Compare commits

...

6 Commits

Author SHA1 Message Date
Jason Wen da6313dbe9 Update CHANGELOG.md 2026-06-26 21:33:25 -04:00
Jason Wen 01a843e0ac ui: reset Enforce Torque Control and NNLC if both are enabled (#1863)
* ui: reset Enforce Torque Control and NNLC if both are enabled

* block
2026-06-09 02:15:42 -04:00
Jason Wen 097dd9b5f2 sunnylink: deprecate legacy params metadata (#1862) 2026-06-08 23:50:43 -04:00
github-actions[bot] 122ac986de [bot] Update Python packages (#1861)
Update Python packages

Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
2026-06-08 23:35:54 -04:00
MVL bc27262a92 Revert "deprecate carState.brake" for Honda Gas Interceptor (#1860)
* Return Honda exception removed in commaai#37857

Added handling for brakePressed state with conditions for Honda models.

* Add CAR back to imports

* match

---------

Co-authored-by: Jason Wen <haibin.wen3@gmail.com>
2026-06-08 23:20:51 -04:00
James Vecellio-Grant 066ba92e77 modeld_v2: safe model validation (#1855)
* modeld_v2: safe model validation

* fix string

* numpy

* dumb

* god use full attribute names please

---------

Co-authored-by: Jason Wen <haibin.wen3@gmail.com>
2026-06-07 23:07:59 -04:00
24 changed files with 230 additions and 2392 deletions
+30 -1
View File
@@ -1,5 +1,34 @@
sunnypilot Version 2026.002.000 (2026-xx-xx)
sunnypilot Version 2026.002.000 (2026-06-28)
========================
* What's Changed (sunnypilot/sunnypilot)
* ui: update gates for certain toggles by @sunnyhaibin in https://github.com/sunnypilot/sunnypilot/pull/1830
* release: ignore upstream IsReleaseBranch by @sunnyhaibin in https://github.com/sunnypilot/sunnypilot/pull/1831
* manager: disable DEVELOPMENT_ONLY reset by @sunnyhaibin in https://github.com/sunnypilot/sunnypilot/pull/1833
* sunnylink: fix max time offroad values by @nayan8teen in https://github.com/sunnypilot/sunnypilot/pull/1835
* ui: show default model name by @nayan8teen in https://github.com/sunnypilot/sunnypilot/pull/1837
* sunnylink: add CarParams fallback for brand-specific capabilities by @sunnyhaibin in https://github.com/sunnypilot/sunnypilot/pull/1839
* sunnylink SDUI: tweak DisableUpdate param for clarity by @sunnyhaibin in https://github.com/sunnypilot/sunnypilot/pull/1842
* Revert "DM: Lancia Delta HF Integrale model" by @sunnyhaibin in https://github.com/sunnypilot/sunnypilot/pull/1849
* modeld_v2: safe model validation by @Discountchubbs in https://github.com/sunnypilot/sunnypilot/pull/1855
* Revert "deprecate `carState.brake`" for Honda Gas Interceptor by @mvl-boston in https://github.com/sunnypilot/sunnypilot/pull/1860
* sunnylink: deprecate legacy params metadata by @sunnyhaibin in https://github.com/sunnypilot/sunnypilot/pull/1862
* ui: reset Enforce Torque Control and NNLC if both are enabled by @sunnyhaibin in https://github.com/sunnypilot/sunnypilot/pull/1863
* What's Changed (sunnypilot/opendbc)
* Rivian: suppress ACM hold-the-wheel warning during MADS-only lateral by @lukasloetkolben in https://github.com/sunnypilot/opendbc/pull/465
* Sync: `commaai/opendbc:master``sunnypilot/opendbc:master` by @sunnyhaibin in https://github.com/sunnypilot/opendbc/pull/479
* safety: add option to ignore frequency check for RX checks by @sunnyhaibin in https://github.com/sunnypilot/opendbc/pull/480
* Revert "deprecate carState.brake" for Honda Gas Interceptor by @mvl-boston in https://github.com/sunnypilot/opendbc/pull/481
* New Contributors (sunnypilot/sunnypilot)
* @mvl-boston made their first contribution in https://github.com/sunnypilot/sunnypilot/pull/1860
* Full Changelog: https://github.com/sunnypilot/sunnypilot/compare/v2026.001.007...v2026.002.000
************************
* Synced with commaai's openpilot (v0.11.1)
* master commit 69e2c321e49760e52f7983eaa0a5f77cb95de637 (June 02, 2026)
* New driver monitoring model
* Improved image processing pipeline for driver camera
* Improved thermal policy for comma four
* Acura MDX 2022-24 support thanks to mvl-boston!
* Rivian R1S and R1T 2025 support thanks to lukasloetkolben!
sunnypilot Version 2026.001.000 (2026-05-06)
========================
+14 -3
View File
@@ -13,7 +13,7 @@ from opendbc.car import DT_CTRL, gen_empty_fingerprint, structs
from opendbc.car.can_definitions import CanData
from opendbc.car.car_helpers import FRAME_FINGERPRINT, interfaces
from opendbc.car.fingerprints import MIGRATION
from opendbc.car.honda.values import HondaFlags
from opendbc.car.honda.values import CAR as HONDA, HondaFlags
from opendbc.car.structs import car
from opendbc.car.tests.routes import non_tested_cars, routes, CarTestRoute
from opendbc.car.values import Platform, PLATFORMS
@@ -358,7 +358,13 @@ class TestCarModelBase(unittest.TestCase):
self.assertEqual(CS.gasPressed, self.safety.get_gas_pressed_prev())
if self.safety.get_brake_pressed_prev() != prev_panda_brake:
self.assertEqual(CS.brakePressed, self.safety.get_brake_pressed_prev())
# TODO: remove this exception once this mismatch is resolved
brake_pressed = CS.brakePressed
if CS.brakePressed and not self.safety.get_brake_pressed_prev():
if self.CP.carFingerprint in (HONDA.HONDA_PILOT, HONDA.HONDA_RIDGELINE) and CS.brake > 0.05:
brake_pressed = False
self.assertEqual(brake_pressed, self.safety.get_brake_pressed_prev())
if self.safety.get_regen_braking_prev() != prev_panda_regen_braking:
self.assertEqual(CS.regenBraking, self.safety.get_regen_braking_prev())
@@ -442,7 +448,12 @@ class TestCarModelBase(unittest.TestCase):
