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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
floor_height : float, optional
Height of the floor in meters. If none, it will inherit from `SimOptions`. Defaults to None.
use_implicit_solver : bool, optional
Whether to use the implicit solver. Defaults to False.
Implicit solver is a more stable solver for FEM. It can be used with a l... | Skile space before note. Ideally notes should be moved before parameters. | Parameters
----------
dt : float, optional
Time duration for each simulation step in seconds. If none, it will inherit from `SimOptions`. Defaults to None.
gravity : tuple, optional
Gravity force in N/kg. If none, it will inherit from `SimOptions`. Defaults to None.
damping : float, ... | @@ -612,7 +612,8 @@ class FEMOptions(Options):
Rayleigh Damping factor for the implicit solver. Defaults to 0.5. Only used when `use_implicit_solver` is True.
damping_beta : float, optional
Rayleigh Damping factor for the implicit solver. Defaults to 5e-4. Only used when `use_implicit_solver` is ... | genesis/options/solvers.py | 26 | Python | 0.286 | suggestion | 73 | 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/options/solvers.py
PR Title: [FEATURE] Add Fem fixed constraint for implicit solver
BEFORE CODE: floor_height : float, optional
Height of the floor in meters. If none, it will inherit from `SimOptions`. Defaults to None.
use_implicit_so... | Skile space before note. Ideally notes should be moved before parameters. | 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. |
np.array([0.0, 0.0, 0.2], dtype=np.float32),
atol=0.2,
err_msg=f"Sphere center {center} moves too far from [0.0, 0.0, 0.2].",
)
link_verts = link.get_verts()
center = link_verts.mean(axis=0)
min_pos_z = link_verts[..., 2].min()
assert_allclose(
min_pos_z,
-1.... | Only break line when reaching 120 chars line width limit. | np.array([0.0, 0.0, 0.2], dtype=np.float32),
atol=0.2,
err_msg=f"Sphere center {center} moves too far from [0.0, 0.0, 0.2].",
)
link_verts = link.get_verts()
center = link_verts.mean(axis=0)
min_pos_z = link_verts[..., 2].min()
assert_allclose(
min_pos_z,
-1.... | @@ -582,3 +582,139 @@ def test_fem_articulated(fem_material_linear_corotated_soft, show_viewer):
atol=0.2,
err_msg=f"Link center {center} moves too far from [-0.5, -0.5, 0.04].",
)
+
+
+def test_implicit_hard_vertex_constraint(fem_material_linear_corotated, show_viewer):
+ """
+ Test if a b... | tests/test_fem.py | 26 | Python | 0.214 | suggestion | 57 | 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: tests/test_fem.py
PR Title: [FEATURE] Add Fem fixed constraint for implicit solver
BEFORE CODE: np.array([0.0, 0.0, 0.2], dtype=np.float32),
atol=0.2,
err_msg=f"Sphere center {center} moves too far from [0.0, 0.0, 0.2].",
)
lin... | Only break line when reaching 120 chars line width limit. | 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 show_viewer:
sphere = scene.draw_debug_spheres(poss=initial_target_poss, radius=0.02, color=(1, 0, 1, 0.8))
cube.set_vertex_constraints(verts_idx=verts_idx, target_poss=initial_target_poss)
for _ in range(100):
scene.step()
positions = cube.get_state().pos[0][verts_idx]
assert... | Using `gs.EPS` as tolerance. We do not have stricter requirement. |
if show_viewer:
sphere = scene.draw_debug_spheres(poss=initial_target_poss, radius=0.02, color=(1, 0, 1, 0.8))
cube.set_vertex_constraints(verts_idx=verts_idx, target_poss=initial_target_poss)
for _ in range(100):
scene.step()
positions = cube.get_state().pos[0][verts_idx]
assert... | @@ -582,3 +582,139 @@ def test_fem_articulated(fem_material_linear_corotated_soft, show_viewer):
atol=0.2,
err_msg=f"Link center {center} moves too far from [-0.5, -0.5, 0.04].",
)
+
+
+def test_implicit_hard_vertex_constraint(fem_material_linear_corotated, show_viewer):
+ """
+ Test if a b... | tests/test_fem.py | 26 | Python | 0.357 | suggestion | 65 | 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: tests/test_fem.py
PR Title: [FEATURE] Add Fem fixed constraint for implicit solver
BEFORE CODE:
if show_viewer:
sphere = scene.draw_debug_spheres(poss=initial_target_poss, radius=0.02, color=(1, 0, 1, 0.8))
cube.set_vertex_constraints(verts_i... | Using `gs.EPS` as tolerance. We do not have stricter requirement. | 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=f"{asset_path}/meshes/cube8.obj",
scale=0.1,
pos=np.array([0.0, 0.0, 0.6], dtype=np.float32),
),
material=fem_material_linear_corotated,
)
verts_idx = [0]
initial_target_poss = cube.init_positions[verts_idx]
scene.build()
if show_viewer:
... | `gs.tensor` instead of `torch.tensor`? | file=f"{asset_path}/meshes/cube8.obj",
scale=0.1,
pos=np.array([0.0, 0.0, 0.6], dtype=np.float32),
),
material=fem_material_linear_corotated,
)
verts_idx = [0]
initial_target_poss = cube.init_positions[verts_idx]
scene.build()
if show_viewer:
... | @@ -582,3 +582,139 @@ def test_fem_articulated(fem_material_linear_corotated_soft, show_viewer):
atol=0.2,
err_msg=f"Link center {center} moves too far from [-0.5, -0.5, 0.04].",
)
+
+
+def test_implicit_hard_vertex_constraint(fem_material_linear_corotated, show_viewer):
+ """
+ Test if a b... | tests/test_fem.py | 26 | Python | 0.286 | question | 38 | 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: tests/test_fem.py
PR Title: [FEATURE] Add Fem fixed constraint for implicit solver
BEFORE CODE: file=f"{asset_path}/meshes/cube8.obj",
scale=0.1,
pos=np.array([0.0, 0.0, 0.6], dtype=np.float32),
),
material=f... | `gs.tensor` instead of `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. |
pos=(0.0, 0.0, 0.1),
radius=0.1,
),
material=fem_material_linear_corotated,
)
scene.build()
# return
if show_viewer:
sphere_debug = scene.draw_debug_spheres(poss=initial_target_poss, radius=0.02, color=(1, 0, 1, 0.8))
cube.set_vertex_constraints(vert... | Nothing to check of `x` and `y`? and/or the final velocity? | asset_path = get_hf_dataset(pattern="meshes/cube8.obj")
cube = scene.add_entity(
morph=gs.morphs.Mesh(
file=f"{asset_path}/meshes/cube8.obj",
scale=0.1,
pos=np.array([0.0, 0.0, 0.35], dtype=np.float32),
),
material=fem_material_linear_corotated,
)
... | @@ -582,3 +582,139 @@ def test_fem_articulated(fem_material_linear_corotated_soft, show_viewer):
atol=0.2,
err_msg=f"Link center {center} moves too far from [-0.5, -0.5, 0.04].",
)
+
+
+def test_implicit_hard_vertex_constraint(fem_material_linear_corotated, show_viewer):
+ """
+ Test if a b... | tests/test_fem.py | 26 | Python | 0.357 | question | 59 | 27 | 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: tests/test_fem.py
PR Title: [FEATURE] Add Fem fixed constraint for implicit solver
BEFORE CODE: pos=(0.0, 0.0, 0.1),
radius=0.1,
),
material=fem_material_linear_corotated,
)
scene.build()
# return
if sho... | Nothing to check of `x` and `y`? and/or the final velocity? | 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. |
new_target_poss = initial_target_poss + gs.tensor(
[[0.1, 0.1, 0.1]],
)
cube.update_constraint_targets(verts_idx=verts_idx, target_poss=new_target_poss)
if show_viewer:
scene.clear_debug_object(sphere)
sphere = scene.draw_debug_spheres(poss=new_target_poss, radius=0.02, color=(1,... | Nothing to check for `x` and `y`? and/or the final velocity? | new_target_poss = initial_target_poss + gs.tensor(
[[0.1, 0.1, 0.1]],
)
cube.update_constraint_targets(verts_idx=verts_idx, target_poss=new_target_poss)
if show_viewer:
scene.clear_debug_object(sphere)
sphere = scene.draw_debug_spheres(poss=new_target_poss, radius=0.02, color=(1,... | @@ -582,3 +582,139 @@ def test_fem_articulated(fem_material_linear_corotated_soft, show_viewer):
atol=0.2,
err_msg=f"Link center {center} moves too far from [-0.5, -0.5, 0.04].",
)
+
+
+def test_implicit_hard_vertex_constraint(fem_material_linear_corotated, show_viewer):
+ """
+ Test if a b... | tests/test_fem.py | 26 | Python | 0.357 | question | 60 | 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: tests/test_fem.py
PR Title: [FEATURE] Add Fem fixed constraint for implicit solver
BEFORE CODE: new_target_poss = initial_target_poss + gs.tensor(
[[0.1, 0.1, 0.1]],
)
cube.update_constraint_targets(verts_idx=verts_idx, target_poss=new_targ... | Nothing to check for `x` and `y`? and/or the final velocity? | 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
import numpy as np
import numpy.typing as npt
import gstaichi as ti
import torch
import genesis as gs
import genesis.utils.geom as gu
import genesis.utils.array_class as array_class
import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
from .... import maybe_pure
fro... | I would prefer using `gs.maybe_pure` everywhere if you are fine with it. | from typing import TYPE_CHECKING
import numpy as np
import numpy.typing as npt
import gstaichi as ti
import torch
import genesis as gs
import genesis.utils.geom as gu
import genesis.utils.array_class as array_class
import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
from genesis.engine.solvers.rig... | @@ -9,6 +9,7 @@
import genesis.utils.geom as gu
import genesis.utils.array_class as array_class
import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
+from .... import maybe_pure | genesis/engine/solvers/rigid/constraint_solver_decomp.py | 12 | Python | 0.429 | suggestion | 72 | 37 | 37 | false | [FEATURE] Add initial experimental support of gstaichi fast cache feature. | 1,631 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | hughperkins | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_solver_decomp.py
PR Title: [FEATURE] Add initial experimental support of gstaichi fast cache feature.
BEFORE CODE: from typing import TYPE_CHECKING
import numpy as np
import numpy.typing as npt
import gstaichi as ti
imp... | I would prefer using `gs.maybe_pure` everywhere if you are fine with it. | 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 importlib
import io
import os
import sys
import site
import atexit
import logging as _logging
import traceback
from platform import system
from contextlib import redirect_stdout
from typing import Callable, TypeVar, ParamSpec
# Import gstaichi while collecting its output without printing directly
_ti_outputs = ... | As discussed in DM, I will move the reloading logic in unit test, because it is too fragile and only relevant in this context. | import io
import os
import sys
import site
import atexit
import logging as _logging
import traceback
from platform import system
from contextlib import redirect_stdout
from typing import Callable, TypeVar, ParamSpec
# Import gstaichi while collecting its output without printing directly
_ti_outputs = io.StringIO()
os... | genesis/__init__.py | 1 | Python | 0.429 | nitpick | 126 | 26 | 26 | false | [FEATURE] Add initial experimental support of gstaichi fast cache feature. | 1,631 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | hughperkins | Review this pull request change.
Language: Python
File: genesis/__init__.py
PR Title: [FEATURE] Add initial experimental support of gstaichi fast cache feature.
BEFORE CODE: import importlib
import io
import os
import sys
import site
import atexit
import logging as _logging
import traceback
from platform import sys... | As discussed in DM, I will move the reloading logic in unit test, because it is too fragile and only relevant in this context. | 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. | |
ti.loop_config(serialize=static_rigid_sim_config.para_level < gs.PARA_LEVEL.ALL)
for i_l_ in range(rigid_global_info.n_awake_links[i_b]):
i_l = rigid_global_info.awake_links[i_l_, i_b]
links_state.root_COM[i_l, i_b].fill(0.0)
links_state.mass_sum[i_l, i_b] = 0.0
... | I don't think storing `link_COM` is useful because it is basically `i_pos`. Having multiple aliases for the same quantity is a bad practice and not memory friendly. | n_links = links_info.root_idx.shape[0]
if ti.static(static_rigid_sim_config.use_hibernation):
ti.loop_config(serialize=static_rigid_sim_config.para_level < gs.PARA_LEVEL.ALL)
for i_l_ in range(rigid_global_info.n_awake_links[i_b]):
i_l = rigid_global_info.awake_links[i_l_, i_b]
... | @@ -4253,7 +4272,8 @@ def func_COM_links(
i_r = links_info.root_idx[I_l]
links_state.mass_sum[i_r, i_b] += mass
- links_state.COM[i_r, i_b] += mass * links_state.i_pos[i_l, i_b]
+ links_state.link_COM[i_l, i_b] = links_state.i_pos[i_l, i_b] | genesis/engine/solvers/rigid/rigid_solver_decomp.py | 26 | Python | 0.714 | suggestion | 164 | 51 | 51 | false | [BUG FIX, CHANGING] Fix primitive and mesh's COM. Add 'ref' optional argument to 'get_links_pos'. | 1,638 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/rigid_solver_decomp.py
PR Title: [BUG FIX, CHANGING] Fix primitive and mesh's COM. Add 'ref' optional argument to 'get_links_pos'.
BEFORE CODE: ti.loop_config(serialize=static_rigid_sim_config.para_level < gs.PARA_LEVEL.ALL... | I don't think storing `link_COM` is useful because it is basically `i_pos`. Having multiple aliases for the same quantity is a bad practice and not memory friendly. | 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=(0.4, 0.0, -0.54),
fixed=True,
),
vis_mode="collision",
)
asset_path = get_hf_dataset(pattern="spoon.glb")
obj = scene.add_entity(
gs.morphs.Mesh(
file=f"{asset_path}/spoon.glb",
pos=(0.3, 0, 0.015),
quat=(0.707, 0.707, ... | Why did you have to change the tolerance so much? It looks wrong. | ),
rigid_options=gs.options.RigidOptions(
use_gjk_collision=gjk_collision,
),
show_viewer=show_viewer,
show_FPS=False,
)
asset_path = get_hf_dataset(pattern="work_table.glb")
table = scene.add_entity(
gs.morphs.Mesh(
file=f"{asset_path}... | @@ -1686,8 +1691,9 @@ def test_mesh_repair(convexify, show_viewer, gjk_collision):
assert all(geom.metadata["decomposed"] for geom in obj.geoms)
# MPR collision detection is significantly less reliable than SDF in terms of penetration depth estimation.
- tol_pos = 0.05 if convexify else 1e-6
- tol... | tests/test_rigid_physics.py | 26 | Python | 0.286 | bug | 65 | 51 | 51 | false | [BUG FIX, CHANGING] Fix primitive and mesh's COM. Add 'ref' optional argument to 'get_links_pos'. | 1,638 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [BUG FIX, CHANGING] Fix primitive and mesh's COM. Add 'ref' optional argument to 'get_links_pos'.
BEFORE CODE: pos=(0.4, 0.0, -0.54),
fixed=True,
),
vis_mode="collision",
)
a... | Why did you have to change the tolerance so much? It looks wrong. | 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_entities_state(solver):
shape = solver._batch_shape(solver.n_entities_)
kwargs = {
"hibernated": V(dtype=gs.ti_int, shape=shape),
}
if use_ndarray:
return StructEntitiesState(**kwargs)
else:
@ti.data_oriented
class ClassEntitiesState:
def __init... | I can see why you're going to ask me to make it possible to set `FIELD_METADATA_CACHE_VALUE` conceptually on an entire dataclass at some point :) |
def get_entities_state(solver):
shape = solver._batch_shape(solver.n_entities_)
kwargs = {
"hibernated": V(dtype=gs.ti_int, shape=shape),
}
if use_ndarray:
return StructEntitiesState(**kwargs)
else:
@ti.data_oriented
class ClassEntitiesState:
def __init... | @@ -2004,6 +2005,70 @@ def __init__(self):
return ClassEntitiesState()
+# =========================================== StaticRigidSimConfig ===========================================
+
+
+@dataclasses.dataclass
+class StaticRigidSimDummy:
+ para_level: int = dataclasses.field(metadata={FIELD_METADATA_CA... | genesis/utils/array_class.py | 26 | Python | 0.5 | performance | 145 | 51 | 51 | false | [FEATURE] Use FIELD_METADATA_CACHE_VALUE for static config | 1,639 | Genesis-Embodied-AI/Genesis | 26,702 | Python | hughperkins | YilingQiao | Review this pull request change.
Language: Python
File: genesis/utils/array_class.py
PR Title: [FEATURE] Use FIELD_METADATA_CACHE_VALUE for static config
BEFORE CODE:
def get_entities_state(solver):
shape = solver._batch_shape(solver.n_entities_)
kwargs = {
"hibernated": V(dtype=gs.ti_int, shape=sh... | I can see why you're going to ask me to make it possible to set `FIELD_METADATA_CACHE_VALUE` conceptually on an entire dataclass at some point :) | 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. |
para_level: int = dataclasses.field(metadata={FIELD_METADATA_CACHE_VALUE: True})
use_hibernation: bool = dataclasses.field(metadata={FIELD_METADATA_CACHE_VALUE: True})
use_contact_island: bool = dataclasses.field(metadata={FIELD_METADATA_CACHE_VALUE: True})
batch_links_info: bool = dataclasses.field(met... | There was no request ot support floats, only ints and bools. Whilst it seems plausible tht it might work with floats, note that this is not tested. Please consider creating a PR onto gstaichi to add a test for floats. Something like https://github.com/Genesis-Embodied-AI/gstaichi/blob/a00b3aa5154b4b983df0c181cba3d60e20... | @dataclasses.dataclass
class StaticRigidSimCacheKey:
para_level: int = cache_value()
use_hibernation: bool = cache_value()
batch_links_info: bool = cache_value()
batch_dofs_info: bool = cache_value()
batch_joints_info: bool = cache_value()
enable_mujoco_compatibility: bool = cache_value()
en... | @@ -2004,6 +2005,70 @@ def __init__(self):
return ClassEntitiesState()
+# =========================================== StaticRigidSimConfig ===========================================
+
+
+@dataclasses.dataclass
+class StaticRigidSimDummy:
+ para_level: int = dataclasses.field(metadata={FIELD_METADATA_CA... | genesis/utils/array_class.py | 26 | Python | 0.5 | suggestion | 526 | 51 | 51 | false | [FEATURE] Use FIELD_METADATA_CACHE_VALUE for static config | 1,639 | Genesis-Embodied-AI/Genesis | 26,702 | Python | hughperkins | YilingQiao | Review this pull request change.
Language: Python
File: genesis/utils/array_class.py
PR Title: [FEATURE] Use FIELD_METADATA_CACHE_VALUE for static config
BEFORE CODE: para_level: int = dataclasses.field(metadata={FIELD_METADATA_CACHE_VALUE: True})
use_hibernation: bool = dataclasses.field(metadata={FIELD_ME... | There was no request ot support floats, only ints and bools. Whilst it seems plausible tht it might work with floats, note that this is not tested. Please consider creating a PR onto gstaichi to add a test for floats. Something like https://github.com/Genesis-Embodied-AI/gstaichi/blob/a00b3aa5154b4b983df0c181cba3d60e20... | 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_entities_state(solver):
shape = solver._batch_shape(solver.n_entities_)
kwargs = {
"hibernated": V(dtype=gs.ti_int, shape=shape),
}
if use_ndarray:
return StructEntitiesState(**kwargs)
else:
@ti.data_oriented
class ClassEntitiesState:
def __init_... | Why do we have both `StaticRigidSimConfig` and `StaticRigidSimDummy`. We should get rid of the former?
Besides, the name is not quite right. `StaticRigidSimDummy` does not mean anything. What is this class all about? If it is used for kernel caching disambiguation only, then it should rather be called `RigidSimMetad... | def get_entities_state(solver):
shape = solver._batch_shape(solver.n_entities_)
kwargs = {
"hibernated": V(dtype=gs.ti_int, shape=shape),
}
if use_ndarray:
return StructEntitiesState(**kwargs)
else:
@ti.data_oriented
class ClassEntitiesState:
def __init_... | @@ -2004,6 +2005,51 @@ def __init__(self):
return ClassEntitiesState()
+# =========================================== StaticRigidSimConfig ===========================================
+
+cache_value = partial(dataclasses.field, metadata={FIELD_METADATA_CACHE_VALUE: True})
+
+
+@dataclasses.dataclass
+class ... | genesis/utils/array_class.py | 26 | Python | 0.643 | performance | 398 | 51 | 51 | false | [FEATURE] Use FIELD_METADATA_CACHE_VALUE for static config | 1,639 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/utils/array_class.py
PR Title: [FEATURE] Use FIELD_METADATA_CACHE_VALUE for static config
BEFORE CODE: def get_entities_state(solver):
shape = solver._batch_shape(solver.n_entities_)
kwargs = {
"hibernated": V(dtype=gs.ti_int, shape=sha... | Why do we have both `StaticRigidSimConfig` and `StaticRigidSimDummy`. We should get rid of the former?
