.squeeze()
elif target_channels == "RG":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
else:
texture = texture[:, :, (0, 1)]
elif target_channels == "GB":
... | `l` and `a` are not a good variable names. Try to be more explicit. | texture = texture[:, :, 0]
texture = texture.squeeze()
elif target_channels == "RG":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
else:
texture = texture[:, :, (0, 1)]
elif target_channels == "GB":
... | @@ -87,15 +87,23 @@ def format_texture_source(texture, target_channels="RGB"):
if texture.shape[2] == 1:
texture = np.repeat(texture, 3, axis=2)
elif texture.shape[2] == 2:
- raise ValueError("Cannot reformat 2-channel texture into RGB")
+ l = tex... | genesis/ext/pyrender/utils.py | 26 | Python | 0.357 | suggestion | 67 | 45 | 43 | false | [FEATURE] Support 2-channel (LA) textures in Rasterizer. | 1,519 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | LeonLiu4 | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/utils.py
PR Title: [FEATURE] Support 2-channel (LA) textures in Rasterizer.
BEFORE CODE: texture = texture[:, :, 0]
texture = texture.squeeze()
elif target_channels == "RG":
if texture.shape[2] =... | `l` and `a` are not a good variable names. Try to be more explicit. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
elif target_channels == "RG":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
else:
texture = texture[:, :, (0, 1)]
elif target_channels == "GB":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2... | Same `tx` us not a good variable name. | elif target_channels == "RG":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
else:
texture = texture[:, :, (0, 1)]
elif target_channels == "GB":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2... | @@ -87,15 +87,23 @@ def format_texture_source(texture, target_channels="RGB"):
if texture.shape[2] == 1:
texture = np.repeat(texture, 3, axis=2)
elif texture.shape[2] == 2:
- raise ValueError("Cannot reformat 2-channel texture into RGB")
+ l = tex... | genesis/ext/pyrender/utils.py | 26 | Python | 0.214 | suggestion | 38 | 43 | 41 | false | [FEATURE] Support 2-channel (LA) textures in Rasterizer. | 1,519 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | LeonLiu4 | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/utils.py
PR Title: [FEATURE] Support 2-channel (LA) textures in Rasterizer.
BEFORE CODE: elif target_channels == "RG":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
else:
... | Same `tx` us not a good variable name. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
else:
texture = texture[:, :, (0, 1)]
elif target_channels == "GB":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
elif texture.shape[2] > 2:
texture = texture[:, :, (1, 2)]
elif target_channels == "RG... | What about just calling it `texture` at this point ? You already exacted the values so it should be fine. | else:
texture = texture[:, :, (0, 1)]
elif target_channels == "GB":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
elif texture.shape[2] > 2:
texture = texture[:, :, (1, 2)]
elif target_channels == "RG... | @@ -87,15 +87,21 @@ def format_texture_source(texture, target_channels="RGB"):
if texture.shape[2] == 1:
texture = np.repeat(texture, 3, axis=2)
elif texture.shape[2] == 2:
- raise ValueError("Cannot reformat 2-channel texture into RGB")
+ # Conve... | genesis/ext/pyrender/utils.py | 26 | Python | 0.571 | question | 105 | 38 | 38 | false | [FEATURE] Support 2-channel (LA) textures in Rasterizer. | 1,519 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | LeonLiu4 | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/utils.py
PR Title: [FEATURE] Support 2-channel (LA) textures in Rasterizer.