checks['steeringAngleDeg'] += (angle_can > (self.safety.get_angle_meas_max() + 1) or
angle_can < (self.safety.get_angle_meas_min() - 1))
checks['brakePressed'] += CS.brakePressed != self.safety.get_brake_pressed_prev()
# TODO: remove this exception once this mismatch is resolved
brake_pressed = CS.brakePressed
if CS.brakePressed and not self.safety.get_brake_pressed_prev():
if self.CP.carFingerprint in (HONDA.HONDA_PILOT, HONDA.HONDA_RIDGELINE) and CS.brakeDEPRECATED > 0.05:
brake_pressed = False
checks['brakePressed'] += brake_pressed != self.safety.get_brake_pressed_prev()
checks['regenBraking'] += CS.regenBraking != self.safety.get_regen_braking_prev()
checks['steeringDisengage'] += CS.steeringDisengage != self.safety.get_steering_disengage_prev()
@@ -134,6 +134,11 @@ class SteeringLayout(Widget):
enforce_torque_enabled = self._torque_control_toggle.action_item.get_state()
nnlc_enabled = self._nnlc_toggle.action_item.get_state()
if enforce_torque_enabled and nnlc_enabled:
self._torque_control_toggle.action_item.set_state(False)
self._nnlc_toggle.action_item.set_state(False)
enforce_torque_enabled = False
nnlc_enabled = False
self._nnlc_toggle.action_item.set_enabled(ui_state.is_offroad() and torque_allowed and not enforce_torque_enabled)
self._torque_control_toggle.action_item.set_enabled(ui_state.is_offroad() and torque_allowed and not nnlc_enabled)
self._torque_customization_button.action_item.set_enabled(self._torque_control_toggle.action_item.get_state())
+4
View File
@@ -179,6 +179,10 @@ class UIStateSP:
CP = self.CP
if CP is not None:
if self.params.get_bool("EnforceTorqueControl") and self.params.get_bool("NeuralNetworkLateralControl"):
self.params.put_bool("EnforceTorqueControl", False, block=True)
self.params.put_bool("NeuralNetworkLateralControl", False, block=True)
# Angle steering: no torque-based lateral controls
if CP.steerControlType == car.CarParams.SteerControlType.angle:
self.params.remove("EnforceTorqueControl")
+10 -10
View File
@@ -188,7 +188,7 @@ def make_supercombo_input_queues(input_shapes, frame_skip, device):
n_frames = img_shape[1] // 6
img_buf_shape = (frame_skip * (n_frames - 1) + 1, 6, img_shape[2], img_shape[3])
npy_keys = {}
numpy_keys = {}
queue_keys = {}
for key, shape in input_shapes.items():
@@ -196,7 +196,7 @@ def make_supercombo_input_queues(input_shapes, frame_skip, device):
continue
if len(shape) == 3 and shape[1] > 1:
if key.startswith('desire'):
npy_keys[key] = np.zeros(shape[2], dtype=np.float32)
numpy_keys[key] = np.zeros(shape[2], dtype=np.float32)
queue_keys[f'{key}_q'] = Tensor(
np.zeros((frame_skip * shape[1], shape[0], shape[2]), dtype=np.float32),
device=device).contiguous().realize()
@@ -205,24 +205,24 @@ def make_supercombo_input_queues(input_shapes, frame_skip, device):
np.zeros((frame_skip * (shape[1] - 1) + 1, shape[0], shape[2]), dtype=np.float32),
device=device).contiguous().realize()
else:
npy_keys[key] = np.zeros(shape, dtype=np.float32)
numpy_keys[key] = np.zeros(shape, dtype=np.float32)
elif len(shape) == 2:
npy_keys[key] = np.zeros(shape, dtype=np.float32)
numpy_keys[key] = np.zeros(shape, dtype=np.float32)
if 'traffic_convention' not in npy_keys:
if 'traffic_convention' not in numpy_keys:
tc_shape = input_shapes.get('traffic_convention', (1, 2))
npy_keys['traffic_convention'] = np.zeros(tc_shape, dtype=np.float32)
numpy_keys['traffic_convention'] = np.zeros(tc_shape, dtype=np.float32)
npy_keys['tfm'] = np.zeros((3, 3), dtype=np.float32)
npy_keys['big_tfm'] = np.zeros((3, 3), dtype=np.float32)
numpy_keys['tfm'] = np.zeros((3, 3), dtype=np.float32)
numpy_keys['big_tfm'] = np.zeros((3, 3), dtype=np.float32)
input_queues = {
'img_q': Tensor(np.zeros(img_buf_shape, dtype=np.uint8), device=device).contiguous().realize(),
'big_img_q': Tensor(np.zeros(img_buf_shape, dtype=np.uint8), device=device).contiguous().realize(),
**queue_keys,
**{k: Tensor(v, device='NPY').realize() for k, v in npy_keys.items()},
**{k: Tensor(v, device='NPY').realize() for k, v in numpy_keys.items()},
}
return input_queues, npy_keys
return input_queues, numpy_keys
def make_run_supercombo(model_runner, nv12: NV12Frame, model_w, model_h,
+11 -16
View File
@@ -62,11 +62,6 @@ def _find_driving_pkl(bundle):
if _pkl_exists(pkl_path):
return pkl_path
fallback = os.path.join(model_root, 'driving_tinygrad.pkl')
if _pkl_exists(fallback):
return fallback
return None
class FrameMeta:
frame_id: int = 0
@@ -125,7 +120,7 @@ class ModelState(ModelStateBase):
self._vision_input_names = [k for k in model_metadata['input_shapes'] if 'img' in k]
from openpilot.sunnypilot.modeld_v2.compile_modeld import make_supercombo_input_queues
frame_skip = derive_frame_skip({}, model_metadata['input_shapes'])
self.input_queues, self.npy = make_supercombo_input_queues(model_metadata['input_shapes'], frame_skip, device=self.DEV)
self.input_queues, self.numpy_inputs = make_supercombo_input_queues(model_metadata['input_shapes'], frame_skip, device=self.DEV)
else:
vision_metadata = metadata['vision']
policy_keys = [k for k in metadata if k != 'vision']
@@ -143,7 +138,7 @@ class ModelState(ModelStateBase):
policy_input_shapes = first_policy_metadata['input_shapes']
self._vision_input_names = [k for k in vision_input_shapes if 'img' in k]
frame_skip = derive_frame_skip(vision_input_shapes, policy_input_shapes)
self.input_queues, self.npy = make_split_input_queues(vision_input_shapes, policy_input_shapes, frame_skip, device=self.DEV)