Besides, the name is not quite right. `StaticRigidSimDummy` does not mean anything. What is this class all about? If it is used for kernel caching disambiguation only, then it should rather be called `RigidSimMetad... | 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 that using `backend` instead of `test_backend`, breaks genesis pytest...
@pytest.mark.parametrize("test_backend", ["cpu", "gpu"]) # should not affect result
@pytest.mark.parametrize("use_ndarray", [False, True])
def test_gs_num_envs(use_ndarray: bool, enable_pure: bool, test_backend: str, tmp_path: pathlib.Path... | we have fe-ll-cache? Is it newly added? | @pytest.mark.parametrize("test_backend", ["cpu", "gpu"]) # should not affect result
@pytest.mark.parametrize("use_ndarray", [False, True])
def test_gs_num_envs(use_ndarray: bool, enable_pure: bool, test_backend: str, tmp_path: pathlib.Path) -> None:
if sys.platform == "Darwin" and test_backend == "gpu" and use_nda... | @@ -92,7 +93,92 @@ def test_gs_static(
env = dict(os.environ)
env["GS_BETA_PURE"] = "1" if enable_pure else "0"
env["TI_OFFLINE_CACHE_FILE_PATH"] = str(tmp_path)
+
+
+def gs_num_envs_child(args: list[str]):
+ parser = argparse.ArgumentParser()
+ parser.add_argument("--test-backend", typ... | tests/test_fastcache_unified.py | 26 | Python | 0.071 | performance | 39 | 51 | 51 | false | [MISC] Add test of num envs versus recompilation | 1,655 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | hughperkins | Review this pull request change.
Language: Python
File: tests/test_fastcache_unified.py
PR Title: [MISC] Add test of num envs versus recompilation
BEFORE CODE: # note that using `backend` instead of `test_backend`, breaks genesis pytest...
@pytest.mark.parametrize("test_backend", ["cpu", "gpu"]) # should not affec... | we have fe-ll-cache? Is it newly added? | 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.add_entity(
gs.morphs.Box(
size=(0.4, 0.4, 0.4),
pos=(0.0, 0.0, 0.18),
)
)
scene.build(n_envs=args.num_env, env_spacing=(0.5, 0.5))
scene.rigid_solver.collider.detection()
gs.ti.sync()
from genesis.engine.solvers.rigid.rigid_solver_decomp imp... | ```python
pytest.skip(
"fast cache not supported on mac gpus when using ndarray, because mac gpu only supports up to 31 kernel"
" parameters, and we need more than that."
)
``` | scene.add_entity(
gs.morphs.Box(
size=(0.4, 0.4, 0.4),
pos=(0.0, 0.0, 0.18),
)
)
scene.build(n_envs=args.num_env, env_spacing=(0.5, 0.5))
scene.rigid_solver.collider.detection()
gs.ti.sync()
from genesis.engine.solvers.rigid.rigid_solver_decomp import ke... | @@ -92,7 +93,92 @@ def test_gs_static(
env = dict(os.environ)
env["GS_BETA_PURE"] = "1" if enable_pure else "0"
env["TI_OFFLINE_CACHE_FILE_PATH"] = str(tmp_path)
+
+
+def gs_num_envs_child(args: list[str]):
+ parser = argparse.ArgumentParser()
+ parser.add_argument("--test-backend", typ... | tests/test_fastcache_unified.py | 26 | Python | 0.857 | performance | 222 | 51 | 51 | false | [MISC] Add test of num envs versus recompilation | 1,655 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | hughperkins | Review this pull request change.
Language: Python
File: tests/test_fastcache_unified.py
PR Title: [MISC] Add test of num envs versus recompilation
BEFORE CODE: )
scene.add_entity(
gs.morphs.Box(
size=(0.4, 0.4, 0.4),
pos=(0.0, 0.0, 0.18),
)
)
scene.build(n_en... | ```python
pytest.skip(
"fast cache not supported on mac gpus when using ndarray, because mac gpu only supports up to 31 kernel"
" parameters, and we need more than that."
)
``` | 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.add_entity(
gs.morphs.Box(
size=(0.4, 0.4, 0.4),
pos=(0.0, 0.0, 0.18),
)
)
scene.build(n_envs=args.num_env, env_spacing=(0.5, 0.5))
scene.rigid_solver.collider.detection()
gs.ti.sync()
from genesis.engine.solvers.rigid.rigid_solver_decomp imp... | We use `sys.platform != "darwin"` everywhere else. | scene.add_entity(
gs.morphs.Box(
size=(0.4, 0.4, 0.4),
pos=(0.0, 0.0, 0.18),
)
)
scene.build(n_envs=args.num_env, env_spacing=(0.5, 0.5))
scene.rigid_solver.collider.detection()
gs.ti.sync()
from genesis.engine.solvers.rigid.rigid_solver_decomp import ke... | @@ -92,7 +93,92 @@ def test_gs_static(
env = dict(os.environ)
env["GS_BETA_PURE"] = "1" if enable_pure else "0"
env["TI_OFFLINE_CACHE_FILE_PATH"] = str(tmp_path)
+
+
+def gs_num_envs_child(args: list[str]):
+ parser = argparse.ArgumentParser()
+ parser.add_argument("--test-backend", typ... | tests/test_fastcache_unified.py | 26 | Python | 0.429 | suggestion | 50 | 51 | 51 | false | [MISC] Add test of num envs versus recompilation | 1,655 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | hughperkins | Review this pull request change.
Language: Python
File: tests/test_fastcache_unified.py
PR Title: [MISC] Add test of num envs versus recompilation
BEFORE CODE: )
scene.add_entity(
gs.morphs.Box(
size=(0.4, 0.4, 0.4),
pos=(0.0, 0.0, 0.18),
)
)
scene.build(n_en... | We use `sys.platform != "darwin"` everywhere else. | 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. |
euler : tuple, shape (3,), optional
The euler angle of the entity in degrees. This follows scipy's extrinsic x-y-z rotation convention.
Defaults to (0.0, 0.0, 0.0).
quat : tuple, shape (4,), optional
The quaternion (w-x-y-z convention) of the entity. If specified, `euler` will be ignored... | `tile_size` would make more sense? | euler : tuple, shape (3,), optional
The euler angle of the entity in degrees. This follows scipy's extrinsic x-y-z rotation convention.
Defaults to (0.0, 0.0, 0.0).
quat : tuple, shape (4,), optional
The quaternion (w-x-y-z convention) of the entity. If specified, `euler` will be ignored... | @@ -429,10 +429,16 @@ class Plane(Primitive):
conaffinity : int, optional
The 32-bit integer bitmasks used for contact filtering of contact pairs. When the conaffinity of one geom and
the contype of the other geom share a common bit set to 1, two geoms can collide. Defaults to 0xFFFF.
+ size: ... | genesis/options/morphs.py | 26 | Python | 0.214 | question | 34 | 51 | 51 | false | [MISC] Expose parameters for ground plane tiling. | 1,657 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | yuhongyi | Review this pull request change.
Language: Python
File: genesis/options/morphs.py
PR Title: [MISC] Expose parameters for ground plane tiling.
BEFORE CODE: euler : tuple, shape (3,), optional
The euler angle of the entity in degrees. This follows scipy's extrinsic x-y-z rotation convention.
Defau... | `tile_size` would make more 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. |
pos : tuple, shape (3,), optional
The center position of the plane in meters. Defaults to (0.0, 0.0, 0.0).
euler : tuple, shape (3,), optional
The euler angle of the entity in degrees. This follows scipy's extrinsic x-y-z rotation convention.
Defaults to (0.0, 0.0, 0.0).
quat : tuple... | `plane_size` may be? `size` is not very clear. | pos : tuple, shape (3,), optional
The center position of the plane in meters. Defaults to (0.0, 0.0, 0.0).
euler : tuple, shape (3,), optional
The euler angle of the entity in degrees. This follows scipy's extrinsic x-y-z rotation convention.
Defaults to (0.0, 0.0, 0.0).
quat : tuple... | @@ -429,10 +429,16 @@ class Plane(Primitive):
conaffinity : int, optional
The 32-bit integer bitmasks used for contact filtering of contact pairs. When the conaffinity of one geom and
the contype of the other geom share a common bit set to 1, two geoms can collide. Defaults to 0xFFFF.
+ size: ... | genesis/options/morphs.py | 26 | Python | 0.214 | suggestion | 46 | 51 | 51 | false | [MISC] Expose parameters for ground plane tiling. | 1,657 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | yuhongyi | Review this pull request change.
Language: Python
File: genesis/options/morphs.py
PR Title: [MISC] Expose parameters for ground plane tiling.
BEFORE CODE: pos : tuple, shape (3,), optional
The center position of the plane in meters. Defaults to (0.0, 0.0, 0.0).
euler : tuple, shape (3,), optional
... | `plane_size` may be? `size` is not very clear. | 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. |
face_normals.append(sphere_face_normals)
n_verts += len(sphere_vertices)
vertices = np.concatenate(vertices)
faces = np.concatenate(faces)
face_normals = np.concatenate(face_normals)
else:
vertices, faces, face_normals = _create_unit_box_impl()
vertic... | extract `half_x, half_y = (size[0] * 0.5, size[1] * 0.5)` | face_normals.append(sphere_face_normals)
n_verts += len(sphere_vertices)
vertices = np.concatenate(vertices)
faces = np.concatenate(faces)
face_normals = np.concatenate(face_normals)
else:
vertices, faces, face_normals = _create_unit_box_impl()
vertic... | @@ -889,21 +889,21 @@ def create_box(extents=None, color=(1.0, 1.0, 1.0, 1.0), bounds=None, wireframe=
return mesh
-def create_plane(size=1e3, color=None, normal=(0.0, 0.0, 1.0)):
+def create_plane(normal=(0.0, 0.0, 1.0), size=(1e3, 1e3), n_tiles=(1e3, 1e3), color=None):
thickness = 1e-2 # for safety
- ... | genesis/utils/mesh.py | 26 | Python | 0.429 | refactor | 57 | 51 | 51 | false | [MISC] Expose parameters for ground plane tiling. | 1,657 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | yuhongyi | Review this pull request change.
Language: Python
File: genesis/utils/mesh.py
PR Title: [MISC] Expose parameters for ground plane tiling.
BEFORE CODE: face_normals.append(sphere_face_normals)
n_verts += len(sphere_vertices)
vertices = np.concatenate(vertices)
faces = np.conc... | extract `half_x, half_y = (size[0] * 0.5, size[1] * 0.5)` | 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. |
half = (size[0] * 0.5, size[1] * 0.5)
verts = np.array(
[
[-half[0], -half[1], 0.0],
[half[0], -half[1], 0.0],
[half[0], half[1], 0.0],
[-half[0], -half[1], 0.0],
[half[0], half[1], 0.0],
[-half[0], half[1], 0.0],
],
... | Same. I would rather export `x` and `y` components. More readable I think. | half_x, half_y = (plane_size[0] * 0.5, plane_size[1] * 0.5)
verts = np.array(
[
[-half_x, -half_y, 0.0],
[half_x, -half_y, 0.0],
[half_x, half_y, 0.0],
[-half_x, -half_y, 0.0],
[half_x, half_y, 0.0],
[-half_x, half_y, 0.0],
... | @@ -916,11 +916,11 @@ def create_plane(size=1e3, color=None, normal=(0.0, 0.0, 1.0)):
uv=np.array(
[
[0, 0],
- [size, 0],
- [size, size],
+ [n_tiles[0], 0],
+ [n_tiles[0], n_tiles[1]],
... | genesis/utils/mesh.py | 26 | Python | 0.357 | suggestion | 74 | 51 | 51 | false | [MISC] Expose parameters for ground plane tiling. | 1,657 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | yuhongyi | Review this pull request change.
Language: Python
File: genesis/utils/mesh.py
PR Title: [MISC] Expose parameters for ground plane tiling.
BEFORE CODE: half = (size[0] * 0.5, size[1] * 0.5)
verts = np.array(
[
[-half[0], -half[1], 0.0],
[half[0], -half[1], 0.0],
[h... | Same. I would rather export `x` and `y` components. More readable 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. |
for i in range(100):
if is_done:
is_done = False
break
time.sleep(0.1)
else:
raise AssertionError("Keyboard event not processed before timeout")
else:
pyrender_viewer.dispatch_pending_events()
pyrender_viewer.dispatc... | Do not mark this test as required, it is too long:
<img width="997" height="131" alt="image" src="https://github.com/user-attachments/assets/6789beca-a9d2-4f6f-91db-7ef661b679dd" />
| for i in range(100):
if is_done:
is_done = False
break
time.sleep(0.1)
else:
raise AssertionError("Keyboard event not processed before timeout")
else:
pyrender_viewer.dispatch_pending_events()
pyrender_viewer.dispatc... | @@ -777,3 +777,55 @@ def on_key_press(self, symbol: int, modifiers: int):
# Make sure that the result is valid
with open(IMAGE_FILENAME, "rb") as f:
assert f.read() == png_snapshot
+
+
+@pytest.mark.required | tests/test_render.py | 26 | Python | 0.429 | refactor | 184 | 51 | 51 | false | [MISC] Expose parameters for ground plane tiling. | 1,657 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | yuhongyi | Review this pull request change.
Language: Python
File: tests/test_render.py
PR Title: [MISC] Expose parameters for ground plane tiling.
BEFORE CODE: for i in range(100):
if is_done:
is_done = False
break
time.sleep(0.1)
else:
raise... | Do not mark this test as required, it is too long:
<img width="997" height="131" alt="image" src="https://github.com/user-attachments/assets/6789beca-a9d2-4f6f-91db-7ef661b679dd" />
| 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. |
lookat=(0.0, 0.0, 0.5),
fov=45,
GUI=False,
)
scene.add_light(
pos=[0.0, 0.0, 1.5],
dir=[1.0, 1.0, -2.0],
directional=1,
castshadow=1,
cutoff=45.0,
intensity=0.5,
)
scene.build()
... | Remove all the comments. This test is simple. No need to explain what is going on. | import itertools
import os
import re
import sys
import time
import enum
import numpy as np
import pyglet
import pytest
import torch
import genesis as gs
import genesis.utils.geom as gu
from genesis.utils import set_random_seed
from genesis.utils.image_exporter import FrameImageExporter
from genesis.utils.misc import ... | @@ -777,3 +777,55 @@ def on_key_press(self, symbol: int, modifiers: int):
# Make sure that the result is valid
with open(IMAGE_FILENAME, "rb") as f:
assert f.read() == png_snapshot
+
+
+@pytest.mark.required
+@pytest.mark.parametrize(
+ "renderer_type",
+ [RENDERER_TYPE.BATCHRENDER_RASTERIZER],... | tests/test_render.py | 26 | Python | 0.286 | suggestion | 82 | 29 | 817 | false | [MISC] Expose parameters for ground plane tiling. | 1,657 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | yuhongyi | Review this pull request change.
Language: Python
File: tests/test_render.py
PR Title: [MISC] Expose parameters for ground plane tiling.
BEFORE CODE: lookat=(0.0, 0.0, 0.5),
fov=45,
GUI=False,
)
scene.add_light(
pos=[0.0, 0.0, 1.5],
dir=[1.... | Remove all the comments. This test is simple. No need to explain what is going on. | 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. |
camera = scene.add_camera(
res=CAM_RES,
pos=(1.5, 0.5, 1.5),
lookat=(0.0, 0.0, 0.5),
fov=45,
GUI=False,
)
scene.add_light(
pos=[0.0, 0.0, 1.5],
dir=[1.0, 1.0, -2.0],
directional=1,
castsha... | Remove for loop and do not call `scene.step` at all. It is not necessary and useful in this context. | import itertools
import os
import re
import sys
import time
import enum
import numpy as np
import pyglet
import pytest
import torch
import genesis as gs
import genesis.utils.geom as gu
from genesis.utils import set_random_seed
from genesis.utils.image_exporter import FrameImageExporter
from genesis.utils.misc import ... | @@ -777,3 +777,55 @@ def on_key_press(self, symbol: int, modifiers: int):
# Make sure that the result is valid
with open(IMAGE_FILENAME, "rb") as f:
assert f.read() == png_snapshot
+
+
+@pytest.mark.required
+@pytest.mark.parametrize(
+ "renderer_type",
+ [RENDERER_TYPE.BATCHRENDER_RASTERIZER],... | tests/test_render.py | 26 | Python | 0.571 | suggestion | 100 | 32 | 817 | false | [MISC] Expose parameters for ground plane tiling. | 1,657 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | yuhongyi | Review this pull request change.
Language: Python
File: tests/test_render.py
PR Title: [MISC] Expose parameters for ground plane tiling.
BEFORE CODE: camera = scene.add_camera(
res=CAM_RES,
pos=(1.5, 0.5, 1.5),
lookat=(0.0, 0.0, 0.5),
fov=45,
GUI=F... | Remove for loop and do not call `scene.step` at all. It is not necessary and useful in this context. | 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. |
break
time.sleep(0.1)
else:
raise AssertionError("Keyboard event not processed before timeout")
else:
pyrender_viewer.dispatch_pending_events()
pyrender_viewer.dispatch_events()
# Skip the rest of the test if necessary
if sys.platform == "linu... | Use `RENDERER_TYPE.RASTERIZER` if possible, because it works on all platforms unlike `RENDERER_TYPE.BATCHRENDER_*` | break
time.sleep(0.1)
else:
raise AssertionError("Keyboard event not processed before timeout")
else:
pyrender_viewer.dispatch_pending_events()
pyrender_viewer.dispatch_events()
# Skip the rest of the test if necessary
if sys.platform == "linu... | @@ -777,3 +777,55 @@ def on_key_press(self, symbol: int, modifiers: int):
# Make sure that the result is valid
with open(IMAGE_FILENAME, "rb") as f:
assert f.read() == png_snapshot
+
+
+@pytest.mark.required
+@pytest.mark.parametrize(
+ "renderer_type",
+ [RENDERER_TYPE.BATCHRENDER_RASTERIZER], | tests/test_render.py | 26 | Python | 0.714 | performance | 114 | 51 | 51 | false | [MISC] Expose parameters for ground plane tiling. | 1,657 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | yuhongyi | Review this pull request change.
Language: Python
File: tests/test_render.py
PR Title: [MISC] Expose parameters for ground plane tiling.
BEFORE CODE: break
time.sleep(0.1)
else:
raise AssertionError("Keyboard event not processed before timeout")
else:
pyre... | Use `RENDERER_TYPE.RASTERIZER` if possible, because it works on all platforms unlike `RENDERER_TYPE.BATCHRENDER_*` | 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. |
# Skip the rest of the test if necessary
if sys.platform == "linux":
glinfo = pyrender_viewer.context.get_info()
renderer = glinfo.get_renderer()
if "llvmpipe" in renderer:
llvm_version = re.search(r"LLVM\s+([\d.]+)", renderer).group(1)
if llvm_version < "20":
... | Are you sure you need a full grid? What about:
```
for test_idx, (size, n_tiles) in enumerate((
((1.5, 1.8), (0.5, 0.75)),
((10.0, 7.0), (5.0, 3.0)),
((10.0, 7.0), None),
)):
``` |
# Skip the rest of the test if necessary
if sys.platform == "linux":
glinfo = pyrender_viewer.context.get_info()
renderer = glinfo.get_renderer()
if "llvmpipe" in renderer:
llvm_version = re.search(r"LLVM\s+([\d.]+)", renderer).group(1)
if llvm_version < "20":
... | @@ -777,3 +777,55 @@ def on_key_press(self, symbol: int, modifiers: int):
# Make sure that the result is valid
with open(IMAGE_FILENAME, "rb") as f:
assert f.read() == png_snapshot
+
+
+@pytest.mark.required
+@pytest.mark.parametrize(
+ "renderer_type",
+ [RENDERER_TYPE.BATCHRENDER_RASTERIZER],... | tests/test_render.py | 26 | Python | 0.714 | suggestion | 197 | 51 | 51 | false | [MISC] Expose parameters for ground plane tiling. | 1,657 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | yuhongyi | Review this pull request change.
Language: Python
File: tests/test_render.py
PR Title: [MISC] Expose parameters for ground plane tiling.