BEFORE CODE: else:
texture = texture[:, :, (0, 1)]
elif target_channels == "GB":
if texture.shape[2] == 1:
... | What about just calling it `texture` at this point ? You already exacted the values so it should be fine. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
else:
texture = texture[:, :, (0, 1)]
elif target_channels == "GB":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
elif texture.sha... | `rgba[:, :, :3] = luminance` does not work? I think it should. | if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
else:
texture = texture[:, :, (0, 1)]
elif target_channels == "GB":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
elif texture.sha... | @@ -87,15 +87,21 @@ def format_texture_source(texture, target_channels="RGB"):
if texture.shape[2] == 1:
texture = np.repeat(texture, 3, axis=2)
elif texture.shape[2] == 2:
- raise ValueError("Cannot reformat 2-channel texture into RGB")
+ # Conve... | genesis/ext/pyrender/utils.py | 26 | Python | 0.429 | suggestion | 62 | 40 | 40 | false | [FEATURE] Support 2-channel (LA) textures in Rasterizer. | 1,519 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | LeonLiu4 | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/utils.py
PR Title: [FEATURE] Support 2-channel (LA) textures in Rasterizer.
BEFORE CODE: if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
else:
texture = texture[:, :... | `rgba[:, :, :3] = luminance` does not work? I think it should. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
texture = texture.squeeze()
elif target_channels == "RG":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
else:
texture = texture[:, :, (0, 1)]
elif target_channels == "GB":
if texture.shape[2] == 1:
... | A small comment as you did right above would be nice. | texture = texture.squeeze()
elif target_channels == "RG":
if texture.shape[2] == 1:
texture = np.repeat(texture, 2, axis=2)
else:
texture = texture[:, :, (0, 1)]
elif target_channels == "GB":
if texture.shape[2] == 1:
... | @@ -87,15 +87,21 @@ def format_texture_source(texture, target_channels="RGB"):
if texture.shape[2] == 1:
texture = np.repeat(texture, 3, axis=2)
elif texture.shape[2] == 2:
- raise ValueError("Cannot reformat 2-channel texture into RGB")
+ # Conve... | genesis/ext/pyrender/utils.py | 26 | Python | 0.214 | suggestion | 53 | 42 | 42 | false | [FEATURE] Support 2-channel (LA) textures in Rasterizer. | 1,519 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | LeonLiu4 | Review this pull request change.
Language: Python
File: genesis/ext/pyrender/utils.py
PR Title: [FEATURE] Support 2-channel (LA) textures in Rasterizer.
BEFORE CODE: texture = texture.squeeze()
elif target_channels == "RG":
if texture.shape[2] == 1:
texture = np.repea... | A small comment as you did right above would be nice. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
env_idx : int, optional
The index of the environment to track the camera.
camera_type : CAMERA_TYPE
The type of camera. Options are CAMERA_TYPE.RASTERIZER, CAMERA_TYPE.RAYTRACER, CAMERA_TYPE.BATCH_RENDERER.
"""
def __init__(
self,
visualizer,
idx=0,
model... | What about defaulting to None, then replace None using the same logic as we are currently using in `render`, basically BatchRenderer if applicable, if not, RayTracer, if not Rasterizer ? | env_idx : int, optional
The index of the environment to track the camera.
debug : bool, optional
Whether to use the debug camera. It enables to create cameras that can used to monitor / debug the
simulation without being part of the "sensors". Their output is rendered by the usual simple... | @@ -87,6 +92,7 @@ def __init__(
far=100.0,
transform=None,
env_idx=None,
+ camera_type=CAMERA_TYPE.RASTERIZER, | genesis/vis/camera.py | 26 | Python | 0.571 | performance | 186 | 51 | 51 | false | [FEATURE] Add debug camera for monitoring. | 1,521 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Kashu7100 | Review this pull request change.
Language: Python
File: genesis/vis/camera.py
PR Title: [FEATURE] Add debug camera for monitoring.
BEFORE CODE: env_idx : int, optional
The index of the environment to track the camera.
camera_type : CAMERA_TYPE
The type of camera. Options are CAMERA_TYPE.RAST... | What about defaulting to None, then replace None using the same logic as we are currently using in `render`, basically BatchRenderer if applicable, if not, RayTracer, if not Rasterizer ? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
env_idx : int, optional
The index of the environment to track the camera.
camera_type : CAMERA_TYPE
The type of camera. Options are CAMERA_TYPE.RASTERIZER, CAMERA_TYPE.RAYTRACER, CAMERA_TYPE.BATCH_RENDERER.