self.input_queues, self.numpy_inputs = make_split_input_queues(vision_input_shapes, policy_input_shapes, frame_skip, device=self.DEV)
from openpilot.sunnypilot.modeld_v2.parse_model_outputs_split import Parser as SplitParser
from openpilot.sunnypilot.modeld_v2.parse_model_outputs import Parser as CombinedParser
@@ -183,7 +178,7 @@ class ModelState(ModelStateBase):
@property
def desire_key(self) -> str:
return next(k for k in self.npy if k.startswith('desire'))
return next(k for k in self.numpy_inputs if k.startswith('desire'))
def run(self, bufs: dict[str, VisionBuf], transforms: dict[str, np.ndarray],
inputs: dict[str, np.ndarray], prepare_only: bool) -> dict[str, np.ndarray] | None:
@@ -199,16 +194,16 @@ class ModelState(ModelStateBase):
desire_key = self.desire_key
inputs[desire_key][0] = 0
self.npy[desire_key][:] = np.where(inputs[desire_key] - self.prev_desire > .99, inputs[desire_key], 0)
self.numpy_inputs[desire_key][:] = np.where(inputs[desire_key] - self.prev_desire > .99, inputs[desire_key], 0)
self.prev_desire[:] = inputs[desire_key]
for key in ('traffic_convention', 'lateral_control_params'):
if key in self.npy and key in inputs:
self.npy[key][:] = inputs[key]
if key in self.numpy_inputs and key in inputs:
self.numpy_inputs[key][:] = inputs[key]
road_key = next(n for n in bufs if 'big' not in n)
wide_key = next(n for n in bufs if 'big' in n)
self.npy['tfm'][:, :] = transforms[road_key].reshape(3, 3)
self.npy['big_tfm'][:, :] = transforms[wide_key].reshape(3, 3)
self.numpy_inputs['tfm'][:, :] = transforms[road_key].reshape(3, 3)
self.numpy_inputs['big_tfm'][:, :] = transforms[wide_key].reshape(3, 3)
if prepare_only:
self._warp_enqueue(**self.input_queues, frame=self.full_frames[road_key], big_frame=self.full_frames[wide_key])
@@ -236,8 +231,8 @@ class ModelState(ModelStateBase):
if 'planplus' in outputs and 'plan' in outputs:
outputs['plan'] = outputs['plan'] + outputs['planplus']
if 'desired_curvature' in outputs and 'prev_desired_curv' in self.npy:
buf = self.npy['prev_desired_curv']
if 'desired_curvature' in outputs and 'prev_desired_curv' in self.numpy_inputs:
buf = self.numpy_inputs['prev_desired_curv']
buf[0, :-1] = buf[0, 1:]
buf[0, -1, :] = outputs['desired_curvature'][0, :] if not self.mlsim else 0
@@ -409,7 +404,7 @@ def main(demo=False):
'traffic_convention': traffic_convention,
}
if 'lateral_control_params' in model.npy:
if 'lateral_control_params' in model.numpy_inputs:
inputs['lateral_control_params'] = np.array([v_ego, lat_delay], dtype=np.float32)
mt1 = time.perf_counter()
-62
View File
@@ -1,62 +0,0 @@
## Neural networks in openpilot
To view the architecture of the ONNX networks, you can use [netron](https://netron.app/)
## Supercombo
### Supercombo input format (Full size: 799906 x float32)
* **image stream**
* Two consecutive images (256 * 512 * 3 in RGB) recorded at 20 Hz : 393216 = 2 * 6 * 128 * 256
* Each 256 * 512 image is represented in YUV420 with 6 channels : 6 * 128 * 256
* Channels 0,1,2,3 represent the full-res Y channel and are represented in numpy as Y[::2, ::2], Y[::2, 1::2], Y[1::2, ::2], and Y[1::2, 1::2]
* Channel 4 represents the half-res U channel
* Channel 5 represents the half-res V channel
* **wide image stream**
* Two consecutive images (256 * 512 * 3 in RGB) recorded at 20 Hz : 393216 = 2 * 6 * 128 * 256
* Each 256 * 512 image is represented in YUV420 with 6 channels : 6 * 128 * 256
* Channels 0,1,2,3 represent the full-res Y channel and are represented in numpy as Y[::2, ::2], Y[::2, 1::2], Y[1::2, ::2], and Y[1::2, 1::2]
* Channel 4 represents the half-res U channel
* Channel 5 represents the half-res V channel
* **desire**
* one-hot encoded buffer to command model to execute certain actions, bit needs to be sent for the past 5 seconds (at 20FPS) : 100 * 8
* **traffic convention**
* one-hot encoded vector to tell model whether traffic is right-hand or left-hand traffic : 2
* **feature buffer**
* A buffer of intermediate features that gets appended to the current feature to form a 5 seconds temporal context (at 20FPS) : 99 * 512
### Supercombo output format (Full size: XXX x float32)
Read [here](https://github.com/commaai/openpilot/blob/90af436a121164a51da9fa48d093c29f738adf6a/selfdrive/modeld/models/driving.h#L236) for more.
## Driver Monitoring Model
* .onnx model can be run with onnx runtimes
* .dlc file is a pre-quantized model and only runs on qualcomm DSPs
### input format
* single image W = 1440 H = 960 luminance channel (Y) from the planar YUV420 format:
* full input size is 1440 * 960 = 1382400
* normalized ranging from 0.0 to 1.0 in float32 (onnx runner) or ranging from 0 to 255 in uint8 (snpe runner)
* camera calibration angles (roll, pitch, yaw) from liveCalibration: 3 x float32 inputs
### output format
* 84 x float32 outputs = 2 + 41 * 2 ([parsing example](https://github.com/commaai/openpilot/blob/22ce4e17ba0d3bfcf37f8255a4dd1dc683fe0c38/selfdrive/modeld/models/dmonitoring.cc#L33))
* for each person in the front seats (2 * 41)
* face pose: 12 = 6 + 6
* face orientation [pitch, yaw, roll] in camera frame: 3
* face position [dx, dy] relative to image center: 2
* normalized face size: 1
* standard deviations for above outputs: 6
* face visible probability: 1
* eyes: 20 = (8 + 1) + (8 + 1) + 1 + 1
* eye position and size, and their standard deviations: 8
* eye visible probability: 1
* eye closed probability: 1
* wearing sunglasses probability: 1
* face occluded probability: 1
* touching wheel probability: 1