BEFORE CODE:
# Skip the rest of the test if necessary
if sys.platform == "linux":
glinfo = pyrender_viewer.context.get_info()
renderer = glinfo.get_render... | Are you sure you need a full grid? What about:
```
for test_idx, (size, n_tiles) in enumerate((
((1.5, 1.8), (0.5, 0.75)),
((10.0, 7.0), (5.0, 3.0)),
((10.0, 7.0), 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. |
@pytest.mark.required
@pytest.mark.parametrize(
"renderer_type",
[RENDERER_TYPE.BATCHRENDER_RASTERIZER],
)
def test_render_planes(tmp_path, png_snapshot, renderer):
# Manually run this test with different parameters to make sure they run sequentially
# Running in parallel may crash the test OOM.
siz... | What about using a "perfect" top view camera angle? It would make more sense. It would even be possible to check if tiling it right based on the RGB data. | @pytest.mark.parametrize(
"renderer_type",
[RENDERER_TYPE.RASTERIZER],
)
def test_render_planes(tmp_path, png_snapshot, renderer):
for test_idx, (plane_size, tile_size) in enumerate(
(
((3, 4.5), (0.5, 0.75)),
((3.0, 5.0), (5.0, 3.0)),
((4.0, 4.0), (1.0, 1.0)),
... | @@ -777,3 +777,55 @@ def on_key_press(self, symbol: int, modifiers: int):
# Make sure that the result is valid
with open(IMAGE_FILENAME, "rb") as f:
assert f.read() == png_snapshot
+
+
+@pytest.mark.required
+@pytest.mark.parametrize(
+ "renderer_type",
+ [RENDERER_TYPE.BATCHRENDER_RASTERIZER],... | tests/test_render.py | 26 | Python | 0.357 | question | 154 | 51 | 36 | false | [MISC] Expose parameters for ground plane tiling. | 1,657 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | yuhongyi | Review this pull request change.
Language: Python
File: tests/test_render.py
PR Title: [MISC] Expose parameters for ground plane tiling.
BEFORE CODE: @pytest.mark.required
@pytest.mark.parametrize(
"renderer_type",
[RENDERER_TYPE.BATCHRENDER_RASTERIZER],
)
def test_render_planes(tmp_path, png_snapshot, rend... | What about using a "perfect" top view camera angle? It would make more sense. It would even be possible to check if tiling it right based on the RGB data. | 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. |
# Skip the rest of the test if necessary
if sys.platform == "linux":
glinfo = pyrender_viewer.context.get_info()
renderer = glinfo.get_renderer()
if "llvmpipe" in renderer:
llvm_version = re.search(r"LLVM\s+([\d.]+)", renderer).group(1)
if llvm_version < "20":
... | `CAM_RES = (256, 256)` would be faster to render and good enough for this. | # Skip the rest of the test if necessary
if sys.platform == "linux":
glinfo = pyrender_viewer.context.get_info()
renderer = glinfo.get_renderer()
if "llvmpipe" in renderer:
llvm_version = re.search(r"LLVM\s+([\d.]+)", renderer).group(1)
if llvm_version < "20":
... | @@ -777,3 +777,55 @@ def on_key_press(self, symbol: int, modifiers: int):
# Make sure that the result is valid
with open(IMAGE_FILENAME, "rb") as f:
assert f.read() == png_snapshot
+
+
+@pytest.mark.required
+@pytest.mark.parametrize(
+ "renderer_type",
+ [RENDERER_TYPE.BATCHRENDER_RASTERIZER],... | tests/test_render.py | 26 | Python | 0.357 | suggestion | 74 | 51 | 50 | false | [MISC] Expose parameters for ground plane tiling. | 1,657 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | yuhongyi | Review this pull request change.
Language: Python
File: tests/test_render.py
PR Title: [MISC] Expose parameters for ground plane tiling.
BEFORE CODE: # Skip the rest of the test if necessary
if sys.platform == "linux":
glinfo = pyrender_viewer.context.get_info()
renderer = glinfo.get_rendere... | `CAM_RES = (256, 256)` would be faster to render and good enough for this. | 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. |
# Running in parallel may crash the test OOM.
sizes = [(1.5, 1.8), (10.0, 7.0)]
n_tiles = [(0.5, 0.75), (5.0, 3.0), None]
for test_idx, (size, n_tiles) in enumerate(itertools.product(sizes, n_tiles)):
CAM_RES = (512, 512)
NUM_STEPS = 1
scene = gs.Scene(
renderer=rende... | This is not good enough I think. You should ideally check that the pattern is correct, or at least that all the images are different. (this is done in advanced render api test). | ((3, 4.5), (0.5, 0.75)),
((3.0, 5.0), (5.0, 3.0)),
((4.0, 4.0), (1.0, 1.0)),
)
):
CAM_RES = (256, 256)
scene = gs.Scene(
renderer=renderer,
)
plane = scene.add_entity(
gs.morphs.Plane(plane_size=plane_size, tile_size... | @@ -777,3 +777,55 @@ def on_key_press(self, symbol: int, modifiers: int):
# Make sure that the result is valid
with open(IMAGE_FILENAME, "rb") as f:
assert f.read() == png_snapshot
+
+
+@pytest.mark.required
+@pytest.mark.parametrize(
+ "renderer_type",
+ [RENDERER_TYPE.BATCHRENDER_RASTERIZER],... | tests/test_render.py | 26 | Python | 0.429 | suggestion | 177 | 44 | 29 | false | [MISC] Expose parameters for ground plane tiling. | 1,657 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | yuhongyi | Review this pull request change.
Language: Python
File: tests/test_render.py
PR Title: [MISC] Expose parameters for ground plane tiling.
BEFORE CODE: # Running in parallel may crash the test OOM.
sizes = [(1.5, 1.8), (10.0, 7.0)]
n_tiles = [(0.5, 0.75), (5.0, 3.0), None]
for test_idx, (size, n_tiles... | This is not good enough I think. You should ideally check that the pattern is correct, or at least that all the images are different. (this is done in advanced render api test). | 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=(0.1, 0.1, 0.1), pos=(0.5, 0, 0.05)),
material=gs.materials.Rigid(),
)
sphere = scene.add_entity(
gs.morphs.Sphere(radius=0.05, pos=(-0.5, 0, 0.05)),
material=gs.materials.Rigid(),
)
scene.build()
all_aabbs = scene.sim.rigid_solver.get_aabb()
... | Can you also check the actual value of the aabbs for box and sphere since you already know their pos and size? | gs.morphs.Box(size=(0.1, 0.1, 0.1), pos=(0.5, 0, 0.05)),
material=gs.materials.Rigid(),
)
sphere = scene.add_entity(
gs.morphs.Sphere(radius=0.05, pos=(-0.5, 0, 0.05)),
material=gs.materials.Rigid(),
)
scene.build()
all_aabbs = scene.sim.rigid_solver.get_aabb()
... | @@ -2910,3 +2910,46 @@ def test_mesh_primitive_COM(show_viewer, tol):
# root and link COM should be the same for single link
assert_allclose(root_bunny_z, bunny_z, atol=tol)
assert_allclose(root_cube_z, cube_z, atol=tol)
+
+
+@pytest.mark.required
+@pytest.mark.parametrize("backend", [gs.cpu])
+def test_... | tests/test_rigid_physics.py | 26 | Python | 0.357 | question | 110 | 29 | 39 | false | [FEATURE] Batched AABB | 1,658 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | LeonLiu4 | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [FEATURE] Batched AABB
BEFORE CODE: gs.morphs.Box(size=(0.1, 0.1, 0.1), pos=(0.5, 0, 0.05)),
material=gs.materials.Rigid(),
)
sphere = scene.add_entity(
gs.morphs.Sphere(radius=0.05, pos=(-0... | Can you also check the actual value of the aabbs for box and sphere since you already know their pos and size? | 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... | I’m wondering if it is possible to write a simpler unit test without robot that only focuses on assuming noslip.
What about considering 2 cubes touching each over on one side, with one of them being fixed but not the other, then you apply a “strong” normal force on the moving one and you check that it does not fall. | import math
import os
import sys
import tempfile
from typing import cast
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_... | tests/test_rigid_physics.py | 1 | Python | 0.5 | suggestion | 320 | 26 | 26 | false | [FEATURE] Add noslip solver | 1,669 | 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] Add noslip solver
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
import pytest
import torch
import trim... | I’m wondering if it is possible to write a simpler unit test without robot that only focuses on assuming noslip.
What about considering 2 cubes touching each over on one side, with one of them being fixed but not the other, then you apply a “strong” normal force on the moving one and you check that it does not fall. | 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. | |
Integrator type. Current supported integrators are 'gs.integrator.Euler', 'gs.integrator.implicitfast' and
'gs.integrator.approximate_implicitfast'. 'Euler' and 'implicitfast' are consistent with their Mujoco
counterpart. 'approximate_implicitfast' is an even faster approximation of 'implicitfas... | I don't think we should say that. Maybe something like
```
This option should only be enabled if necessary because it is still experimental and it will slow down the simulation.
``` | Integrator type. Current supported integrators are 'gs.integrator.Euler', 'gs.integrator.implicitfast' and
'gs.integrator.approximate_implicitfast'. 'Euler' and 'implicitfast' are consistent with their Mujoco
counterpart. 'approximate_implicitfast' is an even faster approximation of 'implicitfas... | @@ -248,6 +248,13 @@ class RigidOptions(Options):
Number of line search iterations for the constraint solver. Defaults to 50.
ls_tolerance : float, optional
Tolerance for the line search. Defaults to 1e-2.
+ noslip_iterations : int, optional
+ Number of iterations for the noslip solver.... | genesis/options/solvers.py | 26 | Python | 0.786 | performance | 184 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/options/solvers.py
PR Title: [FEATURE] Add noslip solver
BEFORE CODE: Integrator type. Current supported integrators are 'gs.integrator.Euler', 'gs.integrator.implicitfast' and
'gs.integrator.approximate_implicitfast'. 'Euler' and 'impl... | I don't think we should say that. Maybe something like
```
This option should only be enabled if necessary because it is still experimental and it will slow down the simulation.
``` | 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. |
cg_beta: V_ANNOTATION
cg_pg_dot_pMg: V_ANNOTATION
# Optional Newton fields
nt_H: V_ANNOTATION
nt_vec: V_ANNOTATION
def get_constraint_state(constraint_solver, solver):
f_batch = solver._batch_shape
len_constraints = constraint_solver.len_constraints
len_constraints_ = constraint_solver... | Only move to new line if necessary:
```python
gs.logger.warning(
f"efc_AR shape {len_constraints_}x{len_constraints_}x{f_batch()[0]} is very large. Consider manually "
f"set a smaller 'max_collision_pairs' in RigidOptions to reduce the size of reserved memory. "
... | cg_beta: V_ANNOTATION
cg_pg_dot_pMg: V_ANNOTATION
# Optional Newton fields
nt_H: V_ANNOTATION
nt_vec: V_ANNOTATION
def get_constraint_state(constraint_solver, solver):
f_batch = solver._batch_shape
len_constraints = constraint_solver.len_constraints
len_constraints_ = constraint_solver... | @@ -169,6 +171,19 @@ def get_constraint_state(constraint_solver, solver):
"Consider reducing the number of constraints or the number of degrees of freedom."
)
+ if solver._static_rigid_sim_config.noslip_iterations > 0:
+ if len_constraints_ * len_constraints_ * f_batch()[0] > 2 * 10**9... | genesis/utils/array_class.py | 26 | Python | 1 | performance | 330 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/utils/array_class.py
PR Title: [FEATURE] Add noslip solver
BEFORE CODE: cg_beta: V_ANNOTATION
cg_pg_dot_pMg: V_ANNOTATION
# Optional Newton fields
nt_H: V_ANNOTATION
nt_vec: V_ANNOTATION
def get_constraint_state(constraint_solver,... | Only move to new line if necessary:
```python
gs.logger.warning(
f"efc_AR shape {len_constraints_}x{len_constraints_}x{f_batch()[0]} is very large. Consider manually "
f"set a smaller 'max_collision_pairs' in RigidOptions to reduce the size of reserved memory. "
... | 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. |
assert_allclose(cube_z, cube.get_links_pos(links_idx_local=[0], ref="link_com")[0, 2], atol=gs.EPS)
assert_allclose(root_bunny_z, bunny.get_links_pos(links_idx_local=[0], ref="root_com")[0, 2], atol=gs.EPS)
assert_allclose(root_cube_z, cube.get_links_pos(links_idx_local=[0], ref="root_com")[0, 2], atol=gs.E... | Are these options really useful? | scene.step()
assert_allclose(box2.get_pos(), box3.get_pos(), atol=1e-3)
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.... | @@ -2928,6 +2928,52 @@ def test_mesh_primitive_COM(show_viewer, tol):
assert_allclose(root_cube_z, cube_z, atol=tol)
+@pytest.mark.required
+@pytest.mark.parametrize("scale", [0.1, 10.0])
+@pytest.mark.parametrize("backend", [gs.cpu, gs.gpu])
+def test_noslip_iterations(scale, show_viewer, tol):
+ scene = g... | tests/test_rigid_physics.py | 26 | Python | 0.143 | question | 32 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | 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] Add noslip solver
BEFORE CODE: assert_allclose(cube_z, cube.get_links_pos(links_idx_local=[0], ref="link_com")[0, 2], atol=gs.EPS)
assert_allclose(root_bunny_z, bunny.get_links_pos(links_idx_local=[0], re... | Are these options really useful? | 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
def test_batched_aabb(tol):
scene = gs.Scene()
plane = scene.add_entity(
gs.morphs.Plane(
normal=(0, 0, 1),
pos=(0, 0, 0),
),
)
box = scene.add_entity(
gs.morphs.Box(
size=(0.1, 0.1, 0.1),
pos=(0.5, 0, 0.05),... | Use floats for floats: `(i * scale, 0.0, 0.0)` | assert_allclose(root_bunny_z, bunny_z, atol=tol)
assert_allclose(root_cube_z, cube_z, atol=tol)
@pytest.mark.required
@pytest.mark.parametrize("scale", [0.1, 10.0])
@pytest.mark.parametrize("backend", [gs.cpu, gs.gpu])
def test_noslip_iterations(scale, show_viewer, tol):
scene = gs.Scene(
sim_opti... | @@ -2928,6 +2928,52 @@ def test_mesh_primitive_COM(show_viewer, tol):
assert_allclose(root_cube_z, cube_z, atol=tol)
+@pytest.mark.required
+@pytest.mark.parametrize("scale", [0.1, 10.0])
+@pytest.mark.parametrize("backend", [gs.cpu, gs.gpu])
+def test_noslip_iterations(scale, show_viewer, tol):
+ scene = g... | tests/test_rigid_physics.py | 26 | Python | 0.286 | suggestion | 46 | 44 | 51 | false | [FEATURE] Add noslip solver | 1,669 | 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] Add noslip solver
BEFORE CODE:
@pytest.mark.required
def test_batched_aabb(tol):
scene = gs.Scene()
plane = scene.add_entity(
gs.morphs.Plane(
normal=(0, 0, 1),
pos=(0, 0... | Use floats for floats: `(i * scale, 0.0, 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. |
scene = gs.Scene(
viewer_options=gs.options.ViewerOptions(
camera_pos=(2.8, -1, 1.3),
camera_lookat=(0.0, 0.0, 0.5),
camera_fov=30,
res=(960, 640),
max_FPS=600,
),
sim_options=gs.options.SimOptions(
dt=0.01,
),
... | I would recommend adding parentheses: `fixed=(i == 0)` | gs.morphs.Plane(),
)
bunny = scene.add_entity(
gs.morphs.Mesh(
file="meshes/bunny.obj",
pos=(-1.0, -1.0, 1.0),
),
vis_mode="collision",
)
cube = scene.add_entity(
gs.morphs.Box(
size=(0.5, 0.5, 0.5),
pos=(1.0, 1.0, 1... | @@ -2928,6 +2928,52 @@ def test_mesh_primitive_COM(show_viewer, tol):
assert_allclose(root_cube_z, cube_z, atol=tol)
+@pytest.mark.required
+@pytest.mark.parametrize("scale", [0.1, 10.0])
+@pytest.mark.parametrize("backend", [gs.cpu, gs.gpu])
+def test_noslip_iterations(scale, show_viewer, tol):
+ scene = g... | tests/test_rigid_physics.py | 26 | Python | 0.357 | suggestion | 54 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | 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] Add noslip solver
BEFORE CODE: scene = gs.Scene(
viewer_options=gs.options.ViewerOptions(
camera_pos=(2.8, -1, 1.3),
camera_lookat=(0.0, 0.0, 0.5),
camera_fov=30,
... | I would recommend adding parentheses: `fixed=(i == 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. |
boxes = []
for i in range(3):
boxes.append(
scene.add_entity(
gs.morphs.Box(
size=(1 * scale, 1 * scale, 1 * scale),
pos=(i * scale, 0, 0),
fixed=i == 0,
),
)
)
scene.build()
... | I'm ok with using 3 boxes, but 2 boxes should be enough. | scene.build()
rigid = scene.sim.rigid_solver
for i in range(500):
scene.step()
link_COM = rigid.get_links_pos(ref="link_com")
root_COM = rigid.get_links_pos(ref="root_com")
bunny_z = link_COM[1, 2]
cube_z = link_COM[2, 2]
root_bunny_z = root_COM[1, 2]
root_cube_z = root_COM[... | @@ -2928,6 +2928,52 @@ def test_mesh_primitive_COM(show_viewer, tol):
assert_allclose(root_cube_z, cube_z, atol=tol)
+@pytest.mark.required
+@pytest.mark.parametrize("scale", [0.1, 10.0])
+@pytest.mark.parametrize("backend", [gs.cpu, gs.gpu])
+def test_noslip_iterations(scale, show_viewer, tol):
+ scene = g... | tests/test_rigid_physics.py | 26 | Python | 0.286 | suggestion | 56 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | 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] Add noslip solver
BEFORE CODE:
boxes = []
for i in range(3):
boxes.append(
scene.add_entity(
gs.morphs.Box(
size=(1 * scale, 1 * scale, 1 * scale),... | I'm ok with using 3 boxes, but 2 boxes 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. |
),
sim_options=gs.options.SimOptions(
dt=0.01,
),
profiling_options=gs.options.ProfilingOptions(show_FPS=False),
rigid_options=gs.options.RigidOptions(
noslip_iterations=5,
),
show_viewer=show_viewer,
)
boxes = []
for i in rang... | Ideally you should set the friction coefficient yourself, and set the normal force so that to box does not slip (with some safety factor, i.e. `F_n = 2.0 * box_mass * g / friction_coef`, make sure that the box does not move, then decrease the normal force using `F_n = 0.5 * box_mass * g / friction_coef` and check that... | vis_mode="collision",
)
cube = scene.add_entity(
gs.morphs.Box(
size=(0.5, 0.5, 0.5),
pos=(1.0, 1.0, 1.0),
),
vis_mode="collision",
)
############# build ##########################
scene.build()
rigid = scene.sim.rigid_solver
for i in r... | @@ -2928,6 +2928,52 @@ def test_mesh_primitive_COM(show_viewer, tol):
assert_allclose(root_cube_z, cube_z, atol=tol)
+@pytest.mark.required
+@pytest.mark.parametrize("scale", [0.1, 10.0])
+@pytest.mark.parametrize("backend", [gs.cpu, gs.gpu])
+def test_noslip_iterations(scale, show_viewer, tol):
+ scene = g... | tests/test_rigid_physics.py | 26 | Python | 0.643 | performance | 340 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | 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] Add noslip solver
BEFORE CODE: ),
sim_options=gs.options.SimOptions(
dt=0.01,
),
profiling_options=gs.options.ProfilingOptions(show_FPS=False),
rigid_options=gs... | Ideally you should set the friction coefficient yourself, and set the normal force so that to box does not slip (with some safety factor, i.e. `F_n = 2.0 * box_mass * g / friction_coef`, make sure that the box does not move, then decrease the normal force using `F_n = 0.5 * box_mass * g / friction_coef` and check that... | 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 genesis as gs
import genesis.utils.array_class as array_class
import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
@gs.maybe_pure
@ti.kernel
def kernel_build_efc_AR_b(
dofs_state: array_class.DofsState,
entities_info: array_class.EntitiesInfo,
rigid_global_info: array_class.Rigi... | I would suggest to use either `i, j` or `i_r, i_c` instead of `r, c` | import genesis as gs
import genesis.utils.array_class as array_class
import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
@gs.maybe_pure
@ti.kernel
def kernel_build_efc_AR_b(
dofs_state: array_class.DofsState,
entities_info: array_class.EntitiesInfo,
rigid_global_info: array_class.Rigi... | @@ -0,0 +1,331 @@
+import gstaichi as ti
+
+import genesis as gs
+import genesis.utils.array_class as array_class
+
+import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
+
+
+@gs.maybe_pure
+@ti.kernel
+def kernel_build_efc_AR_b(
+ dofs_state: array_class.DofsState,
+ entities_info: array_class... | genesis/engine/solvers/rigid/constraint_noslip.py | 26 | Python | 0.429 | suggestion | 68 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_noslip.py
PR Title: [FEATURE] Add noslip solver
BEFORE CODE: import genesis as gs
import genesis.utils.array_class as array_class
import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
@gs.maybe_pure
... | I would suggest to use either `i, j` or `i_r, i_c` instead of `r, c` | 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. |
ti.loop_config(serialize=static_rigid_sim_config.para_level < gs.PARA_LEVEL.PARTIAL)
for i_b in range(_B):
nefc = constraint_state.n_constraints[i_b]
# zero AR
for r in range(nefc):
for c in range(nefc):
constraint_state.efc_AR[r, c, i_b] = gs.ti_float(0.0)
... | replace `dot` by `*` for consistency with the above formula (even though it should probably be `@` everywhere instead) |
ti.loop_config(serialize=static_rigid_sim_config.para_level < gs.PARA_LEVEL.PARTIAL)
for i_b in range(_B):
nefc = constraint_state.n_constraints[i_b]
# zero AR
for i_row in range(nefc):
for i_col in range(nefc):
constraint_state.efc_AR[i_row, i_col, i_b] = gs... | @@ -0,0 +1,331 @@
+import gstaichi as ti
+
+import genesis as gs
+import genesis.utils.array_class as array_class
+
+import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
+
+
+@gs.maybe_pure
+@ti.kernel
+def kernel_build_efc_AR_b(
+ dofs_state: array_class.DofsState,
+ entities_info: array_class... | genesis/engine/solvers/rigid/constraint_noslip.py | 26 | Python | 0.571 | suggestion | 118 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_noslip.py
PR Title: [FEATURE] Add noslip solver
BEFORE CODE:
ti.loop_config(serialize=static_rigid_sim_config.para_level < gs.PARA_LEVEL.PARTIAL)
for i_b in range(_B):
nefc = constraint_state.n_constrain... | replace `dot` by `*` for consistency with the above formula (even though it should probably be `@` everywhere 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. |
# build AR = J * inv(M) * J^T
# do it row-by-row: for each row r, tmp = inv(M) * J[r]^T, then AR[r,:] = J * tmp
for r in range(nefc):
# tmp = M^{-1} * Jr^T
for i_d in range(n_dofs):
constraint_state.Mgrad[i_d, i_b] = constraint_state.jac[r, i_d, i_b]
... | Remove this comment if it is irrelevant. Do you think we will support other methods? | # build AR = J * inv(M) * J^T
# do it row-by-row: for each row r, tmp = inv(M) * J[r]^T, then AR[r,:] = J * tmp
for i_row in range(nefc):
# tmp = M^{-1} * Jr^T
for i_d in range(n_dofs):
constraint_state.Mgrad[i_d, i_b] = constraint_state.jac[i_row, i_d, i_... | @@ -0,0 +1,331 @@
+import gstaichi as ti
+
+import genesis as gs
+import genesis.utils.array_class as array_class
+
+import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
+
+
+@gs.maybe_pure
+@ti.kernel
+def kernel_build_efc_AR_b(
+ dofs_state: array_class.DofsState,
+ entities_info: array_class... | genesis/engine/solvers/rigid/constraint_noslip.py | 26 | Python | 0.286 | refactor | 84 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_noslip.py
PR Title: [FEATURE] Add noslip solver
BEFORE CODE: # build AR = J * inv(M) * J^T
# do it row-by-row: for each row r, tmp = inv(M) * J[r]^T, then AR[r,:] = J * tmp
for r in range(nefc):
... | Remove this comment if it is irrelevant. Do you think we will support other methods? | 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 i_d in range(n_dofs):
s += constraint_state.jac[c, i_d, i_b] * constraint_state.Mgrad[i_d, i_b]
constraint_state.efc_AR[r, c, i_b] = s
# we don't need this term if we only use PGS for noslip
# for r in range(nefc):
# constraint_state.e... | Use `n_dofs = constraint_state.jac.shape[1]` for consistency. | for i_d in range(n_dofs):
s += constraint_state.jac[i_col, i_d, i_b] * constraint_state.Mgrad[i_d, i_b]
constraint_state.efc_AR[i_row, i_col, i_b] = s
for i_c in range(constraint_state.n_constraints[i_b]):
v = -constraint_state.aref[i_c, i_b]
... | @@ -0,0 +1,331 @@
+import gstaichi as ti
+
+import genesis as gs
+import genesis.utils.array_class as array_class
+
+import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
+
+
+@gs.maybe_pure
+@ti.kernel
+def kernel_build_efc_AR_b(
+ dofs_state: array_class.DofsState,
+ entities_info: array_class... | genesis/engine/solvers/rigid/constraint_noslip.py | 26 | Python | 0.429 | style | 61 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_noslip.py
PR Title: [FEATURE] Add noslip solver
BEFORE CODE: for i_d in range(n_dofs):
s += constraint_state.jac[c, i_d, i_b] * constraint_state.Mgrad[i_d, i_b]
constra... | Use `n_dofs = constraint_state.jac.shape[1]` for 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. |
ti.loop_config(serialize=static_rigid_sim_config.para_level < gs.PARA_LEVEL.ALL)
for i_b in range(_B):
# temp variables
res = ti.Vector.zero(gs.ti_float, 5)
old_force = ti.Vector.zero(gs.ti_float, 5)
bc = ti.Vector.zero(gs.ti_float, 5)
Ac = ti.Vector.zero(gs.ti_float, 9... | just use `constraint_state.efc_force[i_c, i_b] ** 2`, it will be fine.