"""
def __init__(
self,
visualizer,
idx=0,
model... | What about a simple boolean flag like `debug`, `force_rasterizer` or `is_sensor` ?
Do you think there is an actual use case for mixing multiple rendering backends except rasterizer + anything else ? If so, your approve is definitely the right one. | env_idx : int, optional
The index of the environment to track the camera.
debug : bool, optional
Whether to use the debug camera. It enables to create cameras that can used to monitor / debug the
simulation without being part of the "sensors". Their output is rendered by the usual simple... | @@ -87,6 +92,7 @@ def __init__(
far=100.0,
transform=None,
env_idx=None,
+ camera_type=CAMERA_TYPE.RASTERIZER, | genesis/vis/camera.py | 26 | Python | 0.786 | bug | 250 | 51 | 51 | false | [FEATURE] Add debug camera for monitoring. | 1,521 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Kashu7100 | Review this pull request change.
Language: Python
File: genesis/vis/camera.py
PR Title: [FEATURE] Add debug camera for monitoring.
BEFORE CODE: env_idx : int, optional
The index of the environment to track the camera.
camera_type : CAMERA_TYPE
The type of camera. Options are CAMERA_TYPE.RAST... | What about a simple boolean flag like `debug`, `force_rasterizer` or `is_sensor` ?
Do you think there is an actual use case for mixing multiple rendering backends except rasterizer + anything else ? If so, your approve is definitely the right one. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
cam_idx = len(self._cameras)
camera = Camera(
self,
cam_idx,
model,
res,
pos,
lookat,
up,
fov,
aperture,
focus_dist,
GUI,
spp,
denoise,
... | Are you sure that this is a good idea? Does it pop a warning for every single simulation now or only if the user is calling `add_light` ? In the latter case, what about raising an exception instead? | self,
cam_idx,
model,
res,
pos,
lookat,
up,
fov,
aperture,
focus_dist,
GUI,
spp,
denoise,
env_idx=env_idx,
debug=debug,
)
se... | @@ -127,17 +127,35 @@ def destroy(self):
self.viewer_lock = None
self._renderer = None
- def add_camera(self, res, pos, lookat, up, model, fov, aperture, focus_dist, GUI, spp, denoise, env_idx):
+ def add_camera(
+ self, res, pos, lookat, up, model, fov, aperture, focus_dist, GUI, spp, ... | genesis/vis/visualizer.py | 26 | Python | 0.714 | question | 198 | 51 | 51 | false | [FEATURE] Add debug camera for monitoring. | 1,521 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Kashu7100 | Review this pull request change.
Language: Python
File: genesis/vis/visualizer.py
PR Title: [FEATURE] Add debug camera for monitoring.
BEFORE CODE: cam_idx = len(self._cameras)
camera = Camera(
self,
cam_idx,
model,
res,
pos,
lo... | Are you sure that this is a good idea? Does it pop a warning for every single simulation now or only if the user is calling `add_light` ? In the latter case, what about raising an exception instead? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
model : str
Specifies the camera model. Options are 'pinhole' or 'thinlens'.
res : tuple of int, shape (2,)
The resolution of the camera, specified as a tuple (width, height).
pos : tuple of float, shape (3,)
The position of the camera in the scene, specified ... | Can you clarify the documentation? At this point I have no clue what it is supposed to do.
You should explain that it enables to create cameras that can used to monitor / debug the simulation without being part of the that are not officially considered as part of the "sensors". Their output is rendered by the usual ... | ----------
model : str
Specifies the camera model. Options are 'pinhole' or 'thinlens'.
res : tuple of int, shape (2,)
The resolution of the camera, specified as a tuple (width, height).
pos : tuple of float, shape (3,)
The position of the camera in th... | @@ -566,14 +567,16 @@ def add_camera(
Whether to denoise the camera's rendered image. Only available when using the RayTracer renderer. Defaults
to True. If OptiX denoiser is not available in your platform, consider enabling the OIDN denoiser option
when building the RayTracer.