* paying attention probability: 1
* (deprecated) distracted probabilities: 2
* using phone probability: 1
* distracted probability: 1
* common outputs 2
* poor camera vision probability: 1
* left hand drive probability: 1
@@ -1,101 +0,0 @@
// clang++ -O2 repro.cc && ./a.out
#include <sched.h>
#include <sys/types.h>
#include <unistd.h>
#include <cstdint>
#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <ctime>
static inline double millis_since_boot() {
struct timespec t;
clock_gettime(CLOCK_BOOTTIME, &t);
return t.tv_sec * 1000.0 + t.tv_nsec * 1e-6;
}
#define MODEL_WIDTH 320
#define MODEL_HEIGHT 640
// null function still breaks it
#define input_lambda(x) x
// this is copied from models/dmonitoring.cc, and is the code that triggers the issue
void inner(uint8_t *resized_buf, float *net_input_buf) {
int resized_width = MODEL_WIDTH;
int resized_height = MODEL_HEIGHT;
// one shot conversion, O(n) anyway
// yuvframe2tensor, normalize
for (int r = 0; r < MODEL_HEIGHT/2; r++) {
for (int c = 0; c < MODEL_WIDTH/2; c++) {
// Y_ul
net_input_buf[(c*MODEL_HEIGHT/2) + r] = input_lambda(resized_buf[(2*r*resized_width) + (2*c)]);
// Y_ur
net_input_buf[(c*MODEL_HEIGHT/2) + r + (2*(MODEL_WIDTH/2)*(MODEL_HEIGHT/2))] = input_lambda(resized_buf[(2*r*resized_width) + (2*c+1)]);
// Y_dl
net_input_buf[(c*MODEL_HEIGHT/2) + r + ((MODEL_WIDTH/2)*(MODEL_HEIGHT/2))] = input_lambda(resized_buf[(2*r*resized_width+1) + (2*c)]);
// Y_dr
net_input_buf[(c*MODEL_HEIGHT/2) + r + (3*(MODEL_WIDTH/2)*(MODEL_HEIGHT/2))] = input_lambda(resized_buf[(2*r*resized_width+1) + (2*c+1)]);
// U
net_input_buf[(c*MODEL_HEIGHT/2) + r + (4*(MODEL_WIDTH/2)*(MODEL_HEIGHT/2))] = input_lambda(resized_buf[(resized_width*resized_height) + (r*resized_width/2) + c]);
// V
net_input_buf[(c*MODEL_HEIGHT/2) + r + (5*(MODEL_WIDTH/2)*(MODEL_HEIGHT/2))] = input_lambda(resized_buf[(resized_width*resized_height) + ((resized_width/2)*(resized_height/2)) + (r*resized_width/2) + c]);
}
}
}
float trial() {
int resized_width = MODEL_WIDTH;
int resized_height = MODEL_HEIGHT;
int yuv_buf_len = (MODEL_WIDTH/2) * (MODEL_HEIGHT/2) * 6; // Y|u|v -> y|y|y|y|u|v
// allocate the buffers
uint8_t *resized_buf = (uint8_t*)malloc(resized_width*resized_height*3/2);
float *net_input_buf = (float*)malloc(yuv_buf_len*sizeof(float));
printf("allocate -- %p 0x%x -- %p 0x%lx\n", resized_buf, resized_width*resized_height*3/2, net_input_buf, yuv_buf_len*sizeof(float));
// test for bad buffers
static int CNT = 20;
float avg = 0.0;
for (int i = 0; i < CNT; i++) {
double s4 = millis_since_boot();
inner(resized_buf, net_input_buf);
double s5 = millis_since_boot();
avg += s5-s4;
}
avg /= CNT;
// once it's bad, it's reliably bad
if (avg > 10) {
printf("HIT %f\n", avg);
printf("BAD\n");
for (int i = 0; i < 200; i++) {
double s4 = millis_since_boot();
inner(resized_buf, net_input_buf);
double s5 = millis_since_boot();
printf("%.2f ", s5-s4);
}
printf("\n");
exit(0);
}
// don't free so we get a different buffer each time
//free(resized_buf);
//free(net_input_buf);
return avg;
}
int main() {
while (true) {
float ret = trial();
printf("got %f\n", ret);
}
}
@@ -46,16 +46,6 @@ class TestFindDrivingPkl:
assert result is not None
assert 'driving_fof_tinygrad.pkl' in result
def test_finds_fallback_driving_tinygrad(self, tmp_path, monkeypatch):
(tmp_path / 'driving_tinygrad.pkl').write_bytes(b'fake')
from openpilot.system.hardware import hw
monkeypatch.setattr(hw.Paths, 'model_root', staticmethod(lambda: str(tmp_path)))
bundle = DummyBundle(models=[DummyModel('vision', 'nonexistent.pkl')])
result = _find_driving_pkl(bundle)
assert result is not None
assert 'driving_tinygrad.pkl' in result
# Init — assertion guard
@@ -84,8 +74,8 @@ class TestStockEquivalence:
skip_keys = {'action_t'}
assert set(state.input_queues.keys()) == set(stock_queues.keys()) - skip_keys, \
f"Queue keys differ: v2={set(state.input_queues.keys())}, stock={set(stock_queues.keys())}"
assert set(state.npy.keys()) == set(stock_npy.keys()) - skip_keys, \
f"Npy keys differ: v2={set(state.npy.keys())}, stock={set(stock_npy.keys())}"
assert set(state.numpy_inputs.keys()) == set(stock_npy.keys()) - skip_keys, \
f"Npy keys differ: v2={set(state.numpy_inputs.keys())}, stock={set(stock_npy.keys())}"
def test_split_queue_keys_work_with_desire_key(self, model_state_factory):
from openpilot.sunnypilot.modeld_v2.compile_modeld import derive_frame_skip, make_split_input_queues
@@ -188,16 +178,16 @@ class TestInputQueueCreation:
def test_npy_contains_transforms(self, archetype_name, model_state_factory):
arch = ARCHETYPES[archetype_name]
state = model_state_factory(arch)
assert 'tfm' in state.npy, f"{arch.name}: 'tfm' missing from npy"
assert 'big_tfm' in state.npy, f"{arch.name}: 'big_tfm' missing from npy"
assert state.npy['tfm'].shape == (3, 3)
assert state.npy['big_tfm'].shape == (3, 3)
assert 'tfm' in state.numpy_inputs, f"{arch.name}: 'tfm' missing from npy"
assert 'big_tfm' in state.numpy_inputs, f"{arch.name}: 'big_tfm' missing from npy"
assert state.numpy_inputs['tfm'].shape == (3, 3)
assert state.numpy_inputs['big_tfm'].shape == (3, 3)
@pytest.mark.parametrize("archetype_name", ARCHETYPE_NAMES)
def test_npy_contains_desire(self, archetype_name, model_state_factory):
arch = ARCHETYPES[archetype_name]
state = model_state_factory(arch)
assert arch.expected_desire_key in state.npy, \
assert arch.expected_desire_key in state.numpy_inputs, \
f"{arch.name}: '{arch.expected_desire_key}' missing from npy"