You may use `* constraint_state.diag[i_c, i_b]` in place of `/ constraint_state.efc_D[i_c, i_b]` to avoid division, which is much slower than multiplication. | res = ti.Vector.zero(gs.ti_float, 5)
old_force = ti.Vector.zero(gs.ti_float, 5)
bc = ti.Vector.zero(gs.ti_float, 5)
Ac = ti.Vector.zero(gs.ti_float, 9)
n_con = collider_state.n_contacts[i_b]
ne = constraint_state.n_constraints_equality[i_b]
nf = constraint_state.... | @@ -0,0 +1,331 @@
+import gstaichi as ti
+
+import genesis as gs
+import genesis.utils.array_class as array_class
+
+import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
+
+
+@gs.maybe_pure
+@ti.kernel
+def kernel_build_efc_AR_b(
+ dofs_state: array_class.DofsState,
+ entities_info: array_class... | genesis/engine/solvers/rigid/constraint_noslip.py | 26 | Python | 0.643 | performance | 231 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_noslip.py
PR Title: [FEATURE] Add noslip solver
BEFORE CODE:
ti.loop_config(serialize=static_rigid_sim_config.para_level < gs.PARA_LEVEL.ALL)
for i_b in range(_B):
# temp variables
res = ti.Vect... | just use `constraint_state.efc_force[i_c, i_b] ** 2`, it will be fine.
You may use `* constraint_state.diag[i_c, i_b]` in place of `/ constraint_state.efc_D[i_c, i_b]` to avoid division, which is much slower than multiplication. | 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 i_c in range(constraint_state.n_constraints[i_b]):
constraint_state.qfrc_constraint[i_d, i_b] = (
constraint_state.qfrc_constraint[i_d, i_b]
+ constraint_state.jac[i_c, i_d, i_b] * constraint_state.efc_force[i_c, i_b]
)
ri... | You should rename these two functions to be more specific. |
for i_d in range(n_dofs):
constraint_state.qacc[i_d, i_b] = constraint_state.qacc[i_d, i_b] + dofs_state.acc_smooth[i_d, i_b]
dofs_state.acc[i_d, i_b] = constraint_state.qacc[i_d, i_b]
dofs_state.qf_constraint[i_d, i_b] = constraint_state.qfrc_constraint[i_d, i_b]
... | @@ -0,0 +1,331 @@
+import gstaichi as ti
+
+import genesis as gs
+import genesis.utils.array_class as array_class
+
+import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
+
+
+@gs.maybe_pure
+@ti.kernel
+def kernel_build_efc_AR_b(
+ dofs_state: array_class.DofsState,
+ entities_info: array_class... | genesis/engine/solvers/rigid/constraint_noslip.py | 26 | Python | 0.286 | suggestion | 58 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_noslip.py
PR Title: [FEATURE] Add noslip solver
BEFORE CODE:
for i_c in range(constraint_state.n_constraints[i_b]):
constraint_state.qfrc_constraint[i_d, i_b] = (
constrai... | You should rename these two functions to be more specific. | 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 j2 in range(2):
j_efc = base + j2 * 2
res = func_residual(
res=res,
i_b=i_b,
i_efc=j_efc,
dim=2,
flg_subR=True,
cons... | Remove ` + 0`, I don't think it brings any value in this context. | Ac = func_extract_block_matrix_from_AR(
Ac=Ac,
i_b=i_b,
start=j_efc,
n=2,
constraint_state=constraint_state,
)
for j in ti.static(range(2)):... | @@ -0,0 +1,331 @@
+import gstaichi as ti
+
+import genesis as gs
+import genesis.utils.array_class as array_class
+
+import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
+
+
+@gs.maybe_pure
+@ti.kernel
+def kernel_build_efc_AR_b(
+ dofs_state: array_class.DofsState,
+ entities_info: array_class... | genesis/engine/solvers/rigid/constraint_noslip.py | 26 | Python | 0.429 | suggestion | 65 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_noslip.py
PR Title: [FEATURE] Add noslip solver
BEFORE CODE: for j2 in range(2):
j_efc = base + j2 * 2
res = func_residual(
res=res,
... | Remove ` + 0`, I don't think it brings any value in this context. | 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 = const_start + i_col * 4
mu = collider_state.contact_data.friction[i_col, i_b]
for j2 in range(2):
j_efc = base + j2 * 2
res = func_residual(
res=res,
i_b=i_b,
... | Prefer `i, j, k, l, ...` over `i1, i2, i3, i4, ...` | for i2 in ti.static(range(2)):
old_force[i2] = constraint_state.efc_force[j_efc + i2, i_b]
Ac = func_extract_block_matrix_from_AR(
Ac=Ac,
i_b=i_b,
start=j_efc,
... | @@ -0,0 +1,331 @@
+import gstaichi as ti
+
+import genesis as gs
+import genesis.utils.array_class as array_class
+
+import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
+
+
+@gs.maybe_pure
+@ti.kernel
+def kernel_build_efc_AR_b(
+ dofs_state: array_class.DofsState,
+ entities_info: array_class... | genesis/engine/solvers/rigid/constraint_noslip.py | 26 | Python | 0.429 | suggestion | 52 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_noslip.py
PR Title: [FEATURE] Add noslip solver
BEFORE CODE: base = const_start + i_col * 4
mu = collider_state.contact_data.friction[i_col, i_b]
for j2 in range(2):
... | Prefer `i, j, k, l, ...` over `i1, i2, i3, i4, ...` | 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. |
)
rigid_solver.func_solve_mass_batched(
vec=constraint_state.qfrc_constraint,
out=constraint_state.qacc,
i_b=i_b,
entities_info=entities_info,
rigid_global_info=rigid_global_info,
static_rigid_sim_config=static_rigid_sim_co... | Remove this parameter if unused. I can add it back later if necessary. | dofs_state.qf_constraint[i_d, i_b] = constraint_state.qfrc_constraint[i_d, i_b]
dofs_state.force[i_d, i_b] = dofs_state.qf_smooth[i_d, i_b] + constraint_state.qfrc_constraint[i_d, i_b]
@ti.func
def func_extract_block_matrix_from_AR(
Ac,
i_b: int,
start: int,
n: int,
constra... | @@ -0,0 +1,331 @@
+import gstaichi as ti
+
+import genesis as gs
+import genesis.utils.array_class as array_class
+
+import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
+
+
+@gs.maybe_pure
+@ti.kernel
+def kernel_build_efc_AR_b(
+ dofs_state: array_class.DofsState,
+ entities_info: array_class... | genesis/engine/solvers/rigid/constraint_noslip.py | 26 | Python | 0.286 | refactor | 70 | 51 | 51 | false | [FEATURE] Add noslip solver | 1,669 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_noslip.py
PR Title: [FEATURE] Add noslip solver
BEFORE CODE: )
rigid_solver.func_solve_mass_batched(
vec=constraint_state.qfrc_constraint,
out=constraint_state.qacc,
... | Remove this parameter if unused. I can add it back later if 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. |
return res
@ti.func
def func_cost_change(
i_b: int,
Ac,
force: array_class.V_ANNOTATION,
force_start: int,
old_force,
res,
dim: int,
):
change = gs.ti_float(0.0)
if dim == 1:
delta = force[force_start + 0, i_b] - old_force[0]
change = 0.5 * Ac[0] * delta * delta... | Avoid hard-coded constant in the implementation. What is this supposed to be? It should either be some exposed parameter, `gs.EPS`, or at least some global constant defined in uppercase right after the imports. | change += delta[i] * res[i]
if change > gs.EPS:
for i in range(dim):
force[force_start + i, i_b] = old_force[i]
change = 0.0
return change
@gs.maybe_pure
@ti.kernel
def compute_A_diag(
rigid_global_info: array_class.RigidGlobalInfo,
constraint_state: array_class... | @@ -0,0 +1,331 @@
+import gstaichi as ti
+
+import genesis as gs
+import genesis.utils.array_class as array_class
+
+import genesis.engine.solvers.rigid.rigid_solver_decomp as rigid_solver
+
+
+@gs.maybe_pure
+@ti.kernel
+def kernel_build_efc_AR_b(
+ dofs_state: array_class.DofsState,
+ entities_info: array_class... | genesis/engine/solvers/rigid/constraint_noslip.py | 26 | Python | 0.643 | suggestion | 210 | 51 | 46 | false | [FEATURE] Add noslip solver | 1,669 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/constraint_noslip.py
PR Title: [FEATURE] Add noslip solver
BEFORE CODE: return res
@ti.func
def func_cost_change(
i_b: int,
Ac,
force: array_class.V_ANNOTATION,
force_start: int,
old_force,
res,
di... | Avoid hard-coded constant in the implementation. What is this supposed to be? It should either be some exposed parameter, `gs.EPS`, or at least some global constant defined in uppercase right after the 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. |
@pytest.mark.parametrize("n_cubes", [5])
@pytest.mark.parametrize("enable_island", [False])
@pytest.mark.parametrize("n_envs", [2048])
@pytest.mark.parametrize(
"gjk, enable_mujoco_compatibility",
[
(False, True), # MPR
(False, False), # MPR+SDF
(True, False), # GJK
],
)
def test_... | * Why are you adding this in benchmarks?
* You should mark this test as required
* You should explicitly check for something, or document why it is not necessary
* Are you sure doing one step is necessary? Build is not enough? | import hashlib
import numbers
import os
import time
from enum import Enum
from pathlib import Path
from typing import Any
import numpy as np
import pytest
import torch
import wandb
import genesis as gs
from .utils import (
get_hardware_fingerprint,
get_platform_fingerprint,
get_git_commit_timestamp,
... | @@ -684,3 +684,20 @@ def test_box_pyramid(factory_logger, request, n_cubes, solver, enable_island, n_
}
) as logger:
logger.write(request.getfixturevalue("box_pyramid"))
+
+
+@pytest.mark.parametrize("batch_links_info", [False, True]) | tests/test_rigid_benchmarks.py | 26 | Python | 0.5 | question | 229 | 41 | 687 | false | [BUG FIX] Fix entity idx to be always int when loading terrain with batch_links_info. | 1,691 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Kashu7100 | Review this pull request change.
Language: Python
File: tests/test_rigid_benchmarks.py
PR Title: [BUG FIX] Fix entity idx to be always int when loading terrain with batch_links_info.
BEFORE CODE: @pytest.mark.parametrize("n_cubes", [5])
@pytest.mark.parametrize("enable_island", [False])
@pytest.mark.parametrize("n_... | * Why are you adding this in benchmarks?
* You should mark this test as required
* You should explicitly check for something, or document why it is not necessary
* Are you sure doing one step is necessary? Build is not 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. |
import pathlib
import pytest
import subprocess
import sys
import os
import argparse
import genesis as gs
TEST_RAN = "test ran"
def gs_static_child(args: list[str]):
parser = argparse.ArgumentParser()
parser.add_argument("--enable-multi-contact", action="store_true")
parser.add_argument("--expected-num-... | Cross-referencing my comment here: https://github.com/Genesis-Embodied-AI/gstaichi/pull/177/files#r2337243300 | import pathlib
import pytest
import subprocess
import sys
import os
import argparse
import genesis as gs
RET_SUCCESS = 42
def gs_static_child(args: list[str]):
parser = argparse.ArgumentParser()
parser.add_argument("--enable-multi-contact", action="store_true")
parser.add_argument("--expected-num-conta... | tests/test_fastcache_unified.py | 1 | Python | 0.357 | suggestion | 109 | 26 | 26 | false | [BUG FIX] Improve robustness of gstaichi fast-cache unit tests. | 1,709 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | hughperkins | Review this pull request change.
Language: Python
File: tests/test_fastcache_unified.py
PR Title: [BUG FIX] Improve robustness of gstaichi fast-cache unit tests.
BEFORE CODE: import pathlib
import pytest
import subprocess
import sys
import os
import argparse
import genesis as gs
TEST_RAN = "test ran"
def gs_st... | Cross-referencing my comment here: https://github.com/Genesis-Embodied-AI/gstaichi/pull/177/files#r2337243300 | 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. | |
"""
Copy the tensor into the next position of the ring buffer, and advance the index pointer.
Parameters
----------
tensor : torch.Tensor
The tensor to copy into the ring buffer.
"""
self.buffer[self._idx_ptr.value].copy_(tensor)
self._idx_ptr... | I need to figure this out still.. It works when i do arange.. When using :batch_size and reshapes or squeeze(1) or having ifs i could not get it to work for both batched/unbatched case. I was tired when I tried last night though. Help is appreciated 🥲 | def append(self, tensor: torch.Tensor):
"""
Copy the tensor into the next position of the ring buffer, and advance the index pointer.
Parameters
----------
tensor : torch.Tensor
The tensor to copy into the ring buffer.
"""
self.buffer[self._idx].c... | @@ -67,7 +67,8 @@ def at(self, idx: int | torch.Tensor) -> torch.Tensor:
rel_idx = (self._idx_ptr.value - idx) % self.N
if isinstance(idx, torch.Tensor):
batch_size = len(idx) if idx.ndim > 0 else 1
- return self.buffer[rel_idx, :batch_size]
+ batch_indices = torch.a... | genesis/utils/ring_buffer.py | 26 | Python | 0.5 | performance | 251 | 51 | 51 | false | [BUG FIX] Add tests for adding multiple sensors and envs. | 1,726 | Genesis-Embodied-AI/Genesis | 26,702 | Python | Milotrince | Milotrince | Review this pull request change.
Language: Python
File: genesis/utils/ring_buffer.py
PR Title: [BUG FIX] Add tests for adding multiple sensors and envs.
BEFORE CODE: """
Copy the tensor into the next position of the ring buffer, and advance the index pointer.
Parameters
----------
... | I need to figure this out still.. It works when i do arange.. When using :batch_size and reshapes or squeeze(1) or having ifs i could not get it to work for both batched/unbatched case. I was tired when I tried last night though. Help is appreciated 🥲 | 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. |
"""
Copy the tensor into the next position of the ring buffer, and advance the index pointer.
Parameters
----------
tensor : torch.Tensor
The tensor to copy into the ring buffer.
"""
self.buffer[self._idx_ptr.value].copy_(tensor)
self._idx_ptr... | I think there is still something funky going on with the indexing and/or delay handling. I will try to add another test tomorrow. | def append(self, tensor: torch.Tensor):
"""
Copy the tensor into the next position of the ring buffer, and advance the index pointer.
Parameters
----------
tensor : torch.Tensor
The tensor to copy into the ring buffer.
"""
self.buffer[self._idx].c... | @@ -67,7 +67,8 @@ def at(self, idx: int | torch.Tensor) -> torch.Tensor:
rel_idx = (self._idx_ptr.value - idx) % self.N
if isinstance(idx, torch.Tensor):
batch_size = len(idx) if idx.ndim > 0 else 1
- return self.buffer[rel_idx, :batch_size]
+ # using [tensor_indices... | genesis/utils/ring_buffer.py | 26 | Python | 0.357 | suggestion | 129 | 51 | 51 | false | [BUG FIX] Add tests for adding multiple sensors and envs. | 1,726 | Genesis-Embodied-AI/Genesis | 26,702 | Python | Milotrince | Milotrince | Review this pull request change.
Language: Python
File: genesis/utils/ring_buffer.py
PR Title: [BUG FIX] Add tests for adding multiple sensors and envs.
BEFORE CODE: """
Copy the tensor into the next position of the ring buffer, and advance the index pointer.