+ ... | genesis/engine/scene.py | 26 | Python | 0.786 | bug | 574 | 51 | 51 | false | [FEATURE] Add debug camera for monitoring. | 1,521 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Kashu7100 | Review this pull request change.
Language: Python
File: genesis/engine/scene.py
PR Title: [FEATURE] Add debug camera for monitoring.
BEFORE CODE: model : str
Specifies the camera model. Options are 'pinhole' or 'thinlens'.
res : tuple of int, shape (2,)
The resolution of the ... | Can you clarify the documentation? At this point I have no clue what it is supposed to do.
You should explain that it enables to create cameras that can used to monitor / debug the simulation without being part of the that are not officially considered as part of the "sensors". Their output is rendered by the usual ... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
The rendered RGB image(s).
depth_arr : np.ndarray
The rendered depth image(s).
seg_arr : np.ndarray
The rendered segmentation mask(s).
normal_arr : np.ndarray
The rendered surface normal(s).
"""
rgb_arr, depth_arr, seg_arr, seg_colo... | Remove this message, or using debug logging. | depth : bool, optional
Whether to render depth image(s).
segmentation : bool, optional
Whether to render the segmentation mask(s).
colorize_seg : bool, optional
If True, the segmentation mask will be colorized.
normal : bool, optional
Wheth... | @@ -388,6 +395,7 @@ def render(
self, False, depth, segmentation, normal=normal
)
else:
+ print("Camera is debug camera") | genesis/vis/camera.py | 26 | Python | 0.286 | bug | 44 | 51 | 51 | false | [FEATURE] Add debug camera for monitoring. | 1,521 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Kashu7100 | Review this pull request change.
Language: Python
File: genesis/vis/camera.py
PR Title: [FEATURE] Add debug camera for monitoring.
BEFORE CODE: The rendered RGB image(s).
depth_arr : np.ndarray
The rendered depth image(s).
seg_arr : np.ndarray
The rendered segment... | Remove this message, or using debug logging. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
def get_transform(self, env_idx=None):
"""
The current transform matrix of the camera.
"""
env_idx = () if env_idx is None else env_idx
transform = self._multi_env_transform_tensor[env_idx]
if self._batch_renderer is None:
transform = transform.clone()
... | I would rather keep this inside `_batch_render` rather than make it an official property at camera level. This kind of circular dependency is not great.
Besides, I don't think caching the result is relevant for something as simple as finding an index in a list. | if self._batch_renderer is None:
transform = transform.clone()
transform[..., :3, 3] += self._envs_offset[env_idx]
return transform
def _repr_brief(self):
return f"{self._repr_type()}: idx: {self._idx}, pos: {self.pos}, lookat: {self.lookat}"
@property
def i... | @@ -763,9 +771,18 @@ def is_built(self):
@property
def idx(self):
- """The integer index of the camera."""
+ """The global integer index of the camera."""
return self._idx
+ @property
+ @lru_cache(maxsize=1) | genesis/vis/camera.py | 26 | Python | 0.571 | performance | 264 | 51 | 51 | false | [FEATURE] Add debug camera for monitoring. | 1,521 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Kashu7100 | Review this pull request change.
Language: Python
File: genesis/vis/camera.py
PR Title: [FEATURE] Add debug camera for monitoring.
BEFORE CODE: def get_transform(self, env_idx=None):
"""
The current transform matrix of the camera.
"""
env_idx = () if env_idx is None else env_idx
... | I would rather keep this inside `_batch_render` rather than make it an official property at camera level. This kind of circular dependency is not great.
Besides, I don't think caching the result is relevant for something as simple as finding an index in a list. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
The rendered RGB image(s).
depth_arr : np.ndarray
The rendered depth image(s).
seg_arr : np.ndarray
The rendered segmentation mask(s).
normal_arr : np.ndarray
The rendered surface normal(s).