@@ -1,2 +0,0 @@
#!/usr/bin/env bash
clang++ -I /home/batman/one/external/tensorflow/include/ -L /home/batman/one/external/tensorflow/lib -Wl,-rpath=/home/batman/one/external/tensorflow/lib main.cc -ltensorflow
@@ -1,69 +0,0 @@
#include <cassert>
#include <cstdio>
#include <cstdlib>
#include "tensorflow/c/c_api.h"
void* read_file(const char* path, size_t* out_len) {
FILE* f = fopen(path, "r");
if (!f) {
return NULL;
}
fseek(f, 0, SEEK_END);
long f_len = ftell(f);
rewind(f);
char* buf = (char*)calloc(f_len, 1);
assert(buf);
size_t num_read = fread(buf, f_len, 1, f);
fclose(f);
if (num_read != 1) {
free(buf);
return NULL;
}
if (out_len) {
*out_len = f_len;
}
return buf;
}
static void DeallocateBuffer(void* data, size_t) {
free(data);
}
int main(int argc, char* argv[]) {
TF_Buffer* buf;
TF_Graph* graph;
TF_Status* status;
char *path = argv[1];
// load model
{
size_t model_size;
char tmp[1024];
snprintf(tmp, sizeof(tmp), "%s.pb", path);
printf("loading model %s\n", tmp);
uint8_t *model_data = (uint8_t *)read_file(tmp, &model_size);
buf = TF_NewBuffer();
buf->data = model_data;
buf->length = model_size;
buf->data_deallocator = DeallocateBuffer;
printf("loaded model of size %d\n", model_size);
}
// import graph
status = TF_NewStatus();
graph = TF_NewGraph();
TF_ImportGraphDefOptions *opts = TF_NewImportGraphDefOptions();
TF_GraphImportGraphDef(graph, buf, opts, status);
TF_DeleteImportGraphDefOptions(opts);
TF_DeleteBuffer(buf);
if (TF_GetCode(status) != TF_OK) {
printf("FAIL: %s\n", TF_Message(status));
} else {
printf("SUCCESS\n");
}
}
@@ -1,8 +0,0 @@
#!/usr/bin/env python3
import sys
import tensorflow as tf
with open(sys.argv[1], "rb") as f:
graph_def = tf.compat.v1.GraphDef()
graph_def.ParseFromString(f.read())
#tf.io.write_graph(graph_def, '', sys.argv[1]+".try")
@@ -1,38 +0,0 @@
#!/usr/bin/env python3
import os
import time
import numpy as np
import cereal.messaging as messaging
from openpilot.system.manager.process_config import managed_processes
N = int(os.getenv("N", "5"))
TIME = int(os.getenv("TIME", "30"))
if __name__ == "__main__":
sock = messaging.sub_sock('modelV2', conflate=False, timeout=1000)
execution_times = []
for _ in range(N):
os.environ['LOGPRINT'] = 'debug'
managed_processes['modeld'].start()
time.sleep(5)
t = []
start = time.monotonic()
while time.monotonic() - start < TIME:
msgs = messaging.drain_sock(sock, wait_for_one=True)
for m in msgs:
t.append(m.modelV2.modelExecutionTime)
execution_times.append(np.array(t[10:]) * 1000)
managed_processes['modeld'].stop()
print("\n\n")
print(f"ran modeld {N} times for {TIME}s each")
for _, t in enumerate(execution_times):
print(f"\tavg: {sum(t)/len(t):0.2f}ms, min: {min(t):0.2f}ms, max: {max(t):0.2f}ms")
print("\n\n")
+117 -85
View File
@@ -6,80 +6,138 @@ See the LICENSE.md file in the root directory for more details.
"""
import hashlib
import os
import pickle
from pathlib import Path
import numpy as np
from openpilot.common.params import Params
from cereal import custom
from openpilot.sunnypilot.models.constants import Meta, MetaTombRaider, MetaSimPose
from openpilot.common.params import Params
from openpilot.common.swaglog import cloudlog
from openpilot.sunnypilot.models.constants import Meta, MetaSimPose, MetaTombRaider
from openpilot.system.hardware.hw import Paths
from pathlib import Path
# see the README.md for more details on the model selector versioning
CURRENT_SELECTOR_VERSION = 15
REQUIRED_MIN_SELECTOR_VERSION = 14
# SET ME TO THE EXACT JSON VERSION WE SET IN SUNNYPILOT_MODELS REPO
REQUIRED_JSON_VERSION = 15
CUSTOM_MODEL_PATH = Paths.model_root()
METADATA_PATH = Path(__file__).parent / '../models/supercombo_metadata.pkl'
ModelManager = custom.ModelManagerSP
_LAST_VALIDATED_RAW = None
def _compute_hash(file_path: str) -> str | None:
from openpilot.common.file_chunker import read_file_chunked
try:
return hashlib.sha256(read_file_chunked(file_path)).hexdigest().lower()
except FileNotFoundError:
return None
async def verify_file(file_path: str, expected_hash: str) -> bool:
from openpilot.common.file_chunker import read_file_chunked
try:
data = read_file_chunked(file_path)
except FileNotFoundError:
return False
return hashlib.sha256(data).hexdigest().lower() == expected_hash.lower()
file_hash = _compute_hash(file_path)
return file_hash == expected_hash.lower() if file_hash else False
def _verify_file(file_path: str, expected_hash: str) -> bool:
file_hash = _compute_hash(file_path)
return file_hash == expected_hash.lower() if file_hash else False
def is_bundle_version_compatible(bundle: dict) -> bool:
"""
Checks whether the model bundle is compatible with the current selector version constraints.
The bundle specifies a `minimum_selector_version`, which defines the minimum selector version
The bundle parsed from the json specifies a `minimum_selector_version`, which defines the minimum selector version
required to load the model. This function ensures that:
1. The model is not too old: the bundle must require at least `REQUIRED_MIN_SELECTOR_VERSION`.
2. The model is not too new: it must support the current selector version (`CURRENT_SELECTOR_VERSION`).
This allows the selector to enforce both a minimum and maximum range of supported models,
even if a model would otherwise be compatible.
:param bundle: Dictionary containing `minimum_selector_version`, as defined by the model bundle.
:type bundle: Dict
:return: True if the selector version is within the accepted range for the bundle; otherwise False.