Parameters
----------
... | I think there is still something funky going on with the indexing and/or delay handling. I will try to add another test tomorrow. | 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._shared_metadata.expanded_links_idx = concat_with_tensor(
self._shared_metadata.expanded_links_idx,
self._options.entity_idx + self._options.link_idx_local,
expand=(self._manager._sim._B, 1, -1),
dim=-1,
)
def _get_return_format(self) -> tuple[i... | Concatenating is probably less efficient that loop over the two arrays separately I think. |
link_start = self._shared_metadata.solver.entities[self._options.entity_idx].link_start
self._shared_metadata.expanded_links_idx = concat_with_tensor(
self._shared_metadata.expanded_links_idx,
link_start + self._options.link_idx_local,
expand=(1,),
dim=0,... | @@ -127,9 +123,12 @@ def _update_shared_ground_truth_cache(
cls, shared_metadata: ContactSensorMetadata, shared_ground_truth_cache: torch.Tensor
):
all_contacts = shared_metadata.solver.collider.get_contacts(as_tensor=True, to_torch=True)
- contact_links = torch.cat([all_contacts["link_a"]... | genesis/sensors/contact_force.py | 26 | Python | 0.214 | performance | 90 | 51 | 51 | false | [BUG FIX] Add tests for adding multiple sensors and envs. | 1,726 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/contact_force.py
PR Title: [BUG FIX] Add tests for adding multiple sensors and envs.
BEFORE CODE:
self._shared_metadata.expanded_links_idx = concat_with_tensor(
self._shared_metadata.expanded_links_idx,
self._op... | Concatenating is probably less efficient that loop over the two arrays separately 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. |
codec_options: dict[str, str]
Additional low-level codec options that will be pass to ffmpeg. Empty by default.
save_on_reset: bool, optional
Whether to save the data on reset. If True, a counter will be added to the filename and incremented on each
reset. Defaults to False.
"""
... | this line will run during import time, which will trigger error if `IS_PYAV_AVAILABLE=False` | codec_options: dict[str, str]
Additional low-level codec options that will be pass to ffmpeg. Empty by default.
save_on_reset: bool, optional
Whether to save the data on reset. If True, a counter will be added to the filename and incremented on each
reset. Defaults to False.
"""
... | @@ -119,19 +119,14 @@ def validate(self):
@register_recording(VideoFileWriterOptions)
class VideoFileWriter(BaseFileWriter):
- video_container: av.container.OutputContainer | None | genesis/recorders/file_writers.py | 26 | Python | 0.5 | suggestion | 92 | 51 | 51 | false | [BUG FIX] fix single_franka_envs example | 1,743 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | YilingQiao | Review this pull request change.
Language: Python
File: genesis/recorders/file_writers.py
PR Title: [BUG FIX] fix single_franka_envs example
BEFORE CODE: codec_options: dict[str, str]
Additional low-level codec options that will be pass to ffmpeg. Empty by default.
save_on_reset: bool, optional
... | this line will run during import time, which will trigger error if `IS_PYAV_AVAILABLE=False` | 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. |
with gs.logger.timer("Running convex decomposition."):
mesh = coacd.Mesh(mesh.vertices, mesh.faces)
args = coacd_options
result = coacd.run_coacd(
mesh,
threshold=args.threshold,
max_convex_hull=args.max_convex_hull,
... | I don't know what you are trying to do here, but this is clearly invalid code. | with gs.logger.timer("Running convex decomposition."):
mesh = coacd.Mesh(mesh.vertices, mesh.faces)
args = coacd_options
result = coacd.run_coacd(
mesh,
threshold=args.threshold,
max_convex_hull=args.max_convex_hull,
... | @@ -300,6 +333,7 @@ def convex_decompose(mesh, coacd_options):
mesh_parts = []
for vs, fs in result:
mesh_parts.append(trimesh.Trimesh(vs, fs))
+ mesh_parts.append(mesh_scale) | genesis/utils/mesh.py | 26 | Python | 0.214 | suggestion | 78 | 51 | 51 | false | [FEATURE] Make convex decomposition cache scale-invariant. | 1,810 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | SonSang | Review this pull request change.
Language: Python
File: genesis/utils/mesh.py
PR Title: [FEATURE] Make convex decomposition cache scale-invariant.
BEFORE CODE: with gs.logger.timer("Running convex decomposition."):
mesh = coacd.Mesh(mesh.vertices, mesh.faces)
args = coacd_options
... | I don't know what you are trying to do here, but this is clearly invalid code. | 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="meshes/duck.obj",
scale=second_scale,
pos=(1.0, 0, 1.0),
quat=(1, 0, 0, 0),
),
)
scene.build()
assert len(seen_paths) == 2, "Expected 2 cache paths, got %d" % len(seen_paths)
assert len(seen_results) == 2, "Expected 2 convex decomposition re... | You have `assert_array_equal` for this but no big deal. | import os
import sys
import numpy as np
import pytest
import trimesh
import genesis as gs
import genesis.utils.gltf as gltf_utils
import genesis.utils.usda as usda_utils
import genesis.utils.mesh as mesh_utils
from .utils import assert_allclose, assert_array_equal, get_hf_dataset
VERTICES_TOL = 1e-05 # Transforma... | @@ -510,3 +511,79 @@ def test_splashsurf_surface_reconstruction(show_viewer):
)
scene.build()
cam.render(rgb=True, depth=False, segmentation=False, colorize_seg=False, normal=False)
+
+
+@pytest.mark.required
+def test_convex_decompose_cache(monkeypatch):
+ # Check if the convex decomposition cache is... | tests/test_mesh.py | 26 | Python | 0.357 | suggestion | 55 | 27 | 587 | false | [FEATURE] Make convex decomposition cache scale-invariant. | 1,810 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | SonSang | Review this pull request change.
Language: Python
File: tests/test_mesh.py
PR Title: [FEATURE] Make convex decomposition cache scale-invariant.
BEFORE CODE: file="meshes/duck.obj",
scale=second_scale,
pos=(1.0, 0, 1.0),
quat=(1, 0, 0, 0),
),
)
scene.bu... | You have `assert_array_equal` for this but no big deal. | 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. |
second_scale = 4.0
duck = scene.add_entity(
morph=gs.morphs.Mesh(
file="meshes/duck.obj",
scale=second_scale,
pos=(1.0, 0, 1.0),
quat=(1, 0, 0, 0),
),
)
scene.build()
assert len(seen_paths) == 2, "Expected 2 cache paths, got %d" % len(... | I don't think it is necessary to add this assert comments. It is self-explanatory here, and right and left side values of equality checks while be displayed automatically by pytest. | second_scale = 4.0
duck = scene.add_entity(
morph=gs.morphs.Mesh(
file="meshes/duck.obj",
scale=second_scale,
pos=(1.0, 0, 1.0),
quat=(1, 0, 0, 0),
),
)
scene.build()
assert len(seen_paths) == 2
assert len(seen_results) == 2
#... | @@ -510,3 +511,79 @@ def test_splashsurf_surface_reconstruction(show_viewer):
)
scene.build()
cam.render(rgb=True, depth=False, segmentation=False, colorize_seg=False, normal=False)
+
+
+@pytest.mark.required
+def test_convex_decompose_cache(monkeypatch):
+ # Check if the convex decomposition cache is... | tests/test_mesh.py | 26 | Python | 0.357 | suggestion | 181 | 30 | 27 | false | [FEATURE] Make convex decomposition cache scale-invariant. | 1,810 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | SonSang | Review this pull request change.
Language: Python
File: tests/test_mesh.py
PR Title: [FEATURE] Make convex decomposition cache scale-invariant.
BEFORE CODE: second_scale = 4.0
duck = scene.add_entity(
morph=gs.morphs.Mesh(
file="meshes/duck.obj",
scale=second_scale,
... | I don't think it is necessary to add this assert comments. It is self-explanatory here, and right and left side values of equality checks while be displayed automatically by 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. |
d,
i_g,
i_b,
):
"""
Count the number of valid support points for the given world direction.
"""
d_mesh = gu.ti_transform_by_quat(d, gu.ti_inv_quat(geoms_state.quat[i_g, i_b]))
return _func_count_supports_mesh(
geoms_state, geoms_info, support_field_info, support_field_static_conf... | changing comments are not necessary | d,
i_g,
i_b,
):
"""
Count the number of valid support points for the given world direction.
"""
d_mesh = gu.ti_transform_by_quat(d, gu.ti_inv_quat(geoms_state.quat[i_g, i_b]))
return _func_count_supports_mesh(
geoms_state, geoms_info, support_field_info, support_field_static_conf... | @@ -350,14 +350,14 @@ def _func_count_supports_mesh(
"""
Count the number of valid support points for a mesh in the given direction.
"""
- theta = ti.atan2(d_mesh[1], d_mesh[0]) # [-pi, pi]
- phi = ti.acos(d_mesh[2]) # [0, pi]
+ theta = ti.atan2(d_mesh[1], d_mesh[0]) # [-math.pi, math.pi] | genesis/engine/solvers/rigid/support_field_decomp.py | 26 | Python | 0.071 | suggestion | 35 | 51 | 51 | false | [MISC] Import math module instead of constants to avoid violating gstaichi pure checker. | 1,812 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | hughperkins | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/support_field_decomp.py
PR Title: [MISC] Import math module instead of constants to avoid violating gstaichi pure checker.
BEFORE CODE: d,
i_g,
i_b,
):
"""
Count the number of valid support points for the given ... | changing comments are not 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. |
i_g,
i_b,
):
"""
support position for a world direction
"""
g_pos = geoms_state.pos[i_g, i_b]
g_quat = geoms_state.quat[i_g, i_b]
d_mesh = gu.ti_transform_by_quat(d, gu.ti_inv_quat(g_quat))
v_, vid = _func_support_mesh(support_field_info, support_field_static_config, d_mesh, i_g)
... | changing comments are not necessary | i_g,
i_b,
):
"""
support position for a world direction
"""
g_pos = geoms_state.pos[i_g, i_b]
g_quat = geoms_state.quat[i_g, i_b]
d_mesh = gu.ti_transform_by_quat(d, gu.ti_inv_quat(g_quat))
v_, vid = _func_support_mesh(support_field_info, support_field_static_config, d_mesh, i_g)
... | @@ -154,16 +154,16 @@ def _func_support_mesh(
"""
support point at mesh frame coordinate.
"""
- theta = ti.atan2(d_mesh[1], d_mesh[0]) # [-pi, pi]
- phi = ti.acos(d_mesh[2]) # [0, pi]
+ theta = ti.atan2(d_mesh[1], d_mesh[0]) # [-math.pi, math.pi] | genesis/engine/solvers/rigid/support_field_decomp.py | 26 | Python | 0.071 | suggestion | 35 | 51 | 51 | false | [MISC] Import math module instead of constants to avoid violating gstaichi pure checker. | 1,812 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | hughperkins | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/support_field_decomp.py
PR Title: [MISC] Import math module instead of constants to avoid violating gstaichi pure checker.
BEFORE CODE: i_g,
i_b,
):
"""
support position for a world direction
"""
g_pos = ge... | changing comments are not 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. |
# Set default rotor armature inertia
if default_armature is not None:
mj.dof_armature[:] = default_armature
mj.body_invweight0[:] = 0.0
mj.dof_invweight0[:] = 0.0
else:
# Apply only those defaults th... | I don't like this entire logic duplication. I hope there should be a better way. We put a lot of effort into unifying URDF and MJCF loading, it would be a shame to do a step back. | mj.body_invweight0[:] = 0.0
mj.dof_invweight0[:] = 0.0
elif isinstance(xml, mujoco.MjModel):
mj = xml
else:
raise gs.raise_exception(f"'{xml}' is not a valid MJCF file.")
return mj
def parse_xml(morph, surface):
# Always merge fixed links unless... | @@ -138,6 +145,26 @@ def build_model(xml, discard_visual, default_armature=None, merge_fixed_links=Fa
mj.dof_armature[:] = default_armature
mj.body_invweight0[:] = 0.0
mj.dof_invweight0[:] = 0.0
+ else:
+ # Apply only those default... | genesis/utils/mjcf.py | 26 | Python | 0.429 | suggestion | 179 | 51 | 51 | false | [BUG FIX] Fix loading MJCF file with includes. | 1,838 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/utils/mjcf.py
PR Title: [BUG FIX] Fix loading MJCF file with includes.
BEFORE CODE:
# Set default rotor armature inertia
if default_armature is not None:
mj.dof_armature[:] = default_armature
... | I don't like this entire logic duplication. I hope there should be a better way. We put a lot of effort into unifying URDF and MJCF loading, it would be a shame to do a step back. | 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 .patterns import DepthCameraPattern
from .raycaster import RaycasterData, RaycasterOptions, RaycasterSensor, RaycasterSharedMetadata
class DepthCameraOptions(RaycasterOptions):
"""
Depth camera that uses ray casting to obtain depth images.
Parameters
----------
pattern: DepthCameraPattern
... | This is not readable. You should rather do:
```python
batch_shape = (self._manager._sim._B,) if self._manager._sim.n_envs > 0 else ()
self._shape = (*batch_shape, self._options.pattern.height, self._options.pattern.width)
``` | from .patterns import DepthCameraPattern
from .raycaster import RaycasterData, RaycasterOptions, RaycasterSensor, RaycasterSharedMetadata
class DepthCameraOptions(RaycasterOptions):
"""
Depth camera that uses ray casting to obtain depth images.
Parameters
----------
pattern: DepthCameraPattern
... | @@ -21,6 +21,14 @@ class DepthCameraOptions(RaycasterOptions):
@register_sensor(DepthCameraOptions, RaycasterSharedMetadata, RaycasterData)
class DepthCameraSensor(RaycasterSensor):
+
+ def build(self):
+ super().build()
+ self._shape = ((self._manager._sim._B,) if self._manager._sim.n_envs > 0 els... | genesis/sensors/raycaster/depth_camera.py | 26 | Python | 0.857 | performance | 225 | 40 | 38 | false | [BUG FIX] Fix DepthCamera sensor. | 1,842 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | Milotrince | Review this pull request change.
Language: Python
File: genesis/sensors/raycaster/depth_camera.py
PR Title: [BUG FIX] Fix DepthCamera sensor.
BEFORE CODE: from .patterns import DepthCameraPattern
from .raycaster import RaycasterData, RaycasterOptions, RaycasterSensor, RaycasterSharedMetadata
class DepthCameraOpti... | This is not readable. You should rather do:
```python
batch_shape = (self._manager._sim._B,) if self._manager._sim.n_envs > 0 else ()
self._shape = (*batch_shape, self._options.pattern.height, self._options.pattern.width)
``` | 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. |
# Update the differentiable contact data.
gjk_state.contact_pos[i_b, insert_id] = contact_pos
gjk_state.normal[i_b, insert_id] = contact_normal
gjk_state.diff_penetration[i_b, insert_id] = diff_penetration
# Update the non-differentiable contact data for the backward pass.
... | `gjk_state.is_col[i_b] = n_contacts > 0` | # Update the differentiable contact data.
gjk_state.contact_pos[i_b, insert_id] = contact_pos
gjk_state.normal[i_b, insert_id] = contact_normal
gjk_state.diff_penetration[i_b, insert_id] = diff_penetration
# Update the non-differentiable contact data for the backward pass.
... | @@ -342,7 +342,7 @@ def func_gjk_contact(
break
gjk_state.n_contacts[i_b] = n_contacts
- gjk_state.is_col[i_b] = 1 if n_contacts > 0 else 0
+ gjk_state.is_col[i_b] = True if n_contacts > 0 else False | genesis/engine/solvers/rigid/diff_gjk_decomp.py | 26 | Python | 0.214 | suggestion | 40 | 51 | 51 | false | [BUG FIX] Fixed minor bug in differentiable contact detection. | 1,864 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | SonSang | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/diff_gjk_decomp.py
PR Title: [BUG FIX] Fixed minor bug in differentiable contact detection.
BEFORE CODE: # Update the differentiable contact data.
gjk_state.contact_pos[i_b, insert_id] = contact_pos
gjk_stat... | `gjk_state.is_col[i_b] = n_contacts > 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. |
gjk_state.contact_pos[i_b, insert_id] = contact_pos
gjk_state.normal[i_b, insert_id] = contact_normal
gjk_state.diff_penetration[i_b, insert_id] = diff_penetration
# Update the non-differentiable contact data for the backward pass.
gjk_state.diff_contact_input.local_pos1_a[i_b, ... | This one should also be a boolean. | gjk_state.contact_pos[i_b, insert_id] = contact_pos
gjk_state.normal[i_b, insert_id] = contact_normal
gjk_state.diff_penetration[i_b, insert_id] = diff_penetration
# Update the non-differentiable contact data for the backward pass.
gjk_state.diff_contact_input.local_pos1_a[i_b, ... | @@ -342,7 +342,7 @@ def func_gjk_contact(
break
gjk_state.n_contacts[i_b] = n_contacts
- gjk_state.is_col[i_b] = 1 if n_contacts > 0 else 0
+ gjk_state.is_col[i_b] = True if n_contacts > 0 else False
gjk_state.multi_contact_flag[i_b] = 1 | genesis/engine/solvers/rigid/diff_gjk_decomp.py | 26 | Python | 0.143 | suggestion | 34 | 51 | 51 | false | [BUG FIX] Fixed minor bug in differentiable contact detection. | 1,864 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | SonSang | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/diff_gjk_decomp.py
PR Title: [BUG FIX] Fixed minor bug in differentiable contact detection.
BEFORE CODE: gjk_state.contact_pos[i_b, insert_id] = contact_pos
gjk_state.normal[i_b, insert_id] = contact_normal
... | This one should also be a boolean. | 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. |
E=3.0e4,
nu=0.45,
rho=1000.0,
model="stable_neohookean",
),
)
scene.build(n_envs=2)
mpm_cube_pos = mpm_cube.get_particles_pos()
cloth_pos = cloth.get_particles_pos()
water_pos = water.get_particles_pos()
mpm_cube_active = mpm_cube.get_par... | It is not checking much, just that the API is available. Not really useful. The optional arguments are not tested and the returned value is completely ignored. It would be nice to check that the value is correct, and if too complicated, at least checking that it makes sense (i.e. non-zero and in valid range). You shoul... | )
scene.build(n_envs=2)
init_mpm_cube_pos = mpm_cube.get_particles_pos()
init_cloth_pos = cloth.get_particles_pos()
init_water_pos = water.get_particles_pos()
scene.get_state()
for i in range(1500):
scene.step()
final_mpm_cube_pos = mpm_cube.get_particles_pos()
final_cloth... | @@ -191,6 +191,25 @@ def test_deformable_parallel(show_viewer):
)
scene.build(n_envs=2)
+ mpm_cube_pos = mpm_cube.get_particles_pos()
+ cloth_pos = cloth.get_particles_pos()
+ water_pos = water.get_particles_pos()
+
+ mpm_cube_active = mpm_cube.get_particles_active()
+ cloth_active = cloth.ge... | tests/test_deformable_physics.py | 26 | Python | 0.5 | suggestion | 476 | 51 | 51 | false | [MISC] Add get_particles_pos to sph | 1,871 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_deformable_physics.py
PR Title: [MISC] Add get_particles_pos to sph
BEFORE CODE: E=3.0e4,
nu=0.45,
rho=1000.0,
model="stable_neohookean",
),
)
scene.build(n_envs=2)
mpm_cube_pos = ... | It is not checking much, just that the API is available. Not really useful. The optional arguments are not tested and the returned value is completely ignored. It would be nice to check that the value is correct, and if too complicated, at least checking that it makes sense (i.e. non-zero and in valid range). You shoul... | 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. |
assert_allclose(state.vel, 0.0, tol=2e-2)
# Landed vertically
pos = tensor_to_array(state.pos.squeeze(0))
BV, *_ = igl.bounding_box(pos)
entity_center = 0.5 * (BV[0] + BV[-1])
if coupler_type == gs.options.SAPCouplerOptions:
tol = 1e-2 if material_model == "l... | hope we can reduce the time and mark them required in the near future | assert_allclose(state.vel, 0.0, tol=2e-2)
# Landed vertically
pos = tensor_to_array(state.pos.squeeze(0))
BV, *_ = igl.bounding_box(pos)
entity_center = 0.5 * (BV[0] + BV[-1])
if coupler_type == gs.options.SAPCouplerOptions:
tol = 1e-2 if material_model == "l... | @@ -255,7 +255,8 @@ def test_implicit_falling_sphere_box(coupler_type, material_model, show_viewer):
assert_allclose(-state.pos[..., 2].min(), penetration_depth_ref, tol=tol)
-@pytest.mark.required
+# This test cannot be flagged as required because it takes 250s to run on CPU. | tests/test_fem.py | 26 | Python | 0.214 | suggestion | 69 | 51 | 51 | false | [BUG FIX] Fix sensor recorder Matplotlib plotter on Linux and Windows. | 1,894 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | duburcqa | Review this pull request change.
Language: Python
File: tests/test_fem.py
PR Title: [BUG FIX] Fix sensor recorder Matplotlib plotter on Linux and Windows.