"""
rgb_arr, depth_arr, seg_arr, seg_colo... | Is the print message correct?
How about
```
if self._debug:
print(...)
``` | depth : bool, optional
Whether to render depth image(s).
segmentation : bool, optional
Whether to render the segmentation mask(s).
colorize_seg : bool, optional
If True, the segmentation mask will be colorized.
normal : bool, optional
Wheth... | @@ -388,6 +395,7 @@ def render(
self, False, depth, segmentation, normal=normal
)
else:
+ print("Camera is debug camera") | genesis/vis/camera.py | 26 | Python | 0.714 | bug | 83 | 51 | 51 | false | [FEATURE] Add debug camera for monitoring. | 1,521 | Genesis-Embodied-AI/Genesis | 26,702 | Python | yuhongyi | Kashu7100 | Review this pull request change.
Language: Python
File: genesis/vis/camera.py
PR Title: [FEATURE] Add debug camera for monitoring.
BEFORE CODE: The rendered RGB image(s).
depth_arr : np.ndarray
The rendered depth image(s).
seg_arr : np.ndarray
The rendered segment... | Is the print message correct?
How about
```
if self._debug:
print(...)
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self._raytracer.destroy()
self._raytracer = None
if self._context is not None:
self._context.destroy()
del self._context
self._context = None
self.viewer_lock = None
self._renderer = None
def add_camera(self, res, pos, lookat, up, ... | Why are you actively ignoring env_idx. It seems weird. | self._raytracer.destroy()
self._raytracer = None
if self._context is not None:
self._context.destroy()
del self._context
self._context = None
self.viewer_lock = None
self._renderer = None
def add_camera(self, res, pos, lookat, up, ... | @@ -127,17 +127,33 @@ def destroy(self):
self.viewer_lock = None
self._renderer = None
- def add_camera(self, res, pos, lookat, up, model, fov, aperture, focus_dist, GUI, spp, denoise, env_idx):
+ def add_camera(self, res, pos, lookat, up, model, fov, aperture, focus_dist, GUI, spp, denoise, e... | genesis/vis/visualizer.py | 26 | Python | 0.214 | suggestion | 54 | 51 | 51 | false | [FEATURE] Add debug camera for monitoring. | 1,521 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Kashu7100 | Review this pull request change.
Language: Python
File: genesis/vis/visualizer.py
PR Title: [FEATURE] Add debug camera for monitoring.
BEFORE CODE: self._raytracer.destroy()
self._raytracer = None
if self._context is not None:
self._context.destroy()
del self.... | Why are you actively ignoring env_idx. It seems weird. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
pos : tuple of float, shape (3,)
The position of the camera in the scene, specified as (x, y, z).
lookat : tuple of float, shape (3,)
The point in the scene that the camera is looking at, specified as (x, y, z).
up : tuple of float, shape (3,)
The up vector of... | Do not skip line unless it would exceed 120 linewidth otherwise. | The resolution of the camera, specified as a tuple (width, height).
pos : tuple of float, shape (3,)
The position of the camera in the scene, specified as (x, y, z).
lookat : tuple of float, shape (3,)
The point in the scene that the camera is looking at, specified as... | @@ -566,14 +567,20 @@ def add_camera(
Whether to denoise the camera's rendered image. Only available when using the RayTracer renderer. Defaults
to True. If OptiX denoiser is not available in your platform, consider enabling the OIDN denoiser option
when building the RayTracer.
+ ... | genesis/engine/scene.py | 26 | Python | 0.214 | suggestion | 65 | 51 | 51 | false | [FEATURE] Add debug camera for monitoring. | 1,521 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Kashu7100 | Review this pull request change.
Language: Python
File: genesis/engine/scene.py
PR Title: [FEATURE] Add debug camera for monitoring.
