| compute_cost |
Ran ran on a synthesized input |
hybridrobotics/car-racing/car_racing/control/lmpc_helper.py code served (permissive licence) · get_code("9a70375264b211c3") How it ranRan in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (5,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.
import numpy as np
# Read the function to understand what it needs:
# - xcurv: a numpy array with shape (n, m) where it accesses xcurv[i, 4]
# So we need at least 5 columns (indices 0-4)
# - u: not used in the function body, but is a parameter
# - lap_length: a scalar that xcurv values are compared against
# Create xcurv: shape (batch_size, num_features) where num_features > 4
# Using shape (5, 6) to have distinct dimensions
xcurv = np.array([
[1.0, 2.0, 3.0, 4.0, 150.0, 6.0], # column 4 value = 150.0
[1.0, 2.0, 3.0, 4.0, 200.0, 6.0], # column 4 value = 200.0
[1.0, 2.0, 3.0, 4.0, 50.0, 6.0], # column 4 value = 50.0
[1.0, 2.0, 3.0, 4.0, 300.0, 6.0], # column 4 value = 300.0
[1.0, 2.0, 3.0, 4.0, 100.0, 6.0], # column 4 value = 100.0
], dtype=np.float32)
# u is not used in function body, so just pass None or a dummy value
u = None
# lap_length is used in comparison with xcurv[:, 4]
# Set it to a value that will create interesting branching
lap_length = 175.0
ARGS = {
"xcurv": xcurv,
"u": u,
"lap_length": lap_length
}
- Setup block drafted by
- a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
- Imports
- numpy — import names as written, not package names to install; a name the repository itself provides is not a package
- Returned
- type ndarray, ndim 1, dtype family float, shape (5,) — shape (5,) as a later run recorded it
- dtype family
- int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
- Environment
- image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
|
| compute_index |
Ran ran on a synthesized input |
hybridrobotics/car-racing/car_racing/control/lmpc_helper.py code served (permissive licence) · get_code("68e9c59f7f55b503") How it ranRan in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.
import numpy as np
import datetime
# Set up small test data
np.random.seed(0)
# Parameters based on function signature and body inspection:
# h: scalar threshold for norm comparison
h = 0.5
# ss_xcurv: 3D array [time, state_features, iterations]
# Shape: (10, 4, 2) - 10 time steps, 4 state features, 2 iterations
ss_xcurv = np.random.randn(10, 4, 2)
# u_ss: 3D array [time, inputs, iterations]
# Shape: (10, 3, 2) - 10 time steps, 3 inputs, 2 iterations
u_ss = np.random.randn(10, 3, 2)
# time_ss: array of time lengths for each iteration
# Shape: (2,) - one entry per iteration
time_ss = np.array([8, 7])
# iter: current iteration index
iter = 0
# x0: initial state vector
# Shape: (4,) - must match state_features dimension
x0 = np.array([0.1, 0.2, 0.3, 0.4])
# state_features: indices of state features to use
# Shape: (4,) - selecting 4 features from ss_xcurv
state_features = np.array([0, 1, 2, 3])
# scaling: scaling matrix applied to differences
# Shape: (4, 4) - must match number of state features when consider_input=False
# or (4+3, 4) when consider_input=True (state_features + inputs)
scaling = np.eye(4)
# max_num_point: maximum number of points to return
max_num_point = 5
# consider_input: boolean flag for including inputs in DataMatrix
consider_input = False
ARGS = {
'h': h,
'ss_xcurv': ss_xcurv,
'u_ss': u_ss,
'time_ss': time_ss,
'iter': iter,
'x0': x0,
'state_features': state_features,
'scaling': scaling,
'max_num_point': max_num_point,
'consider_input': consider_input,
}
- Setup block drafted by
- a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
- Imports
- numpy — import names as written, not package names to install; a name the repository itself provides is not a package
- Returned
- type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
- dtype family
- int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
- Environment
- image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
|
| ego_agent_overlap_checker |
Ran ran on a synthesized input |
hybridrobotics/car-racing/car_racing/planning/planner_helper.py code served (permissive licence) · get_code("bc24854144d3bf14") How it ranRan in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned bool, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.