BEFORE CODE: assert_allclose(state.vel, 0.0, tol=2e-2)
# Landed vertically
pos = tensor_to_array(state.pos.squeeze(0))
BV, *_ = ... | hope we can reduce the time and mark them required in the near future | 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. |
mass_mat_L, mass_mat_D_inv = franka1.get_mass_mat(decompose=True)
mass_mat = mass_mat_L.T @ torch.diag(1.0 / mass_mat_D_inv) @ mass_mat_L
assert_allclose(mass_mat, mass_mat_1, tol=tol)
@pytest.mark.required
@pytest.mark.parametrize("model_name", ["hinge_slide"])
@pytest.mark.parametrize("gs_solver", [gs.c... | You must use our custom `assert_allclose` | mass_mat_L, mass_mat_D_inv = franka1.get_mass_mat(decompose=True)
mass_mat = mass_mat_L.T @ torch.diag(1.0 / mass_mat_D_inv) @ mass_mat_L
assert_allclose(mass_mat, mass_mat_1, tol=tol)
@pytest.mark.required
@pytest.mark.parametrize("model_name", ["hinge_slide"])
@pytest.mark.parametrize("gs_solver", [gs.c... | @@ -1692,12 +1692,16 @@ def test_set_dofs_frictionloss_physics(gs_sim, tol):
robot.set_dofs_velocity(initial_velocity)
robot.set_dofs_frictionloss(np.array([0.0, 0.0]))
+ frictionloss = robot.get_dofs_frictionloss()
+ np.testing.assert_allclose(frictionloss, np.array([0.0, 0.0]), atol=tol)
for _ ... | tests/test_rigid_physics.py | 26 | Python | 0.286 | suggestion | 41 | 51 | 51 | false | [FEATURE] Add Rigid ‘get_dofs_frictionloss’ public API method. | 1,904 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | ax-anoop | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [FEATURE] Add Rigid ‘get_dofs_frictionloss’ public API method.
BEFORE CODE: mass_mat_L, mass_mat_D_inv = franka1.get_mass_mat(decompose=True)
mass_mat = mass_mat_L.T @ torch.diag(1.0 / mass_mat_D_inv) @ mass_mat_L
... | You must use our custom `assert_allclose` | 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. |
(robot,) = gs_sim.entities
initial_velocity = np.array([1.0, 0.0])
robot.set_dofs_velocity(initial_velocity)
robot.set_dofs_frictionloss(np.array([0.0, 0.0]))
frictionloss = robot.get_dofs_frictionloss().cpu().numpy()
np.testing.assert_allclose(frictionloss, np.array([0.0, 0.0]), atol=tol)
... | No need to apply `.cpu().numpy()` if you use `assert_allclose` | (robot,) = gs_sim.entities
initial_velocity = np.array([1.0, 0.0])
robot.set_dofs_velocity(initial_velocity)
robot.set_dofs_frictionloss(np.array([0.0, 0.0]))
frictionloss = robot.get_dofs_frictionloss()
assert_allclose(frictionloss, np.array([0.0, 0.0]), atol=tol)
for _ in range(10):
... | @@ -1707,6 +1711,8 @@ def test_set_dofs_frictionloss_physics(gs_sim, tol):
robot.set_dofs_velocity(initial_velocity)
robot.set_dofs_frictionloss(np.array([0.5]), dofs_idx_local=[0])
+ frictionloss = robot.get_dofs_frictionloss(dofs_idx_local=[0]).cpu().numpy() | tests/test_rigid_physics.py | 26 | Python | 0.429 | suggestion | 62 | 51 | 51 | false | [FEATURE] Add Rigid ‘get_dofs_frictionloss’ public API method. | 1,904 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | ax-anoop | Review this pull request change.
Language: Python
File: tests/test_rigid_physics.py
PR Title: [FEATURE] Add Rigid ‘get_dofs_frictionloss’ public API method.
BEFORE CODE: (robot,) = gs_sim.entities
initial_velocity = np.array([1.0, 0.0])
robot.set_dofs_velocity(initial_velocity)
robot.set_dofs_fric... | No need to apply `.cpu().numpy()` if you use `assert_allclose` | 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. |
# Support field used for mpr and gjk. Rather than having separate support fields for each algorithm, keep only
# one copy here to save memory and maintain cleaner code.
self._support_field = SupportField(rigid_solver)
def _init_static_config(self) -> None:
# Identify the convex col... | when you have time, can you take a look @SonSang | # Support field used for mpr and gjk. Rather than having separate support fields for each algorithm, keep only
# one copy here to save memory and maintain cleaner code.
self._support_field = SupportField(rigid_solver)
def _init_static_config(self) -> None:
# Identify the convex coll... | @@ -86,7 +86,9 @@ def _init_static_config(self) -> None:
geom.is_convex or geom.type == gs.GEOM_TYPE.TERRAIN for geom in self._solver.geoms
),
has_terrain=any(geom.type == gs.GEOM_TYPE.TERRAIN for geom in self._solver.geoms),
- n_contacts_per_pair=20 if self._solver... | genesis/engine/solvers/rigid/collider_decomp.py | 26 | Python | 0.071 | suggestion | 49 | 51 | 51 | false | [MISC] Reduce memory usage due to diff constraint solver. | 1,930 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/collider_decomp.py
PR Title: [MISC] Reduce memory usage due to diff constraint solver.
BEFORE CODE:
# Support field used for mpr and gjk. Rather than having separate support fields for each algorithm, keep only
# o... | when you have time, can you take a look @SonSang | 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 dataclasses
from functools import partial
import gstaichi as ti
import numpy as np
import genesis as gs
if not gs._initialized:
gs.raise_exception("Genesis hasn't been initialized. Did you call `gs.init()`?")
V_ANNOTATION = ti.types.ndarray() if gs.use_ndarray else ti.template
V = ti.ndarra... | I'm not in favour of using cryptic helpers. It just makes everything harder to read. | import math
import dataclasses
from functools import partial
import gstaichi as ti
import numpy as np
import genesis as gs
if not gs._initialized:
gs.raise_exception("Genesis hasn't been initialized. Did you call `gs.init()`?")
V_ANNOTATION = ti.types.ndarray() if gs.use_ndarray else ti.template
V = ti.ndarra... | @@ -20,6 +20,10 @@
DATA_ORIENTED = partial(dataclasses.dataclass, frozen=True) if gs.use_ndarray else ti.data_oriented
+def _s(shape, is_on):
+ return shape if is_on else (1,)
+
+ | genesis/utils/array_class.py | 26 | Python | 0.214 | refactor | 84 | 51 | 51 | false | [MISC] Reduce memory usage due to diff constraint solver. | 1,930 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/utils/array_class.py
PR Title: [MISC] Reduce memory usage due to diff constraint solver.
BEFORE CODE: import math
import dataclasses
from functools import partial
import gstaichi as ti
import numpy as np
import genesis as gs
if not gs._initialized:... | I'm not in favour of using cryptic helpers. It just makes everything harder to read. | 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 dataclasses
from functools import partial
import gstaichi as ti
import numpy as np
import genesis as gs
if not gs._initialized:
gs.raise_exception("Genesis hasn't been initialized. Did you call `gs.init()`?")
V_ANNOTATION = ti.types.ndarray() if gs.use_ndarray else ti.template
V = ti.ndarra... | Besides, if you want to avoid memory allocation for unused variables, you should rather use shape `()` instead of `(1,)`. | import math
import dataclasses
from functools import partial
import gstaichi as ti
import numpy as np
import genesis as gs
if not gs._initialized:
gs.raise_exception("Genesis hasn't been initialized. Did you call `gs.init()`?")
V_ANNOTATION = ti.types.ndarray() if gs.use_ndarray else ti.template
V = ti.ndarra... | @@ -20,6 +20,10 @@
DATA_ORIENTED = partial(dataclasses.dataclass, frozen=True) if gs.use_ndarray else ti.data_oriented
+def _s(shape, is_on):
+ return shape if is_on else (1,) | genesis/utils/array_class.py | 24 | Python | 0.714 | suggestion | 121 | 49 | 49 | false | [MISC] Reduce memory usage due to diff constraint solver. | 1,930 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/utils/array_class.py
PR Title: [MISC] Reduce memory usage due to diff constraint solver.
BEFORE CODE: import math
import dataclasses
from functools import partial
import gstaichi as ti
import numpy as np
import genesis as gs
if not gs._initialized:... | Besides, if you want to avoid memory allocation for unused variables, you should rather use shape `()` instead of `(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. |
geom_a: V_ANNOTATION
geom_b: V_ANNOTATION
# Local positions of the 3 vertices from the two geometries that define the face on the Minkowski difference
local_pos1_a: V_ANNOTATION
local_pos1_b: V_ANNOTATION
local_pos1_c: V_ANNOTATION
local_pos2_a: V_ANNOTATION
local_pos2_b: V_ANNOTATION
... | Define `requires_grad = solver._requires_grad` to reduce line width. It would be more readable. | geom_a: V_ANNOTATION
geom_b: V_ANNOTATION
# Local positions of the 3 vertices from the two geometries that define the face on the Minkowski difference
local_pos1_a: V_ANNOTATION
local_pos1_b: V_ANNOTATION
local_pos1_c: V_ANNOTATION
local_pos2_a: V_ANNOTATION
local_pos2_b: V_ANNOTATION
... | @@ -356,19 +358,21 @@ def get_diff_contact_input(solver, max_contacts_per_pair):
_B = solver._B
return StructDiffContactInput(
- geom_a=V(dtype=gs.ti_int, shape=(_B, max_contacts_per_pair)),
- geom_b=V(dtype=gs.ti_int, shape=(_B, max_contacts_per_pair)),
- local_pos1_a=V_VEC(3, dtype=gs... | genesis/utils/array_class.py | 26 | Python | 0.357 | suggestion | 95 | 51 | 51 | false | [MISC] Reduce memory usage due to diff constraint solver. | 1,930 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/utils/array_class.py
PR Title: [MISC] Reduce memory usage due to diff constraint solver.
BEFORE CODE: geom_a: V_ANNOTATION
geom_b: V_ANNOTATION
# Local positions of the 3 vertices from the two geometries that define the face on the Minkowsk... | Define `requires_grad = solver._requires_grad` to reduce line width. It would be more readable. | 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 dataclasses
from functools import partial
import gstaichi as ti
import numpy as np
import genesis as gs
if not gs._initialized:
gs.raise_exception("Genesis hasn't been initialized. Did you call `gs.init()`?")
V_ANNOTATION = ti.types.ndarray() if gs.use_ndarray else ti.template
V = ti.ndarra... | I would be better to have a more explicit name if you want to keep this pattern. Like `optional_shape` or `cond_shape` or `maybe_shape`. | import math
import dataclasses
from functools import partial
import gstaichi as ti
import numpy as np
import genesis as gs
if not gs._initialized:
gs.raise_exception("Genesis hasn't been initialized. Did you call `gs.init()`?")
V_ANNOTATION = ti.types.ndarray() if gs.use_ndarray else ti.template
V = ti.ndarra... | @@ -20,6 +20,10 @@
DATA_ORIENTED = partial(dataclasses.dataclass, frozen=True) if gs.use_ndarray else ti.data_oriented
+def _shape(shape, is_on): | genesis/utils/array_class.py | 23 | Python | 0.571 | suggestion | 136 | 48 | 48 | false | [MISC] Reduce memory usage due to diff constraint solver. | 1,930 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/utils/array_class.py
PR Title: [MISC] Reduce memory usage due to diff constraint solver.
BEFORE CODE: import math
import dataclasses
from functools import partial
import gstaichi as ti
import numpy as np
import genesis as gs
if not gs._initialized:... | I would be better to have a more explicit name if you want to keep this pattern. Like `optional_shape` or `cond_shape` or `maybe_shape`. | 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 shape_arg in (*shape_args, shape_args[0]):
if tuple(shape_arg) == (3, 3):
np_arg = gu.rotvec_to_R(np.random.randn(*batch_shape, 3).clip(-1.0, 1.0).astype(gs.np_float))
else:
np_arg = np.random.randn(*batch_shape, *shape_arg).clip(-1.0, 1.0).astype(gs.n... | Do not check this. It is not part of our spec. | for shape_arg in (*shape_args, shape_args[0]):
if tuple(shape_arg) == (3, 3):
np_arg = gu.rotvec_to_R(np.random.randn(*batch_shape, 3).clip(-1.0, 1.0).astype(gs.np_float))
else:
np_arg = np.random.randn(*batch_shape, *shape_arg).clip(-1.0, 1.0).astype(gs.n... | @@ -292,3 +292,145 @@ def test_utils_geom_tensor_identity(batch_shape):
np.testing.assert_allclose(np_args[0], np_args[-1], atol=1e2 * gs.EPS)
np.testing.assert_allclose(tensor_to_array(tc_args[0]), tensor_to_array(tc_args[-1]), atol=1e2 * gs.EPS)
+
+
+@pytest.mark.required
+def test_pyrender_vec3()... | tests/test_utils.py | 26 | Python | 0.071 | suggestion | 46 | 51 | 51 | false | [BUG FIX] Fix interactive viewer 'Vec3'. | 1,931 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_utils.py
PR Title: [BUG FIX] Fix interactive viewer 'Vec3'.
BEFORE CODE: for shape_arg in (*shape_args, shape_args[0]):
if tuple(shape_arg) == (3, 3):
np_arg = gu.rotvec_to_R(np.random.randn(*batch_shape, 3).clip(... | Do not check this. It is not part of our spec. | 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. |
np_arg = np.random.randn(*batch_shape, *shape_arg).clip(-1.0, 1.0).astype(gs.np_float)
tc_arg = torch.as_tensor(np_arg, dtype=gs.tc_float, device=gs.device)
np_args.append(np_arg)
tc_args.append(tc_arg)
num_funcs = len(py_funcs)
for i, py_func in enum... | Same: `assert_allclose((v.x, v.y, v.z), (1.0, 2.0, 3.0), tol=gs.EPS)` | np_arg = np.random.randn(*batch_shape, *shape_arg).clip(-1.0, 1.0).astype(gs.np_float)
tc_arg = torch.as_tensor(np_arg, dtype=gs.tc_float, device=gs.device)
np_args.append(np_arg)
tc_args.append(tc_arg)
num_funcs = len(py_funcs)
for i, py_func in enum... | @@ -292,3 +292,145 @@ def test_utils_geom_tensor_identity(batch_shape):
np.testing.assert_allclose(np_args[0], np_args[-1], atol=1e2 * gs.EPS)
np.testing.assert_allclose(tensor_to_array(tc_args[0]), tensor_to_array(tc_args[-1]), atol=1e2 * gs.EPS)
+
+
+@pytest.mark.required
+def test_pyrender_vec3()... | tests/test_utils.py | 26 | Python | 0.357 | suggestion | 69 | 51 | 51 | false | [BUG FIX] Fix interactive viewer 'Vec3'. | 1,931 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_utils.py
PR Title: [BUG FIX] Fix interactive viewer 'Vec3'.
BEFORE CODE: np_arg = np.random.randn(*batch_shape, *shape_arg).clip(-1.0, 1.0).astype(gs.np_float)
tc_arg = torch.as_tensor(np_arg, dtype=gs.tc_float, device=gs... | Same: `assert_allclose((v.x, v.y, v.z), (1.0, 2.0, 3.0), tol=gs.EPS)` | 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. |
num_funcs = len(py_funcs)
for i, py_func in enumerate(py_funcs):
np_args[i + 1][:] = py_func(np_args[i])
tc_args[i + 1][:] = py_func(tc_args[i])
np.testing.assert_allclose(np_args[0], np_args[-1], atol=1e2 * gs.EPS)
np.testing.assert_allclose(tensor_to_array(tc_... | Do not assert dtype, we don't care. |
num_funcs = len(py_funcs)
for i, py_func in enumerate(py_funcs):
np_args[i + 1][:] = py_func(np_args[i])
tc_args[i + 1][:] = py_func(tc_args[i])
np.testing.assert_allclose(np_args[0], np_args[-1], atol=1e2 * gs.EPS)
np.testing.assert_allclose(tensor_to_array(tc_... | @@ -292,3 +292,145 @@ def test_utils_geom_tensor_identity(batch_shape):
np.testing.assert_allclose(np_args[0], np_args[-1], atol=1e2 * gs.EPS)
np.testing.assert_allclose(tensor_to_array(tc_args[0]), tensor_to_array(tc_args[-1]), atol=1e2 * gs.EPS)
+
+
+@pytest.mark.required
+def test_pyrender_vec3()... | tests/test_utils.py | 26 | Python | 0.214 | suggestion | 35 | 51 | 51 | false | [BUG FIX] Fix interactive viewer 'Vec3'. | 1,931 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_utils.py
PR Title: [BUG FIX] Fix interactive viewer 'Vec3'.
BEFORE CODE:
num_funcs = len(py_funcs)
for i, py_func in enumerate(py_funcs):
np_args[i + 1][:] = py_func(np_args[i])
tc_args[i + 1][:] = py_func(tc... | Do not assert dtype, we don't care. | 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. |
# construction helpers enforce shape and dtype
v = Vec3.from_xyz(1.0, 2.0, 3.0)
assert v.v.shape == (3,)
assert v.v.dtype == np.float32
assert np.allclose(v.v, np.array([1.0, 2.0, 3.0], dtype=np.float32))
assert v.x == pytest.approx(1.0)
assert v.y == pytest.approx(2.0)
assert v.z == pyt... | `assert_allclose(Vec3.full(5.5).v, 5.5, tol=gs.EPS)` | tol = 1e-6
# construction helpers enforce shape and dtype
v = Vec3.from_xyz(1.0, 2.0, 3.0)
assert v.v.shape == (3,)
assert_allclose(v.v, np.array([1.0, 2.0, 3.0]), tol=gs.EPS)
assert_allclose((v.x, v.y, v.z), (1.0, 2.0, 3.0), tol=gs.EPS)
# from_array converts various dtypes to float32
v... | @@ -292,3 +292,145 @@ def test_utils_geom_tensor_identity(batch_shape):
np.testing.assert_allclose(np_args[0], np_args[-1], atol=1e2 * gs.EPS)
np.testing.assert_allclose(tensor_to_array(tc_args[0]), tensor_to_array(tc_args[-1]), atol=1e2 * gs.EPS)
+
+
+@pytest.mark.required
+def test_pyrender_vec3()... | tests/test_utils.py | 26 | Python | 0.357 | suggestion | 52 | 51 | 51 | false | [BUG FIX] Fix interactive viewer 'Vec3'. | 1,931 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_utils.py
PR Title: [BUG FIX] Fix interactive viewer 'Vec3'.
BEFORE CODE: # construction helpers enforce shape and dtype
v = Vec3.from_xyz(1.0, 2.0, 3.0)
assert v.v.shape == (3,)
assert v.v.dtype == np.float32
assert np.allclose(v... | `assert_allclose(Vec3.full(5.5).v, 5.5, tol=gs.EPS)` | 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. |
# construction helpers enforce shape and dtype
v = Vec3.from_xyz(1.0, 2.0, 3.0)
assert v.v.shape == (3,)
assert v.v.dtype == np.float32
assert np.allclose(v.v, np.array([1.0, 2.0, 3.0], dtype=np.float32))
assert v.x == pytest.approx(1.0)
assert v.y == pytest.approx(2.0)
assert v.z == py... | `assert_allclose(Vec3.zero().v, 0.0, tol=gs.EPS)`
`assert_allclose(Vec3.one().v, 1.0, tol=gs.EPS)` |
tol = 1e-6
# construction helpers enforce shape and dtype
v = Vec3.from_xyz(1.0, 2.0, 3.0)
assert v.v.shape == (3,)
assert_allclose(v.v, np.array([1.0, 2.0, 3.0]), tol=gs.EPS)
assert_allclose((v.x, v.y, v.z), (1.0, 2.0, 3.0), tol=gs.EPS)
# from_array converts various dtypes to float32
... | @@ -292,3 +292,145 @@ def test_utils_geom_tensor_identity(batch_shape):
np.testing.assert_allclose(np_args[0], np_args[-1], atol=1e2 * gs.EPS)
np.testing.assert_allclose(tensor_to_array(tc_args[0]), tensor_to_array(tc_args[-1]), atol=1e2 * gs.EPS)
+
+
+@pytest.mark.required
+def test_pyrender_vec3()... | tests/test_utils.py | 26 | Python | 0.357 | suggestion | 98 | 51 | 51 | false | [BUG FIX] Fix interactive viewer 'Vec3'. | 1,931 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_utils.py
PR Title: [BUG FIX] Fix interactive viewer 'Vec3'.