BEFORE CODE: pos : tuple of float, shape (3,)
The position of the camera in the scene, specified as (x, y, z).
lookat : tuple of float, shape (3,)
... | Do not skip line unless it would exceed 120 linewidth otherwise. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
self.viewer_lock = None
self._renderer = None
def add_camera(self, res, pos, lookat, up, model, fov, aperture, focus_dist, GUI, spp, denoise, env_idx, debug):
cam_idx = len(self._cameras) if not debug else len(self._debug_cameras)
camera = Camera(
self,
cam_i... | I'm not a great fan of managing cameras and "debug cameras" separately. They are inherently the same things. Just the debug camera are not considered real sensors but only used to monitor the progress of the agent / debugging. Having `camera` getting hiding debug camera by default, and keeping them accessible via `_cam... | self._renderer = None
def add_camera(self, res, pos, lookat, up, model, fov, aperture, focus_dist, GUI, spp, denoise, env_idx, debug):
cam_idx = len([camera for camera in self._cameras if camera.debug == debug])
camera = Camera(
self,
cam_idx,
model,
... | @@ -127,17 +128,36 @@ def destroy(self):
self.viewer_lock = None
self._renderer = None
- def add_camera(self, res, pos, lookat, up, model, fov, aperture, focus_dist, GUI, spp, denoise, env_idx):
- cam_idx = len(self._cameras)
+ def add_camera(self, res, pos, lookat, up, model, fov, aper... | genesis/vis/visualizer.py | 26 | Python | 0.786 | bug | 368 | 51 | 51 | false | [FEATURE] Add debug camera for monitoring. | 1,521 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Kashu7100 | Review this pull request change.
Language: Python
File: genesis/vis/visualizer.py
PR Title: [FEATURE] Add debug camera for monitoring.
BEFORE CODE: self.viewer_lock = None
self._renderer = None
def add_camera(self, res, pos, lookat, up, model, fov, aperture, focus_dist, GUI, spp, denoise, env_i... | I'm not a great fan of managing cameras and "debug cameras" separately. They are inherently the same things. Just the debug camera are not considered real sensors but only used to monitor the progress of the agent / debugging. Having `camera` getting hiding debug camera by default, and keeping them accessible via `_cam... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
B = z.shape[:-1]
if out is None:
out_ = np.empty((*B, 3, 3), dtype=z.dtype)
else:
assert out.shape == (*B, 3, 3)
out_ = out
z_norm = np.sqrt(np.sum(np.square(z.reshape((-1, 3))), -1)).reshape(B)
out_[..., 2] = z
for i in np.ndindex(B):
z_norm_i = z_norm[i]
... | Remove this change. You just need to update the special case:
```python
else:
x[0] = z[2]
x[1] = 0.0
x[2] = -z[0]
``` | B = z.shape[:-1]
if out is None:
out_ = np.empty((*B, 3, 3), dtype=z.dtype)
else:
assert out.shape == (*B, 3, 3)
out_ = out
z_norm = np.sqrt(np.sum(np.square(z.reshape((-1, 3))), -1)).reshape(B)
out_[..., 2] = z
for i in np.ndindex(B):
z_norm_i = z_norm[i]
... | @@ -1162,18 +1162,20 @@ def _np_z_up_to_R(z, up=None, out=None):
z[:] = 0.0, 1.0, 0.0
if up is not None:
- x[:] = np.cross(up[i], z)
+ a = up[i]
else:
- x[0] = z[1]
- x[1] = -z[0]
- x[2] = 0.0
+ a = np.array([0.0, 1.0, 0.... | genesis/utils/geom.py | 26 | Python | 0.786 | refactor | 167 | 51 | 51 | false | [BUG FIX] Correct -Z orientation & vertex color | 1,540 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | ceasor-mao | Review this pull request change.