:rtype: Bool
the bundle MUST match the `REQUIRED_JSON_VERSION` set here in helpers.
"""
return bool(REQUIRED_MIN_SELECTOR_VERSION <= bundle.get("minimumSelectorVersion", 0) <= CURRENT_SELECTOR_VERSION)
return bundle.get("minimumSelectorVersion", 0) == REQUIRED_JSON_VERSION
def get_active_bundle(params: Params = None) -> custom.ModelManagerSP.ModelBundle:
"""Gets the active model bundle from cache"""
if params is None:
params = Params()
def _bundle_artifacts(bundle: custom.ModelManagerSP.ModelBundle) -> list[tuple[str, str]]:
artifacts = []
for model in getattr(bundle, 'models', []) or []:
for artifact in (getattr(model, 'artifact', None), getattr(model, 'metadata', None)):
if artifact and getattr(artifact, 'fileName', None) and getattr(artifact, 'downloadUri', None):
sha256 = getattr(artifact.downloadUri, 'sha256', None)
if sha256:
artifacts.append((artifact.fileName, sha256))
return artifacts
def _bundle_is_valid_locally(bundle: custom.ModelManagerSP.ModelBundle) -> bool:
model_root = Paths.model_root()
return all(_verify_file(os.path.join(model_root, file_name), expected_hash)
for file_name, expected_hash in _bundle_artifacts(bundle))
def _bundle_needs_reset(active_bundle: custom.ModelManagerSP.ModelBundle, available_bundles: list[custom.ModelManagerSP.ModelBundle] | None) -> bool:
if active_bundle is None:
return False
if available_bundles is not None:
matching_bundle = None
for bundle in available_bundles:
if getattr(active_bundle, 'ref', None) and getattr(bundle, 'ref', None):
if active_bundle.ref == bundle.ref:
matching_bundle = bundle
break
elif getattr(active_bundle, 'internalName', None) == getattr(bundle, 'internalName', None):
matching_bundle = bundle
break
if matching_bundle is None:
return True
if active_bundle.minimumSelectorVersion != matching_bundle.minimumSelectorVersion:
return True
active_runner = getattr(active_bundle, 'runner', None)
matching_runner = getattr(matching_bundle, 'runner', None)
if active_runner is not None and matching_runner is not None:
if getattr(active_runner, 'raw', active_runner) != getattr(matching_runner, 'raw', matching_runner):
return True
if set(_bundle_artifacts(active_bundle)) != set(_bundle_artifacts(matching_bundle)):
return True
return not _bundle_is_valid_locally(active_bundle)
def validate_active_bundle(params: Params, available_bundles: list[custom.ModelManagerSP.ModelBundle] | None = None) -> None:
global _LAST_VALIDATED_RAW
raw_bundle = params.get("ModelManager_ActiveBundle")
if not raw_bundle:
return
if raw_bundle == _LAST_VALIDATED_RAW:
return
active_bundle = get_active_bundle(params, raw_bundle_dict=raw_bundle)
if active_bundle is None or _bundle_needs_reset(active_bundle, available_bundles):
cloudlog.warning("Active model bundle invalid; resetting to default")
params.remove("ModelManager_ActiveBundle")
params.put("ModelRunnerTypeCache", int(custom.ModelManagerSP.Runner.stock), block=True)
_LAST_VALIDATED_RAW = None
else:
_LAST_VALIDATED_RAW = raw_bundle
def get_active_bundle(params: Params | None = None, raw_bundle_dict: dict | bytes | None = None) -> "custom.ModelManagerSP.ModelBundle | None":
params = params or Params()
try:
if (active_bundle := params.get("ModelManager_ActiveBundle") or {}) and is_bundle_version_compatible(active_bundle):
return custom.ModelManagerSP.ModelBundle(**active_bundle)
active_bundle_dict = raw_bundle_dict if raw_bundle_dict is not None else (params.get("ModelManager_ActiveBundle") or {})
if active_bundle_dict and is_bundle_version_compatible(active_bundle_dict):
return custom.ModelManagerSP.ModelBundle(**active_bundle_dict)
except Exception:
pass
return None
def get_active_model_runner(params: Params = None, force_check=False) -> int:
if params is None:
params = Params()
def get_active_model_runner(params: Params | None = None, force_check: bool = False) -> int:
params = params or Params()
cached_runner_type = params.get("ModelRunnerTypeCache")
if cached_runner_type is not None and not force_check:
return cached_runner_type
runner_type = custom.ModelManagerSP.Runner.stock
if active_bundle := get_active_bundle(params):
runner_type = active_bundle.runner.raw
@@ -88,66 +146,40 @@ def get_active_model_runner(params: Params = None, force_check=False) -> int:
return runner_type
def _get_model():
if bundle := get_active_bundle():
drive_model = next(model for model in bundle.models if model.type == ModelManager.Model.Type.supercombo)
return drive_model
return None
def load_metadata():
metadata_path = METADATA_PATH
if model := _get_model():
metadata_path = f"{CUSTOM_MODEL_PATH}/{model.metadata.fileName}"
def load_metadata():
model = _get_model()
metadata_path = f"{CUSTOM_MODEL_PATH}/{model.metadata.fileName}" if model else METADATA_PATH
with open(metadata_path, 'rb') as f:
return pickle.load(f)
def prepare_inputs(model_metadata) -> dict[str, np.ndarray]:
# img buffers are managed in openCL transform code so we don't pass them as inputs
inputs = {
k: np.zeros(v, dtype=np.float32).flatten()
for k, v in model_metadata['input_shapes'].items()
if 'img' not in k
def prepare_inputs(model_metadata: dict) -> dict[str, np.ndarray]:
return {
key: np.zeros(shape, dtype=np.float32).flatten()
for key, shape in model_metadata['input_shapes'].items()
if 'img' not in key
}
return inputs
def load_meta_constants(model_metadata: dict):
""" Loads the appropriate meta model class based on key shapes"""
if 'sim_pose' in model_metadata['input_shapes']:
return MetaSimPose