BEFORE CODE:
# construction helpers enforce shape and dtype
v = Vec3.from_xyz(1.0, 2.0, 3.0)
assert v.v.shape == (3,)
assert v.v.dtype == np.float32
assert np.allclose(... | `assert_allclose(Vec3.zero().v, 0.0, tol=gs.EPS)`
`assert_allclose(Vec3.one().v, 1.0, tol=gs.EPS)` | 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.ext.pyrender.interaction.vec3 import Vec3, Quat
# construction helpers enforce shape and dtype
v = Vec3.from_xyz(1.0, 2.0, 3.0)
assert v.v.shape == (3,)
assert v.v.dtype == np.float32
assert np.allclose(v.v, np.array([1.0, 2.0, 3.0], dtype=np.float32))
assert v.x == pytest.appr... | Do not check dtype preservation. | from genesis.ext.pyrender.interaction.vec3 import Vec3, Quat
tol = 1e-6
# construction helpers enforce shape and dtype
v = Vec3.from_xyz(1.0, 2.0, 3.0)
assert v.v.shape == (3,)
assert_allclose(v.v, np.array([1.0, 2.0, 3.0]), tol=gs.EPS)
assert_allclose((v.x, v.y, v.z), (1.0, 2.0, 3.0), tol=... | @@ -292,3 +292,145 @@ def test_utils_geom_tensor_identity(batch_shape):
np.testing.assert_allclose(np_args[0], np_args[-1], atol=1e2 * gs.EPS)
np.testing.assert_allclose(tensor_to_array(tc_args[0]), tensor_to_array(tc_args[-1]), atol=1e2 * gs.EPS)
+
+
+@pytest.mark.required
+def test_pyrender_vec3()... | tests/test_utils.py | 26 | Python | 0.214 | suggestion | 32 | 51 | 51 | false | [BUG FIX] Fix interactive viewer 'Vec3'. | 1,931 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_utils.py
PR Title: [BUG FIX] Fix interactive viewer 'Vec3'.
BEFORE CODE: from genesis.ext.pyrender.interaction.vec3 import Vec3, Quat
# construction helpers enforce shape and dtype
v = Vec3.from_xyz(1.0, 2.0, 3.0)
assert v.v.shape =... | Do not check dtype preservation. | 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. |
v_t = Vec3.from_tensor(torch.tensor([4.0, 5.0, 6.0], dtype=torch.float32))
assert np.allclose(v_t.v, [4.0, 5.0, 6.0])
# constants
assert np.allclose(Vec3.zero().v, np.zeros(3, dtype=np.float32))
assert np.allclose(Vec3.one().v, np.ones(3, dtype=np.float32))
assert np.allclose(Vec3.full(5.5).v, ... | This is not part of our spec. Or is it? It should be clarified. | assert_allclose(Vec3.zero().v, 0.0, tol=gs.EPS)
assert_allclose(Vec3.one().v, 1.0, tol=gs.EPS)
assert_allclose(Vec3.full(5.5).v, 5.5, tol=gs.EPS)
# arithmetic ops and dtype preservation
a = Vec3.from_xyz(1, 2, 3)
b = Vec3.from_xyz(4, 5, 6)
c = a + b
d = b - a
assert_allclose(c.v, np... | @@ -292,3 +292,145 @@ def test_utils_geom_tensor_identity(batch_shape):
np.testing.assert_allclose(np_args[0], np_args[-1], atol=1e2 * gs.EPS)
np.testing.assert_allclose(tensor_to_array(tc_args[0]), tensor_to_array(tc_args[-1]), atol=1e2 * gs.EPS)
+
+
+@pytest.mark.required
+def test_pyrender_vec3()... | tests/test_utils.py | 26 | Python | 0.286 | suggestion | 63 | 51 | 51 | false | [BUG FIX] Fix interactive viewer 'Vec3'. | 1,931 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_utils.py
PR Title: [BUG FIX] Fix interactive viewer 'Vec3'.
BEFORE CODE: v_t = Vec3.from_tensor(torch.tensor([4.0, 5.0, 6.0], dtype=torch.float32))
assert np.allclose(v_t.v, [4.0, 5.0, 6.0])
# constants
assert np.allclose(Vec3.zero(... | This is not part of our spec. Or is it? It should be clarified. | 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. |
assert np.allclose(m2.v, [2, 4, 6])
assert m1.v.dtype == np.float32 and m2.v.dtype == np.float32
assert np.allclose((-a).v, [-1, -2, -3])
# dot and cross
dot_ab = a.dot(b)
assert isinstance(dot_ab, float)
assert dot_ab == pytest.approx(1 * 4 + 2 * 5 + 3 * 6)
cross_ab = a.cross(b)
a... | As everywhere else, use our helpers, and do not checks stuffs that are not part of your spec.
`assert_allclose(cp.v, 0.0, tol=gs.EPS)` | assert_allclose(dot_ab, 1 * 4 + 2 * 5 + 3 * 6, tol=gs.EPS)
cross_ab = a.cross(b)
assert_allclose(cross_ab.v, np.array([-3.0, 6.0, -3.0]), tol=gs.EPS)
# norms
assert_allclose(a.sqr_magnitude(), 1.0 + 4.0 + 9.0, tol=gs.EPS)
assert_allclose(a.magnitude(), np.sqrt(a.sqr_magnitude()), tol=gs.EPS)
... | @@ -292,3 +292,145 @@ def test_utils_geom_tensor_identity(batch_shape):
np.testing.assert_allclose(np_args[0], np_args[-1], atol=1e2 * gs.EPS)
np.testing.assert_allclose(tensor_to_array(tc_args[0]), tensor_to_array(tc_args[-1]), atol=1e2 * gs.EPS)
+
+
+@pytest.mark.required
+def test_pyrender_vec3()... | tests/test_utils.py | 26 | Python | 0.571 | refactor | 134 | 51 | 51 | false | [BUG FIX] Fix interactive viewer 'Vec3'. | 1,931 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_utils.py
PR Title: [BUG FIX] Fix interactive viewer 'Vec3'.
BEFORE CODE: assert np.allclose(m2.v, [2, 4, 6])
assert m1.v.dtype == np.float32 and m2.v.dtype == np.float32
assert np.allclose((-a).v, [-1, -2, -3])
# dot and cross
d... | As everywhere else, use our helpers, and do not checks stuffs that are not part of your spec.
`assert_allclose(cp.v, 0.0, tol=gs.EPS)` | 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. |
assert a.sqr_magnitude() == pytest.approx(1.0 + 4.0 + 9.0)
assert a.magnitude() == pytest.approx(np.sqrt(a.sqr_magnitude()))
na = a.normalized()
assert na.v.dtype == np.float32
assert na.magnitude() == pytest.approx(1.0, rel=1e-6, abs=1e-6)
assert np.allclose(Vec3.zero().normalized().v, [0.0, 0.... | This restriction is useless. I would not assert it. It would be better to do `self.v = np.asarray(v, dtype=np.float32)` | cp = a.copy()
assert cp is not a
cp.v[...] = 0.0
assert_allclose(a.v, np.array([1.0, 2.0, 3.0]), tol=gs.EPS)
assert_allclose(cp.v, 0.0, tol=gs.EPS)
# repr and tensor conversion
t = a.as_tensor()
assert isinstance(t, torch.Tensor)
assert_allclose(t.cpu().numpy(), a.v, tol=gs.EPS)
... | @@ -292,3 +292,145 @@ def test_utils_geom_tensor_identity(batch_shape):
np.testing.assert_allclose(np_args[0], np_args[-1], atol=1e2 * gs.EPS)
np.testing.assert_allclose(tensor_to_array(tc_args[0]), tensor_to_array(tc_args[-1]), atol=1e2 * gs.EPS)
+
+
+@pytest.mark.required
+def test_pyrender_vec3()... | tests/test_utils.py | 26 | Python | 0.714 | suggestion | 119 | 51 | 51 | false | [BUG FIX] Fix interactive viewer 'Vec3'. | 1,931 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_utils.py
PR Title: [BUG FIX] Fix interactive viewer 'Vec3'.
BEFORE CODE: assert a.sqr_magnitude() == pytest.approx(1.0 + 4.0 + 9.0)
assert a.magnitude() == pytest.approx(np.sqrt(a.sqr_magnitude()))
na = a.normalized()
assert na.v.dty... | This restriction is useless. I would not assert it. It would be better to do `self.v = np.asarray(v, dtype=np.float32)` | 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. |
assert cp is not a
cp.v[...] = 0.0
assert np.allclose(a.v, [1.0, 2.0, 3.0])
assert np.allclose(cp.v, [0.0, 0.0, 0.0])
# repr and tensor conversion
assert "Vec3(" in repr(a)
t = a.as_tensor()
assert isinstance(t, torch.Tensor)
assert t.dtype == torch.float32
assert np.allclose(t.... | `assert_allclose((q.w, q.x, q.y, q.z), (1, 0, 0, 0))` | assert_allclose(t.cpu().numpy(), a.v, tol=gs.EPS)
# --- Quat tests ---
q = Quat.from_wxyz(1.0, 0.0, 0.0, 0.0) # identity
assert q.v.shape == (4,)
assert_allclose(np.array([q.w, q.x, q.y, q.z]), np.array([1.0, 0.0, 0.0, 0.0]), tol=gs.EPS)
# from_array converts dtype and enforces shape
q_ar... | @@ -292,3 +292,145 @@ def test_utils_geom_tensor_identity(batch_shape):
np.testing.assert_allclose(np_args[0], np_args[-1], atol=1e2 * gs.EPS)
np.testing.assert_allclose(tensor_to_array(tc_args[0]), tensor_to_array(tc_args[-1]), atol=1e2 * gs.EPS)
+
+
+@pytest.mark.required
+def test_pyrender_vec3()... | tests/test_utils.py | 26 | Python | 0.357 | suggestion | 53 | 51 | 47 | false | [BUG FIX] Fix interactive viewer 'Vec3'. | 1,931 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_utils.py
PR Title: [BUG FIX] Fix interactive viewer 'Vec3'.
BEFORE CODE: assert cp is not a
cp.v[...] = 0.0
assert np.allclose(a.v, [1.0, 2.0, 3.0])
assert np.allclose(cp.v, [0.0, 0.0, 0.0])
# repr and tensor conversion
asse... | `assert_allclose((q.w, q.x, q.y, q.z), (1, 0, 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. |
for _ in range(50):
scene.step()
# Check that the sphere did not move, and the slightly squished
state = sphere.get_state()
center = state.pos.mean(axis=(0, 1))
assert_allclose(center[:2], 0.0, tol=0.01)
assert center[2] < SPHERE_RADIUS - 0.02
# Check that the ant is laying on top ... | let's see how long will this test run. | # Run the simulation
for _ in range(50):
scene.step()
# Check that the sphere did not move, and the slightly squished
state = sphere.get_state()
center = state.pos.mean(axis=(0, 1))
assert_allclose(center[:2], 0.0, tol=0.01)
assert center[2] < SPHERE_RADIUS - 0.02
# Check that ... | @@ -278,3 +278,74 @@ def test_sap_fem_vs_robot(show_viewer):
# Check that the legs of the ants are resting on the sphere
assert_allclose(robot.get_qpos()[-4:].abs(), 1.0, tol=0.1)
+
+
+@pytest.mark.required
+@pytest.mark.parametrize("substeps", [1, 10])
+def test_rigid_mpm_legacy_coupling(substeps, show_vie... | tests/test_hybrid.py | 26 | Python | 0.071 | suggestion | 38 | 51 | 51 | false | [CHANGING][BUG FIX] Fix external coupling forces from other solvers not affecting rigid bodies. | 1,941 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | SonSang | Review this pull request change.
Language: Python
File: tests/test_hybrid.py
PR Title: [CHANGING][BUG FIX] Fix external coupling forces from other solvers not affecting rigid bodies.
BEFORE CODE: for _ in range(50):
scene.step()
# Check that the sphere did not move, and the slightly squished
st... | let's see how long will this test run. | 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. |
color=(1.0, 0.4, 0.4),
vis_mode="visual",
),
)
obj_sand = scene.add_entity(
material=gs.materials.MPM.Liquid(),
morph=gs.morphs.Box(
pos=(0.0, 0.0, 0.25),
size=(0.3, 0.3, 0.3),
),
surface=gs.surfaces.Default(
co... | why is this target_pos the expected pos? Does this mean the sand push the objects in the -y direction? | color=(1.0, 0.4, 0.4),
vis_mode="visual",
),
)
obj_sand = scene.add_entity(
material=gs.materials.MPM.Liquid(),
morph=gs.morphs.Box(
pos=(0.0, 0.0, 0.2),
size=(0.3, 0.3, 0.3),
),
surface=gs.surfaces.Default(
col... | @@ -278,3 +278,74 @@ def test_sap_fem_vs_robot(show_viewer):
# Check that the legs of the ants are resting on the sphere
assert_allclose(robot.get_qpos()[-4:].abs(), 1.0, tol=0.1)
+
+
+@pytest.mark.required
+@pytest.mark.parametrize("substeps", [1, 10])
+def test_rigid_mpm_legacy_coupling(substeps, show_vie... | tests/test_hybrid.py | 26 | Python | 0.357 | question | 102 | 29 | 27 | false | [CHANGING][BUG FIX] Fix external coupling forces from other solvers not affecting rigid bodies. | 1,941 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | SonSang | Review this pull request change.
Language: Python
File: tests/test_hybrid.py
PR Title: [CHANGING][BUG FIX] Fix external coupling forces from other solvers not affecting rigid bodies.
BEFORE CODE: color=(1.0, 0.4, 0.4),
vis_mode="visual",
),
)
obj_sand = scene.add_entity(
... | why is this target_pos the expected pos? Does this mean the sand push the objects in the -y 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. |
geoms_state=geoms_state,
rigid_global_info=rigid_global_info,
static_rigid_sim_config=static_rigid_sim_config,
contact_island_state=contact_island_state,
)
func_update_acc(
update_cacc=False,
dofs_state=dofs_state,
links_info=links_info,
links_stat... | Do not add dedicated clear ingredients kernel. This will significantly affect simulation performance. Just add the clearing code in `func_update_force` kernel, at the very end. | entities_info=entities_info,
dofs_state=dofs_state,
dofs_info=dofs_info,
links_state=links_state,
links_info=links_info,
joints_info=joints_info,
geoms_state=geoms_state,
rigid_global_info=rigid_global_info,
static_rigid_sim_config=static_rigid_sim... | @@ -3810,6 +3810,11 @@ def func_forward_dynamics(
rigid_global_info=rigid_global_info,
static_rigid_sim_config=static_rigid_sim_config,
)
+ func_clear_coupling_force(
+ links_state=links_state,
+ rigid_global_info=rigid_global_info,
+ static_rigid_sim_config=static_rigid_s... | genesis/engine/solvers/rigid/rigid_solver_decomp.py | 26 | Python | 0.643 | performance | 176 | 51 | 51 | false | [CHANGING][BUG FIX] Fix external coupling forces from other solvers not affecting rigid bodies. | 1,941 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | SonSang | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/rigid_solver_decomp.py
PR Title: [CHANGING][BUG FIX] Fix external coupling forces from other solvers not affecting rigid bodies.
BEFORE CODE: geoms_state=geoms_state,
rigid_global_info=rigid_global_info,
sta... | Do not add dedicated clear ingredients kernel. This will significantly affect simulation performance. Just add the clearing code in `func_update_force` kernel, at the very end. | 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.step()
# Check that the sphere did not move, and the slightly squished
state = sphere.get_state()
center = state.pos.mean(axis=(0, 1))
assert_allclose(center[:2], 0.0, tol=0.01)
assert center[2] < SPHERE_RADIUS - 0.02
# Check that the ant is laying on top of the sphere
robot_... | If it is only used once, inline the value on the assert itself. This will improve readability. | for _ in range(50):
scene.step()
# Check that the sphere did not move, and the slightly squished
state = sphere.get_state()
center = state.pos.mean(axis=(0, 1))
assert_allclose(center[:2], 0.0, tol=0.01)
assert center[2] < SPHERE_RADIUS - 0.02
# Check that the ant is laying on top ... | @@ -278,3 +278,74 @@ def test_sap_fem_vs_robot(show_viewer):
# Check that the legs of the ants are resting on the sphere
assert_allclose(robot.get_qpos()[-4:].abs(), 1.0, tol=0.1)
+
+
+@pytest.mark.required
+@pytest.mark.parametrize("substeps", [1, 10])
+def test_rigid_mpm_legacy_coupling(substeps, show_vie... | tests/test_hybrid.py | 26 | Python | 0.286 | suggestion | 94 | 51 | 51 | false | [CHANGING][BUG FIX] Fix external coupling forces from other solvers not affecting rigid bodies. | 1,941 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | SonSang | Review this pull request change.
Language: Python
File: tests/test_hybrid.py
PR Title: [CHANGING][BUG FIX] Fix external coupling forces from other solvers not affecting rigid bodies.
BEFORE CODE: scene.step()
# Check that the sphere did not move, and the slightly squished
state = sphere.get_state()... | If it is only used once, inline the value on the assert itself. This will improve 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. |
CAPSULE = 4
BOX = 5
MESH = 6
TERRAIN = 7
# joint type in rigid solver, ranked by number of dofs
class JOINT_TYPE(IntEnum):
FIXED = 0
REVOLUTE = 1
PRISMATIC = 2
SPHERICAL = 3
FREE = 4
class EQUALITY_TYPE(IntEnum):
CONNECT = 0
WELD = 1
JOINT = 2
class CTRL_MODE(IntEnu... | I don't think internally you need this extra mode. I would just remove `POSITION` and use `POSITION_VELOCITY` everywhere instead. | CAPSULE = 4
BOX = 5
MESH = 6
TERRAIN = 7
# joint type in rigid solver, ranked by number of dofs
class JOINT_TYPE(IntEnum):
FIXED = 0
REVOLUTE = 1
PRISMATIC = 2
SPHERICAL = 3
FREE = 4
class EQUALITY_TYPE(IntEnum):
CONNECT = 0
WELD = 1
JOINT = 2
class CTRL_MODE(IntEnu... | @@ -47,6 +47,7 @@ class CTRL_MODE(IntEnum):
FORCE = 0
VELOCITY = 1
POSITION = 2
+ POSITION_VELOCITY = 3 | genesis/constants.py | 26 | Python | 0.571 | suggestion | 129 | 51 | 51 | false | [FEATURE] Implement position-velocity controller. | 1,948 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | matthieuvigne | Review this pull request change.
Language: Python
File: genesis/constants.py
PR Title: [FEATURE] Implement position-velocity controller.
BEFORE CODE: CAPSULE = 4
BOX = 5
MESH = 6
TERRAIN = 7
# joint type in rigid solver, ranked by number of dofs
class JOINT_TYPE(IntEnum):
FIXED = 0
REVOLUT... | I don't think internally you need this extra mode. I would just remove `POSITION` and use `POSITION_VELOCITY` everywhere 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. |
velocity = velocity.unsqueeze(0)
has_gains = kernel_control_dofs_position_velocity(
position, velocity, dofs_idx, envs_idx, self.dofs_state, self.dofs_info, self._static_rigid_sim_config
)
if not unsafe and not has_gains:
raise gs.raise_exception(
... | `kernel_control_dofs_velocity` should be modified to only consider `kv` and ignore `kp` | dofs_info=self.dofs_info,
geoms_state=self.geoms_state,
geoms_info=self.geoms_info,
entities_info=self.entities_info,
rigid_global_info=self._rigid_global_info,
static_rigid_sim_config=self._static_rigid_sim_config,
... | @@ -1824,7 +1824,7 @@ def control_dofs_velocity(self, velocity, dofs_idx=None, envs_idx=None, *, unsaf
)
if not unsafe and not has_gains: | genesis/engine/solvers/rigid/rigid_solver_decomp.py | 26 | Python | 0.429 | suggestion | 87 | 51 | 51 | false | [FEATURE] Implement position-velocity controller. | 1,948 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | matthieuvigne | Review this pull request change.
Language: Python
File: genesis/engine/solvers/rigid/rigid_solver_decomp.py
PR Title: [FEATURE] Implement position-velocity controller.
BEFORE CODE: velocity = velocity.unsqueeze(0)
has_gains = kernel_control_dofs_position_velocity(
position, velocity,... | `kernel_control_dofs_velocity` should be modified to only consider `kv` and ignore `kp` | 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. |
wave_width = (columns - logger.INFO_length - 11) // 2
if wave_width % 2 == 0:
wave_width -= 1
wave_width = max(0, min(38, wave_width))
bar_width = wave_width * 2 + 9
wave = ("┈┉" * wave_width)[:wave_width]
logger.info(f"~<╭{'─'*(bar_width)}╮>~")
logger.info(f"~<│{wave}>~ ~~~~<Genesis... | if env var has GS_ENABLE_NDARRAY, `is_ndarray_disabled=True`? | wave_width = (columns - logger.INFO_length - 11) // 2
if wave_width % 2 == 0:
wave_width -= 1
wave_width = max(0, min(38, wave_width))
bar_width = wave_width * 2 + 9
wave = ("┈┉" * wave_width)[:wave_width]
logger.info(f"~<╭{'─'*(bar_width)}╮>~")
logger.info(f"~<│{wave}>~ ~~~~<Genesis... | @@ -118,8 +118,7 @@ def init(
# Configure GsTaichi fast cache and array type
global use_ndarray, use_fastcache
- # is_ndarray_disabled = (os.environ.get("GS_ENABLE_NDARRAY") or ("0" if sys.platform == "darwin" else "1")) == "0"
- is_ndarray_disabled = os.environ.get("GS_ENABLE_NDARRAY", "0") == "0"
+ ... | genesis/__init__.py | 26 | Python | 0.357 | question | 62 | 51 | 51 | false | [MISC] Enable back GsTaichi dynamic array mode by default except for MacOS. | 1,977 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | duburcqa | Review this pull request change.