Language: Python
File: genesis/utils/geom.py
PR Title: [BUG FIX] Correct -Z orientation & vertex color
BEFORE CODE: B = z.shape[:-1]
if out is None:
out_ = np.empty((*B, 3, 3), dtype=z.dtype)
else:
assert out.shape == (*B, 3, 3)
out_ = out
z... | Remove this change. You just need to update the special case:
```python
else:
x[0] = z[2]
x[1] = 0.0
x[2] = -z[0]
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
z_norm = np.sqrt(np.sum(np.square(z.reshape((-1, 3))), -1)).reshape(B)
out_[..., 2] = z
for i in np.ndindex(B):
z_norm_i = z_norm[i]
R = out_[i]
x, y, z = R.T
if z_norm_i > gs.EPS:
z /= z_norm_i
else:
z[:] = 0.0, 1.0, 0.0
if up is n... | You should rather adapt the special case to detect whether `x_norm < gs.EPS` before setting it. Basically, you just need to check `z[1] > math.sqrt(1.0 - gs.EPS ** 2)`, then use `axis = (1.0, 0.0, 0.0)` otherwise (virtually, without actually cross product). |
z_norm = np.sqrt(np.sum(np.square(z.reshape((-1, 3))), -1)).reshape(B)
out_[..., 2] = z
for i in np.ndindex(B):
z_norm_i = z_norm[i]
R = out_[i]
x, y, z = R.T
if z_norm_i > gs.EPS:
z /= z_norm_i
else:
if up is None or abs(up[i][1]) < 0.5:
... | @@ -1162,18 +1162,20 @@ def _np_z_up_to_R(z, up=None, out=None):
z[:] = 0.0, 1.0, 0.0
if up is not None:
- x[:] = np.cross(up[i], z)
+ a = up[i]
else:
- x[0] = z[1]
- x[1] = -z[0]
- x[2] = 0.0
+ a = np.array([0.0, 1.0, 0.... | genesis/utils/geom.py | 26 | Python | 0.643 | suggestion | 257 | 51 | 51 | false | [BUG FIX] Correct -Z orientation & vertex color | 1,540 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | ceasor-mao | Review this pull request change.
Language: Python
File: genesis/utils/geom.py
PR Title: [BUG FIX] Correct -Z orientation & vertex color
BEFORE CODE:
z_norm = np.sqrt(np.sum(np.square(z.reshape((-1, 3))), -1)).reshape(B)
out_[..., 2] = z
for i in np.ndindex(B):
z_norm_i = z_norm[i]
R =... | You should rather adapt the special case to detect whether `x_norm < gs.EPS` before setting it. Basically, you just need to check `z[1] > math.sqrt(1.0 - gs.EPS ** 2)`, then use `axis = (1.0, 0.0, 0.0)` otherwise (virtually, without actually cross product). | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
continue
self.rigid_state_dof[i_b, i_d].impulse = 0.0
def compute_preconditioner(self):
if self.fem_solver.is_active():
self.compute_fem_preconditioner()
@ti.kernel
def compute_fem_preconditioner(self):
for i_b, i_v in ti.ndrange(self.fem_solver._B, ... | use `I_v` or `i_vs` to indicate that they are several. | continue
self.rigid_state_dof[i_b, i_d].impulse = 0.0
def compute_preconditioner(self):
if self.fem_solver.is_active():
self.compute_fem_preconditioner()
@ti.kernel
def compute_fem_preconditioner(self):
for i_b, i_v in ti.ndrange(self.fem_solver._B, ... | @@ -729,18 +729,16 @@ def compute_rigid_pcg_matrix_vector_product(self):
)
@ti.func
- def compute_elastic_products(self, i_b, i_e, B, s, i_v0, i_v1, i_v2, i_v3, src):
+ def compute_elastic_products(self, i_b, i_e, S, i_v, src): | genesis/engine/couplers/sap_coupler.py | 26 | Python | 0.429 | suggestion | 54 | 51 | 51 | false | [FEATURE] Add Fem fixed constraint for implicit solver | 1,562 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Libero0809 | Review this pull request change.