# Setup code for ego_agent_overlap_checker
# Analyzing the function:
# - s_ego_min: numeric value (float/int) - minimum s coordinate of ego agent
# - s_ego_max: numeric value (float/int) - maximum s coordinate of ego agent
# - s_veh_min: numeric value (float/int) - minimum s coordinate of vehicle
# - s_veh_max: numeric value (float/int) - maximum s coordinate of vehicle
# - lap_length: numeric value (float/int) - length of the lap for wraparound
# The function performs overlap checking with lap wraparound logic
# All parameters are used in simple numeric comparisons
# Create distinct numeric values for testing
s_ego_min = 10.0
s_ego_max = 20.0
s_veh_min = 25.0
s_veh_max = 35.0
lap_length = 100.0
ARGS = {
"s_ego_min": s_ego_min,
"s_ego_max": s_ego_max,
"s_veh_min": s_veh_min,
"s_veh_max": s_veh_max,
"lap_length": lap_length,
}
- Setup block drafted by
- a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
- Imports
- none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
- Returned
- type bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
- dtype family
- int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
- Environment
- image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
|
| get_agent_range |
Ran ran on a synthesized input |
hybridrobotics/car-racing/car_racing/planning/planner_helper.py code served (permissive licence) · get_code("536c1618e2fa4c1d") How it ranRan in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.
import numpy as np
# Analyze the function: get_agent_range takes 5 scalar parameters
# s_agent: longitudinal position (scalar)
# ey_agent: lateral position (scalar)
# epsi_agent: yaw angle (scalar)
# length: vehicle length (scalar)
# width: vehicle width (scalar)
# All parameters are used in numpy operations (sin, cos, arithmetic)
ARGS = {
"s_agent": 10.0,
"ey_agent": 2.5,
"epsi_agent": 0.3,
"length": 4.5,
"width": 2.0,
}
- Setup block drafted by
- a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
- Imports
- numpy — import names as written, not package names to install; a name the repository itself provides is not a package
- Returned
- type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
- dtype family
- int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
- Environment
- image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
|
| get_bezier_control_points |
Ran ran on a synthesized input |
hybridrobotics/car-racing/car_racing/planning/planner_helper.py code served (permissive licence) · get_code("fee9029a83ea8862") How it ranRan in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 3, dtype family float, shape (3, 4, 2) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.
import numpy as np
from scipy.interpolate import interp1d
class VehicleParam:
def __init__(self, length, width):
self.length = length
self.width = width
class Vehicle:
def __init__(self, length=2.0, width=1.5):
self.param = VehicleParam(length, width)
class AgentInfo:
def __init__(self, max_delta_v):
self.max_delta_v = max_delta_v
class RacingGameParam:
def __init__(self, planning_prediction_factor, safety_factor, bezier_order):
self.planning_prediction_factor = planning_prediction_factor
self.safety_factor = safety_factor
self.bezier_order = bezier_order
class Track:
def __init__(self, lap_length, width):
self.lap_length = lap_length
self.width = width
# Create vehicles_interest dictionary with 2 vehicles
vehicles_interest = {
0: Vehicle(length=2.0, width=1.5),
1: Vehicle(length=2.0, width=1.5)
}
# Create veh_info_list: (num_vehicles, 2) array with position info
veh_info_list = np.array([
[10.0, 0.5],
[12.0, -0.3]
])
# Create agent_info
agent_info = AgentInfo(max_delta_v=5.0)
# Create racing_game_param
racing_game_param = RacingGameParam(
planning_prediction_factor=2.0,
safety_factor=1.2,
bezier_order=3
)
# Create track
track = Track(lap_length=100.0, width=8.0)
# Create optimal_traj_xcurv: (num_points, 6) array
# Columns: [0, 1, 2, 3, 4 (s/xcurv), 5 (ey)]
optimal_traj_xcurv = np.array([
[0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[1.0, 0.1, 0.0, 0.0, 5.0, 0.2],