Language: Python
File: genesis/__init__.py
PR Title: [MISC] Enable back GsTaichi dynamic array mode by default except for MacOS.
BEFORE CODE: wave_width = (columns - logger.INFO_length - 11) // 2
if wave_width % 2 == 0:
wave_width -= 1
wave_width = max(0, min(38,... | if env var has GS_ENABLE_NDARRAY, `is_ndarray_disabled=True`? | 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 collections import defaultdict
import csv
import subprocess
import time
import argparse
def grep(contents: list[str], target):
return [l for l in contents if target in l]
def get_cuda_usage() -> dict[int, int]:
output = subprocess.check_output(["nvidia-smi"]).decode("utf-8")
section = 0
subsec ... | Interesting. Is this the best practice? Can we use pynvml? (asked ChatGPT. I did not try myself) | from collections import defaultdict
import csv
import subprocess
import time
import os
import argparse
import psutil
def grep(contents: list[str], target):
return [l for l in contents if target in l]
def get_cuda_usage() -> dict[int, int]:
output = subprocess.check_output(["nvidia-smi"]).decode("utf-8")
... | @@ -0,0 +1,94 @@
+from collections import defaultdict
+import csv
+import subprocess
+import time
+import argparse
+
+
+def grep(contents: list[str], target):
+ return [l for l in contents if target in l]
+
+
+def get_cuda_usage() -> dict[int, int]:
+ output = subprocess.check_output(["nvidia-smi"]).decode("utf-8... | tests/monitor_test_mem.py | 19 | Python | 0.286 | suggestion | 96 | 44 | 44 | false | [MISC] Add benchmark test memory monitoring. | 1,981 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | hughperkins | Review this pull request change.
Language: Python
File: tests/monitor_test_mem.py
PR Title: [MISC] Add benchmark test memory monitoring.
BEFORE CODE: from collections import defaultdict
import csv
import subprocess
import time
import argparse
def grep(contents: list[str], target):
return [l for l in contents ... | Interesting. Is this the best practice? Can we use pynvml? (asked ChatGPT. I did not try myself) | 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. |
6D rotation representation, of size (*, 6)
[1] http://arxiv.org/abs/1812.07035
"""
if isinstance(R, torch.Tensor):
return R[..., :2, :].flatten(start_dim=-2)
elif isinstance(R, np.ndarray):
return R[..., :2, :].reshape((*R.shape[:-2], 6))
else:
gs.raise_exception(f"t... | Can you explain why this change is necessary? | 6D rotation representation, of size (*, 6)
[1] http://arxiv.org/abs/1812.07035
"""
if isinstance(R, torch.Tensor):
return R[..., :2, :].flatten(start_dim=-2)
elif isinstance(R, np.ndarray):
return R[..., :2, :].reshape((*R.shape[:-2], 6))
else:
gs.raise_exception(f"t... | @@ -500,9 +500,10 @@ def _np_xyz_to_quat(xyz: np.ndarray, rpy: bool = False, out: np.ndarray | None =
"""
assert xyz.ndim >= 1
if out is None:
- out_ = np.empty((*xyz.shape[:-1], 4), dtype=xyz.dtype)
+ out_ = np.empty(xyz.shape[:-1] + (4,), dtype=xyz.dtype) | genesis/utils/geom.py | 26 | Python | 0.143 | question | 45 | 51 | 51 | false | [MISC] Add benchmark test memory monitoring. | 1,981 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | hughperkins | Review this pull request change.
Language: Python
File: genesis/utils/geom.py
PR Title: [MISC] Add benchmark test memory monitoring.
BEFORE CODE: 6D rotation representation, of size (*, 6)
[1] http://arxiv.org/abs/1812.07035
"""
if isinstance(R, torch.Tensor):
return R[..., :2, :].flatt... | Can you explain why this change 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. |
if T is None:
T = torch.zeros((*B, 4, 4), dtype=trans.dtype, device=trans.device)
elif is_numpy:
if T is None:
T = np.zeros((*B, 4, 4), dtype=trans.dtype)
else:
gs.raise_exception(
f"both of the inputs must be torch.Tensor or np.ndarray. got: {type(tra... | I don’t think changing this is necessary because the assert is already self explanatory. But as you want. | T = np.zeros((*B, 4, 4), dtype=trans.dtype)
else:
gs.raise_exception(
f"both of the inputs must be torch.Tensor or np.ndarray. got: {type(trans)=} and {type(quat)=}"
)
if B:
assert T.shape == (*B, 4, 4)
T[..., 3, 3] = 1.0
if trans is not None:
T[.... | @@ -924,7 +928,8 @@ def T_to_trans_quat(T, *, out=None):
@nb.jit(nopython=True, cache=True)
def _np_quat_mul(u, v, out=None):
- assert u.shape == v.shape | genesis/utils/geom.py | 26 | Python | 0.429 | suggestion | 105 | 51 | 51 | false | [MISC] Add benchmark test memory monitoring. | 1,981 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | hughperkins | Review this pull request change.
Language: Python
File: genesis/utils/geom.py
PR Title: [MISC] Add benchmark test memory monitoring.
BEFORE CODE: if T is None:
T = torch.zeros((*B, 4, 4), dtype=trans.dtype, device=trans.device)
elif is_numpy:
if T is None:
T = np.zeros((*... | I don’t think changing this is necessary because the assert is already self explanatory. 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. |
assert rotvec.ndim >= 1
B = rotvec.shape[:-1]
if out is None:
out_ = np.empty(B + (4,), dtype=rotvec.dtype)
else:
if out.shape != B + (4,):
raise ValueError("Shape mismatch")
out_ = out
# Split unit axis and positive angle
angle = np.sqrt(np.sum(np.square(rot... | Well, it is just obvious shape mismatch between cos_theta and axis. | # FIXME: Taylor expansion should be used to handle angle ~ 0.0
axis = rotvec / np.maximum(angle[..., None], gs.EPS)
# Compute the quaternion representation
out_[..., 0] = np.cos(0.5 * angle)
out_[..., 1:] = np.sin(0.5 * angle[..., None]) * axis
return out_
@nb.jit(nopython=True, cache=True)
... | @@ -1429,13 +1440,16 @@ def rotvec_to_quat(rotvec: np.ndarray, out: np.ndarray | None = None) -> np.ndar
@nb.jit(nopython=True, cache=True)
def _np_axis_cos_angle_to_R(axis: np.ndarray, cos_theta: np.ndarray, out: np.ndarray | None = None) -> np.ndarray:
if isinstance(cos_theta, (float, np.float32, np.float64)):... | genesis/utils/geom.py | 26 | Python | 0.214 | suggestion | 67 | 51 | 51 | false | [MISC] Add benchmark test memory monitoring. | 1,981 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | hughperkins | Review this pull request change.
Language: Python
File: genesis/utils/geom.py
PR Title: [MISC] Add benchmark test memory monitoring.
BEFORE CODE: assert rotvec.ndim >= 1
B = rotvec.shape[:-1]
if out is None:
out_ = np.empty(B + (4,), dtype=rotvec.dtype)
else:
if out.shape != B + (4,)... | Well, it is just obvious shape mismatch between cos_theta and axis. | 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. |
os.environ["PYGLET_HEADLESS"] = "1"
IS_INTERACTIVE_VIEWER_AVAILABLE = has_display or has_egl
TOL_SINGLE = 5e-5
TOL_DOUBLE = 1e-9
IMG_STD_ERR_THR = 1.0
IMG_NUM_ERR_THR = 0.001
def pytest_make_parametrize_id(config, val, argname):
if isinstance(val, Enum):
return val.name
if isinstance(val, type):... | I would rather only add the option if supported. | pyglet.options["headless"] = True
os.environ["PYGLET_HEADLESS"] = "1"
IS_INTERACTIVE_VIEWER_AVAILABLE = has_display or has_egl
TOL_SINGLE = 5e-5
TOL_DOUBLE = 1e-9
IMG_STD_ERR_THR = 1.0
IMG_NUM_ERR_THR = 0.001
def is_mem_monitoring_supported():
try:
assert sys.platform.startswith("linux")
... | @@ -64,7 +65,23 @@ def pytest_make_parametrize_id(config, val, argname):
return f"{val}"
-@pytest.hookimpl
+# @pytest.hookimpl(tryfirst=True)
+# def pytest_runtest_setup(item):
+# # Include test name in process title
+# test_name = item.nodeid.replace(" ", "")
+# setproctitle.setproctitle(f"pytest:... | tests/conftest.py | 26 | Python | 0.071 | suggestion | 48 | 51 | 51 | false | [MISC] Add benchmark test memory monitoring. | 1,981 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | hughperkins | Review this pull request change.
Language: Python
File: tests/conftest.py
PR Title: [MISC] Add benchmark test memory monitoring.
BEFORE CODE: os.environ["PYGLET_HEADLESS"] = "1"
IS_INTERACTIVE_VIEWER_AVAILABLE = has_display or has_egl
TOL_SINGLE = 5e-5
TOL_DOUBLE = 1e-9
IMG_STD_ERR_THR = 1.0
IMG_NUM_ERR_THR =... | I would rather only add the option if supported. | 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 pytest_runtest_setup(item):
# # Include test name in process title
# test_name = item.nodeid.replace(" ", "")
# setproctitle.setproctitle(f"pytest: {test_name}")
def validate_mem_option() -> None:
try:
assert sys.platform.startswith("linux")
nvidia_out = subprocess.check_output([... | If think you should check if xdist is enabled or not. If so, you can start it right away. If not, run this command in the main process. See what is done at the very end of the function for benchmarks. |
def pytest_make_parametrize_id(config, val, argname):
if isinstance(val, Enum):
return val.name
if isinstance(val, type):
return ".".join((val.__module__, val.__name__))
return f"{val}"
@pytest.hookimpl(tryfirst=True)
def pytest_cmdline_main(config: pytest.Config) -> None:
# Make sure... | @@ -73,6 +90,19 @@ def pytest_cmdline_main(config: pytest.Config) -> None:
except NameError as e:
raise pytest.UsageError(f"Unknown marker in CLI expression: '{e.name}'")
+ # Only launch memory monitor from the main process, not from xdist workers
+ if config.getoption("--mem-monitoring-filepath")... | tests/conftest.py | 26 | Python | 0.5 | suggestion | 200 | 51 | 51 | false | [MISC] Add benchmark test memory monitoring. | 1,981 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | hughperkins | Review this pull request change.
Language: Python
File: tests/conftest.py
PR Title: [MISC] Add benchmark test memory monitoring.
BEFORE CODE: # def pytest_runtest_setup(item):
# # Include test name in process title
# test_name = item.nodeid.replace(" ", "")
# setproctitle.setproctitle(f"pytest: {test_na... | If think you should check if xdist is enabled or not. If so, you can start it right away. If not, run this command in the main process. See what is done at the very end of the function for benchmarks. | 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. |
global use_ndarray, use_fastcache, use_zerocopy
is_ndarray_disabled = (os.environ.get("GS_ENABLE_NDARRAY") or ("0" if sys.platform == "darwin" else "1")) == "0"
if use_ndarray is None:
_use_ndarray = not (is_ndarray_disabled or performance_mode)
else:
_use_ndarray = use_ndarray
i... | ti.field won't support zero copy? | global use_ndarray, use_fastcache, use_zerocopy
is_ndarray_disabled = (os.environ.get("GS_ENABLE_NDARRAY") or ("0" if backend == gs_backend.metal else "1")) == "0"
if use_ndarray is None:
_use_ndarray = not (is_ndarray_disabled or performance_mode)
else:
_use_ndarray = use_ndarray
... | @@ -136,6 +137,15 @@ def init(
raise_exception("Genesis previous initialized. GsTaichi fast cache mode cannot be disabled anymore.")
use_ndarray, use_fastcache = _use_ndarray, _use_fastcache
+ # Unlike dynamic vs static array mode, and fastcache, zero-copy can be toggle on/off between init withou... | genesis/__init__.py | 26 | Python | 0.071 | question | 33 | 51 | 51 | false | [FEATURE] Leverage GsTaichi zero-copy in data accessors. | 2,011 | Genesis-Embodied-AI/Genesis | 26,702 | Python | YilingQiao | duburcqa | Review this pull request change.
Language: Python
File: genesis/__init__.py
PR Title: [FEATURE] Leverage GsTaichi zero-copy in data accessors.
BEFORE CODE: global use_ndarray, use_fastcache, use_zerocopy
is_ndarray_disabled = (os.environ.get("GS_ENABLE_NDARRAY") or ("0" if sys.platform == "darwin" else "1")... | ti.field won't support zero copy? | 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. |
mass2, com2 = 1.0, np.array([0.0, 0.0, 0.0])
inertia2 = np.array([[1.0, 0.0, 0.0], [0.0, 2.0, 0.0], [0.0, 0.0, 1.0]])
combined_mass, combined_com, combined_inertia = compose_inertial_properties(
mass1, com1, inertia1, mass2, com2, inertia2
)
expected_inertia = np.array([[3.0, 0.0, 0.0], [0... | Just implement a single test case that combines all the effects at once. | import math
from unittest.mock import patch
import pytest
import torch
import numpy as np
import genesis as gs
import genesis.utils.geom as gu
from genesis.utils.tools import FPSTracker
from genesis.utils.misc import tensor_to_array
from genesis.utils import warnings as warnings_mod
from genesis.utils.warnings import... | @@ -435,3 +436,78 @@ def test_fps_tracker():
fps = tracker.step(current_time=10.45)
# num envs * [num steps] / (delta time)
assert math.isclose(fps, n_envs * 4 / 0.14)
+
+
+@pytest.mark.required
+def test_compose_inertial_properties():
+ """Test composition of inertial properties of two bodies."""
+
+... | tests/test_utils.py | 26 | Python | 0.214 | suggestion | 72 | 27 | 469 | false | [CHANGING][FEATURE] Improve inertial estimation if undefined. | 2,100 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: tests/test_utils.py
PR Title: [CHANGING][FEATURE] Improve inertial estimation if undefined.
BEFORE CODE: mass2, com2 = 1.0, np.array([0.0, 0.0, 0.0])
inertia2 = np.array([[1.0, 0.0, 0.0], [0.0, 2.0, 0.0], [0.0, 0.0, 1.0]])
combined_mass, combined_... | Just implement a single test case that combines all the effects at once. | 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. |
"""Rotate inertia tensor I by rotation matrix R."""
return R @ I @ R.T
def compose_inertial_properties(mass1, com1, inertia1, mass2, com2, inertia2):
"""
Compose inertial properties of two bodies.
Args:
mass1: Mass of first body
com1: Center of mass of first body (3,) array
... | Just raise an exception is both masses are not `> gs.EPS` | """Rotate inertia tensor I by rotation matrix R."""
return R @ I @ R.T
def compose_inertial_properties(mass1, com1, inertia1, mass2, com2, inertia2):
"""
Compose inertial properties of two bodies.
Args:
mass1: Mass of first body
com1: Center of mass of first body (3,) array
... | @@ -419,6 +419,39 @@ def rotate_inertia(I, R):
return R @ I @ R.T
+def compose_inertial_properties(mass1, com1, inertia1, mass2, com2, inertia2):
+ """
+ Compose inertial properties of two bodies.
+
+ Args:
+ mass1: Mass of first body
+ com1: Center of mass of first body (3,) array
+ ... | genesis/utils/urdf.py | 26 | Python | 0.5 | suggestion | 57 | 51 | 51 | false | [CHANGING][FEATURE] Improve inertial estimation if undefined. | 2,100 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/utils/urdf.py
PR Title: [CHANGING][FEATURE] Improve inertial estimation if undefined.
BEFORE CODE: """Rotate inertia tensor I by rotation matrix R."""
return R @ I @ R.T
def compose_inertial_properties(mass1, com1, inertia1, mass2, com2, iner... | Just raise an exception is both masses are not `> gs.EPS` | 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._invweight: float | None = invweight
self._visualize_contact = visualize_contact
self._geoms: list[RigidGeom] = gs.List()
self._vgeoms: list[RigidVisGeom] = gs.List()
def _build(self):
for geom in self._geoms:
geom._build()
for vgeom in self._vgeo... | Avoid doing this. Just inline these function calls directly where they are used. | self._invweight: float | None = invweight
self._visualize_contact = visualize_contact
self._geoms: list[RigidGeom] = gs.List()
self._vgeoms: list[RigidVisGeom] = gs.List()
def _build(self):
for geom in self._geoms:
geom._build()
for vgeom in self._vgeo... | @@ -121,82 +122,88 @@ def _build(self):
for vgeom in self._vgeoms:
vgeom._build()
- self._init_mesh = self._compose_init_mesh()
-
- # inertial_mass and inertia_i
- if self._inertial_mass is None:
- if len(self._geoms) == 0 and len(self._vgeoms) == 0:
- ... | genesis/engine/entities/rigid_entity/rigid_link.py | 26 | Python | 0.286 | suggestion | 80 | 51 | 51 | false | [CHANGING][FEATURE] Improve inertial estimation if undefined. | 2,100 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/entities/rigid_entity/rigid_link.py
PR Title: [CHANGING][FEATURE] Improve inertial estimation if undefined.
BEFORE CODE: self._invweight: float | None = invweight
self._visualize_contact = visualize_contact
self._geoms:... | Avoid doing this. Just inline these function calls directly where they are used. | 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._geoms: list[RigidGeom] = gs.List()
self._vgeoms: list[RigidVisGeom] = gs.List()
def _build(self):
for geom in self._geoms:
geom._build()
for vgeom in self._vgeoms:
vgeom._build()
if self._inertial_mass is None or self._inertial_pos is None or... | Use 0.0. There is no reason to add some random mass in the kinematic tree. |
self._geoms: list[RigidGeom] = gs.List()
self._vgeoms: list[RigidVisGeom] = gs.List()
def _build(self):
for geom in self._geoms:
geom._build()
for vgeom in self._vgeoms:
vgeom._build()
if self._inertial_mass is None or self._inertial_pos is None or... | @@ -121,82 +122,88 @@ def _build(self):
for vgeom in self._vgeoms:
vgeom._build()
- self._init_mesh = self._compose_init_mesh()
-
- # inertial_mass and inertia_i
- if self._inertial_mass is None:
- if len(self._geoms) == 0 and len(self._vgeoms) == 0:
- ... | genesis/engine/entities/rigid_entity/rigid_link.py | 26 | Python | 0.286 | suggestion | 74 | 51 | 51 | false | [CHANGING][FEATURE] Improve inertial estimation if undefined. | 2,100 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/entities/rigid_entity/rigid_link.py
PR Title: [CHANGING][FEATURE] Improve inertial estimation if undefined.
BEFORE CODE:
self._geoms: list[RigidGeom] = gs.List()
self._vgeoms: list[RigidVisGeom] = gs.List()
def _build(self)... | Use 0.0. There is no reason to add some random mass in the kinematic tree. | 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._visualize_contact = visualize_contact
self._geoms: list[RigidGeom] = gs.List()
self._vgeoms: list[RigidVisGeom] = gs.List()
def _build(self):
for geom in self._geoms:
geom._build()
for vgeom in self._vgeoms:
vgeom._build()
if self._in... | This is not a fallback. It is the right solution where there is no geometry attached to a link. | self._visualize_contact = visualize_contact
self._geoms: list[RigidGeom] = gs.List()
self._vgeoms: list[RigidVisGeom] = gs.List()
def _build(self):
for geom in self._geoms:
geom._build()
for vgeom in self._vgeoms:
vgeom._build()
if self._in... | @@ -121,82 +122,88 @@ def _build(self):
for vgeom in self._vgeoms:
vgeom._build()
- self._init_mesh = self._compose_init_mesh()
-
- # inertial_mass and inertia_i
- if self._inertial_mass is None:
- if len(self._geoms) == 0 and len(self._vgeoms) == 0:
- ... | genesis/engine/entities/rigid_entity/rigid_link.py | 26 | Python | 0.214 | suggestion | 95 | 51 | 51 | false | [CHANGING][FEATURE] Improve inertial estimation if undefined. | 2,100 | Genesis-Embodied-AI/Genesis | 26,702 | Python | duburcqa | YilingQiao | Review this pull request change.
Language: Python
File: genesis/engine/entities/rigid_entity/rigid_link.py
PR Title: [CHANGING][FEATURE] Improve inertial estimation if undefined.
BEFORE CODE: self._visualize_contact = visualize_contact
self._geoms: list[RigidGeom] = gs.List()
self._vgeoms: ... | This is not a fallback. It is the right solution where there is no geometry attached to a link. | 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.