def load_meta_constants(model_metadata):
"""
Determines and loads the appropriate meta model class based on the metadata provided. The function checks
specific keys and conditions within the provided metadata dictionary to identify the corresponding meta
model class to return.
meta_slice = model_metadata['output_slices']['meta']
if (meta_slice.start, meta_slice.stop, meta_slice.step) == (5868, 5921, None):
return MetaTombRaider
:param model_metadata: Dictionary containing metadata about the model. It includes
details such as input shapes, output slices, and other configurations for identifying
metadata-dependent meta model classes.
:type model_metadata: dict
:return: The appropriate meta model class (Meta, MetaSimPose, or MetaTombRaider)
based on the conditions and metadata provided.
:rtype: type
"""
meta = Meta # Default Meta
if 'sim_pose' in model_metadata['input_shapes'].keys():
# Meta for models with sim_pose input
meta = MetaSimPose
else:
# Meta for Tomb Raider, it does not include sim_pose input but has the same meta slice as previous models
meta_slice = model_metadata['output_slices']['meta']
meta_tf_slice = slice(5868, 5921, None)
if (
meta_slice.start == meta_tf_slice.start and
meta_slice.stop == meta_tf_slice.stop and
meta_slice.step == meta_tf_slice.step
):
meta = MetaTombRaider
return meta
return Meta
# The following method(s) are modeld helper methods
+4 -3
View File
@@ -17,7 +17,7 @@ from openpilot.system.hardware.hw import Paths
from cereal import messaging, custom
from openpilot.sunnypilot.models.fetcher import ModelFetcher
from openpilot.sunnypilot.models.helpers import verify_file, get_active_bundle
from openpilot.sunnypilot.models.helpers import get_active_bundle, validate_active_bundle, verify_file
class ModelManagerSP:
@@ -239,6 +239,7 @@ class ModelManagerSP:
while True:
try:
self.available_models = self.model_fetcher.get_available_bundles()
validate_active_bundle(self.params, self.available_models)
self.active_bundle = get_active_bundle(self.params)
if (index_to_download := self.params.get("ModelManager_DownloadIndex")) is not None:
@@ -252,8 +253,8 @@ class ModelManagerSP:
self.selected_bundle = None
if self.params.get("ModelManager_ClearCache"):
self.clear_model_cache()
self.params.remove("ModelManager_ClearCache")
self.clear_model_cache()
self.params.remove("ModelManager_ClearCache")
self._report_status()
rk.keep_time()
-34
View File
@@ -41,7 +41,6 @@ LOCAL_PORT_WHITELIST = {8022}
SUNNYLINK_LOG_ATTR_NAME = "user.sunny.upload"
SUNNYLINK_RECONNECT_TIMEOUT_S = 70 # FYI changing this will also would require a change on sidebar.cc
DISALLOW_LOG_UPLOAD = threading.Event()
METADATA_PATH = os.path.join(os.path.dirname(os.path.dirname(os.path.abspath(__file__))), "params_metadata.json")
params = Params()
@@ -170,39 +169,6 @@ def getParamsAllKeys() -> list[str]:
return keys
@dispatcher.add_method
def getParamsAllKeysV1() -> dict[str, str]:
try:
with open(METADATA_PATH) as f:
metadata = json.load(f)
except Exception:
cloudlog.exception("sunnylinkd.getParamsAllKeysV1.metadata.exception")
metadata = {}
try:
available_keys: list[str] = [k.decode('utf-8') for k in Params().all_keys()]
params_dict: dict[str, list[dict[str, str | bool | int | object | dict | None]]] = {"params": []}
for key in available_keys:
value = get_param_as_byte(key, get_default=True)
param_entry = {
"key": key,
"type": int(params.get_type(key).value),
"default_value": base64.b64encode(value).decode('utf-8') if value else None,
}
if key in metadata:
meta_copy = metadata[key].copy()
param_entry["_extra"] = meta_copy
params_dict["params"].append(param_entry)
return {"keys": json.dumps(params_dict.get("params", []))}
except Exception:
cloudlog.exception("sunnylinkd.getParamsAllKeysV1.exception")
raise
@dispatcher.add_method
def getParamsMetadata() -> str:
"""Return settings_ui.json + live capabilities as gzip-compressed, base64-encoded string.
+5 -6
View File
@@ -97,12 +97,11 @@ The compiler splices a list-context `$ref` into its parent list. Macros may refe
```
1. common/params_keys.h — add/remove the C++ param key
2. params_metadata.json — automated via update_params_metadata.py
3. settings_ui_src/pages/<page>.yaml — add/edit/remove the item in the right section
4. python sunnypilot/sunnylink/tools/compile_settings_ui.py
5. python sunnypilot/sunnylink/tools/validate_settings_ui.py (or: --check on the compiler)
6. uv run python -m pytest sunnypilot/sunnylink/tests/ # run regression + compiler tests
7. commit
2. settings_ui_src/pages/<page>.yaml — add/edit/remove the item in the right section
3. python sunnypilot/sunnylink/tools/compile_settings_ui.py
4. python sunnypilot/sunnylink/tools/validate_settings_ui.py (or: --check on the compiler)
5. uv run python -m pytest sunnypilot/sunnylink/tests/ # run regression + compiler tests
6. commit
```
CI runs `compile_settings_ui.py --check` to fail on hand-edited `settings_ui.json`.
File diff suppressed because it is too large Load Diff
@@ -1,86 +0,0 @@
"""
Copyright (c) 2021-, Haibin Wen, sunnypilot, and a number of other contributors.
This file is part of sunnypilot and is licensed under the MIT License.
See the LICENSE.md file in the root directory for more details.
"""
import json
from openpilot.sunnypilot.sunnylink.athena.sunnylinkd import getParamsAllKeysV1, METADATA_PATH
def test_get_params_all_keys_v1():
"""
Test the getParamsAllKeysV1 API endpoint.
Why:
This endpoint is used by the UI (and potentially external tools) to fetch the list of