[2.0, 0.2, 0.0, 0.0, 10.0, 0.4],
[3.0, 0.3, 0.0, 0.0, 15.0, 0.3],
[4.0, 0.4, 0.0, 0.0, 20.0, 0.1]
])
# Create sorted_vehicles (list of vehicle indices)
sorted_vehicles = [0, 1]
# Create xcurv_ego: [x, y, vx, vy, s/xcurv, ey]
xcurv_ego = np.array([8.0, 0.0, 1.0, 0.0, 5.0, 0.1])
ARGS = {
"vehicles_interest": vehicles_interest,
"veh_info_list": veh_info_list,
"agent_info": agent_info,
"racing_game_param": racing_game_param,
"track": track,
"optimal_traj_xcurv": optimal_traj_xcurv,
"sorted_vehicles": sorted_vehicles,
"xcurv_ego": xcurv_ego
}
- Setup block drafted by
- a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
- Imports
- numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
- Returned
- type ndarray, ndim 3, dtype family float, shape (3, 4, 2) — shape (3, 4, 2) as a later run recorded it
- dtype family
- int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
- Environment
- image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
|
| get_local_position |
Ran ran on a synthesized input |
hybridrobotics/car-racing/car_racing/utils/racing_env.py code served (permissive licence) · get_code("a619d00ff3635f98") How it ranRan in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.
import numpy as np
# Create a simple point_and_tangent array
# Each row: [x, y, psi, s_start, arc_length, curvature]
# We'll create a track with a straight segment followed by a curved segment
point_and_tangent = np.array([
[0.0, 0.0, 0.0, 0.0, 10.0, 0.0], # straight segment start
[10.0, 0.0, 0.0, 10.0, 10.0, 0.1], # curved segment
[15.0, 5.0, np.pi/4, 20.0, 10.0, 0.0] # another straight segment
])
# Set up parameters
lap_length = 30.0
width = 2.0
# Position in inertial frame - on the track, first segment
x = 5.0
y = 0.0
psi = 0.0
ARGS = {
"lap_length": lap_length,
"width": width,
"point_and_tangent": point_and_tangent,
"x": x,
"y": y,
"psi": psi
}
- Setup block drafted by
- a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
- Imports
- numpy — import names as written, not package names to install; a name the repository itself provides is not a package
- Returned
- type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
- dtype family
- int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
- Environment
- image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
|
| linear_regression |
Ran ran on a synthesized input |
hybridrobotics/car-racing/car_racing/system/system_identification.py code served (permissive licence) · get_code("2a3ad68ace23d315") How it ranRan in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.
import numpy as np
# Construct valid arguments for linear_regression
# The function expects:
# - x: state data with shape (T, 6) where T is time steps
# - u: control input data with shape (T, 2) where T is time steps
# - lamb: regularization coefficient (scalar)
# Looking at the function body:
# Y = x[2:x.shape[0], :] extracts rows 2 onwards
# X = np.hstack((x[1:x.shape[0]-1, :], u[1:x.shape[0]-1, :]))
# creates matrix from x rows 1 to T-2 and u rows 1 to T-2
# So we need at least 3 time steps
# x has 6 state dimensions (based on A = W.T[:, 0:6])
# u has 2 control dimensions (based on B = W.T[:, 6:8])
T = 5 # Number of time steps
state_dim = 6
control_dim = 2
# Create state trajectory x with shape (T, 6)
x = np.random.randn(T, state_dim)
# Create control input u with shape (T, 2)
u = np.random.randn(T, control_dim)
# Regularization coefficient
lamb = 0.01
ARGS = {
'x': x,
'u': u,
'lamb': lamb
}
- Setup block drafted by
- a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
- Imports
- numpy — import names as written, not package names to install; a name the repository itself provides is not a package
- Returned
- type tuple — a tuple output has no array shape of its own; none is stated
- Environment
- image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
|
| repelling_cost_function |
Ran ran on a synthesized input |
hybridrobotics/car-racing/car_racing/control/ilqr_helper.py code served (permissive licence) · get_code("99d4743cb83798b6") How it ranRan in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.