Language: Python
File: genesis/engine/couplers/sap_coupler.py
PR Title: [FEATURE] Add Fem fixed constraint for implicit solver
BEFORE CODE: continue
self.rigid_state_dof[i_b, i_d].impulse = 0.0
def compute_preconditioner(self):
if self.fem_so... | use `I_v` or `i_vs` to indicate that they are several. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
damping_alpha_dt = self._damping_alpha * self._substep_dt
damping_alpha_factor = damping_alpha_dt + 1.0
damping_beta_over_dt = self._damping_beta / self._substep_dt
damping_beta_factor = damping_beta_over_dt + 1.0
# inertia
for i_b, i_v in ti.ndrange(self._B, self.n_verti... | same. `i_v` is not ideal for something that is a vector. | H_St_x_diff[i * 3 : i * 3 + 3] += (
self.elements_el_hessian[i_b, i, j, i_e] @ St_x_diff[j * 3 : j * 3 + 3]
)
self.elements_el_energy[i_b, i_e].energy += 0.5 * damping_beta_over_dt * St_x_diff.dot(H_St_x_diff)
@ti.kernel
def accum... | @@ -682,18 +677,22 @@ def accumulate_vertex_force_preconditioner(self, f: ti.i32):
V = self.elements_i[i_e].V
B = self.elements_i[i_e].B
gradient = self.elements_el_energy[i_b, i_e].gradient
- force = -V * gradient @ B.transpose()
i_v = self.elements_i[i_e]... | genesis/engine/solvers/fem_solver.py | 26 | Python | 0.357 | suggestion | 56 | 51 | 51 | false | [FEATURE] Add Fem fixed constraint for implicit solver | 1,562 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Libero0809 | Review this pull request change.
Language: Python
File: genesis/engine/solvers/fem_solver.py
PR Title: [FEATURE] Add Fem fixed constraint for implicit solver
BEFORE CODE: damping_alpha_dt = self._damping_alpha * self._substep_dt
damping_alpha_factor = damping_alpha_dt + 1.0
damping_beta_over... | same. `i_v` is not ideal for something that is a vector. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
def compute_Ap(self):
damping_alpha_dt = self._damping_alpha * self._substep_dt
damping_alpha_factor = damping_alpha_dt + 1.0
damping_beta_over_dt = self._damping_beta / self._substep_dt
damping_beta_factor = damping_beta_over_dt + 1.0
for i_b, i_v in ti.ndrange(self._B, self... | You have multiple duplication of this pattern. Could this be factored in some private ti.func? | def compute_Ap(self):
damping_alpha_dt = self._damping_alpha * self._substep_dt
damping_alpha_factor = damping_alpha_dt + 1.0
damping_beta_over_dt = self._damping_beta / self._substep_dt
damping_beta_factor = damping_beta_over_dt + 1.0
for i_b, i_v in ti.ndrange(self._B, self... | @@ -758,40 +757,33 @@ def compute_Ap(self):
continue
V = self.elements_i[i_e].V
B = self.elements_i[i_e].B
- s = -B[0, :] - B[1, :] - B[2, :] # s is the negative sum of B rows
+ i_v = self.elements_i[i_e].el2v
+ S = ti.Matrix.zero(gs.ti_float,... | genesis/engine/solvers/fem_solver.py | 26 | Python | 0.286 | question | 94 | 51 | 51 | false | [FEATURE] Add Fem fixed constraint for implicit solver | 1,562 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Libero0809 | Review this pull request change.
Language: Python
File: genesis/engine/solvers/fem_solver.py
PR Title: [FEATURE] Add Fem fixed constraint for implicit solver
BEFORE CODE: def compute_Ap(self):
damping_alpha_dt = self._damping_alpha * self._substep_dt
damping_alpha_factor = damping_alpha_dt + 1.0... | You have multiple duplication of this pattern. Could this be factored in some private ti.func? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.