import numpy as np
# Set seed for reproducibility
np.random.seed(0)
# Create small numpy arrays for the function parameters
# q1: scalar or array that multiplies the exponential
q1 = np.array(2.5)
# q2: scalar or array that scales c in the exponential
q2 = np.array(0.3)
# c: small array (dim 4)
c = np.random.randn(4)
# c_dot: small array matching c's shape (represents derivative of c)
c_dot = np.random.randn(4)
ARGS = {
"q1": q1,
"q2": q2,
"c": c,
"c_dot": c_dot
}
- Setup block drafted by
- a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
- Imports
- numpy — import names as written, not package names to install; a name the repository itself provides is not a package
- Returned
- type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
- dtype family
- int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
- Environment
- image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
|
| vehicle_dynamics |
Ran ran on a synthesized input |
hybridrobotics/car-racing/car_racing/system/vehicle_dynamics.py code served (permissive licence) · get_code("e98907b0a8ef0600") How it ranRan in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.
import numpy as np
class DynamicsParam:
"""Stand-in class for dynamics parameters."""
def __init__(self):
self.m = 1500.0 # mass
self.lf = 1.2 # distance from CoG to front axle
self.lr = 1.6 # distance from CoG to rear axle
self.Iz = 2500.0 # yaw inertia
self.Df = 0.5 # front tire stiffness
self.Cf = 1.3 # front tire shape
self.Bf = 10.0 # front tire slip angle
self.Dr = 0.5 # rear tire stiffness
self.Cr = 1.3 # rear tire shape
self.Br = 10.0 # rear tire slip angle
def get_params(self):
return (self.m, self.lf, self.lr, self.Iz,
self.Df, self.Cf, self.Bf, self.Dr, self.Cr, self.Br)
dynamics_param = DynamicsParam()
curv = 0.01 # road curvature
xglob = np.array([10.0, 0.5, 0.02, 0.05, 100.0, 50.0]) # [vx, vy, wz, psi, X, Y]
xcurv = np.array([10.0, 0.5, 0.02, 0.01, 50.0, 0.1]) # [vx, vy, wz, epsi, s, ey]
delta_t = 0.01 # time step
u = np.array([0.05, 1.0]) # [delta (steering angle), a (acceleration)]
ARGS = {
'dynamics_param': dynamics_param,
'curv': curv,
'xglob': xglob,
'xcurv': xcurv,
'delta_t': delta_t,
'u': u
}
- Setup block drafted by
- a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
- Imports
- numpy — import names as written, not package names to install; a name the repository itself provides is not a package
- Returned
- type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
- dtype family
- int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
- Environment
- image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
|
| get_cost_derivation |
Not yet run unverified (harvested, no recorded run) |
hybridrobotics/car-racing/car_racing/control/ilqr_helper.py code served (permissive licence) · get_code("6c4f235bb6df72d7") How it ran |
| get_global_position |
Not yet run unverified (harvested, no recorded run) |
hybridrobotics/car-racing/car_racing/utils/racing_env.py code served (permissive licence) · get_code("5b17e602cb86206c") How it ran |
| get_orientation |
Not yet run unverified (harvested, no recorded run) |
hybridrobotics/car-racing/car_racing/utils/racing_env.py code served (permissive licence) · get_code("e0acd10d72c27bd5") How it ran |
| get_udata |
Not yet run unverified (harvested, no recorded run) |
hybridrobotics/car-racing/car_racing/system/system_identification.py code served (permissive licence) · get_code("26dea725f8e95762") How it ran |
| lqr |
Not yet run unverified (harvested, no recorded run) |
hybridrobotics/car-racing/car_racing/control/control.py code served (permissive licence) · get_code("efb1f8e9bb437359") How it ran |
| pid |
Not yet run unverified (harvested, no recorded run) |
hybridrobotics/car-racing/car_racing/control/control.py code served (permissive licence) · get_code("46cb55571f7b0e54") How it ran |