available parameters along with their metadata (titles, descriptions, options, constraints).
We need to ensure it returns the correct structure and that the metadata from
params_metadata.json is correctly merged into the response.
Expected:
- The response should contain a "keys" field which is a JSON string of a list of parameters.
- Each parameter object should have "key", "type", "default_value", and optionally "_extra".
- The "_extra" field should contain the rich metadata (title, options, min/max, etc.) matching
the source of truth (params_metadata.json).
"""
response = getParamsAllKeysV1()
assert "keys" in response
keys_json = response["keys"]
params_list = json.loads(keys_json)
assert isinstance(params_list, list)
assert len(params_list) > 0
# Check structure of first item
first_param = params_list[0]
assert "key" in first_param
assert "type" in first_param
assert "default_value" in first_param
if "_extra" in first_param:
assert isinstance(first_param["_extra"], dict)
assert "default" not in first_param["_extra"]
assert "type" not in first_param["_extra"]
# Load the source of truth
with open(METADATA_PATH) as f:
metadata = json.load(f)
# Verify that the API response matches the metadata file for a few sample keys
# This ensures the plumbing is working without being brittle to content changes
# 1. Check a key that should have metadata
keys_with_metadata = [k for k in params_list if k["key"] in metadata]
assert len(keys_with_metadata) > 0, "No parameters found that match metadata keys"
for param in keys_with_metadata[:5]: # Check first 5 matches
key = param["key"]
expected_meta = metadata[key]
assert "_extra" in param, f"Parameter {key} should have _extra field"
actual_meta = param["_extra"]
# Verify all fields in JSON are present in the API response
for meta_key, meta_val in expected_meta.items():
assert meta_key in actual_meta, f"Missing {meta_key} in API response for {key}"
assert actual_meta[meta_key] == meta_val, f"Mismatch for {key}.{meta_key}: expected {meta_val}, got {actual_meta[meta_key]}"
# 2. Check that we are correctly serving options if they exist
params_with_options = [k for k in keys_with_metadata if "options" in k.get("_extra", {})]
if params_with_options:
param = params_with_options[0]
key = param["key"]
assert isinstance(param["_extra"]["options"], list), f"Options for {key} should be a list"
assert param["_extra"]["options"] == metadata[key]["options"]
# 3. Check that we are correctly serving numeric constraints if they exist
params_with_constraints = [k for k in keys_with_metadata if "min" in k.get("_extra", {})]
if params_with_constraints:
param = params_with_constraints[0]
key = param["key"]
assert param["_extra"]["min"] == metadata[key]["min"]
assert param["_extra"]["max"] == metadata[key]["max"]
assert param["_extra"]["step"] == metadata[key]["step"]
@@ -1,284 +0,0 @@
"""
Copyright (c) 2021-, Haibin Wen, sunnypilot, and a number of other contributors.
This file is part of sunnypilot and is licensed under the MIT License.
See the LICENSE.md file in the root directory for more details.
"""
import json
import os
import pytest
from openpilot.common.params import Params
from openpilot.sunnypilot.sunnylink.athena.sunnylinkd import METADATA_PATH
def test_metadata_json_exists():
"""
Test that the params_metadata.json file exists at the expected path.
Why:
The metadata file is the source of truth for parameter descriptions, options, and constraints.
If it's missing, the UI will not be able to display rich information for parameters.
Expected:
The file should exist at sunnypilot/sunnylink/params_metadata.json.
"""
assert os.path.exists(METADATA_PATH), f"Metadata file not found at {METADATA_PATH}"
def test_metadata_json_valid():
"""
Test that the params_metadata.json file contains valid JSON.
Why:
Invalid JSON will cause the metadata loading to fail, potentially crashing the UI or
resulting in missing metadata.
Expected:
The file content should be parseable as a JSON object (dictionary).
"""
with open(METADATA_PATH) as f:
try:
data = json.load(f)
except json.JSONDecodeError:
pytest.fail("Metadata file is not valid JSON")
assert isinstance(data, dict), "Metadata root must be a dictionary"
def test_all_params_have_metadata():
"""
Test that every parameter in the codebase has a corresponding entry in params_metadata.json.
Why:
We want to ensure 100% coverage of parameter metadata. Any parameter added to the codebase
should also be documented in the metadata file.
Expected:
There should be no parameters in Params() that are missing from the metadata file.
If this fails, run 'python3 sunnypilot/sunnylink/tools/update_params_metadata.py'.
"""
params = Params()
all_keys = [k.decode('utf-8') for k in params.all_keys()]
with open(METADATA_PATH) as f:
metadata = json.load(f)
missing_keys = [key for key in all_keys if key not in metadata]
if missing_keys:
pytest.fail(
f"The following parameters are missing from metadata: {missing_keys}. "
+ "Please run 'python3 sunnypilot/sunnylink/tools/update_params_metadata.py' to update."
)
def test_metadata_keys_exist_in_params():
"""
Test that all keys in params_metadata.json actually exist in the codebase.
Why:
We want to avoid stale metadata for parameters that have been removed or renamed.
This keeps the metadata file clean and relevant.
Expected:
There should be no keys in the metadata file that are not present in Params().
This prints a warning rather than failing, as it's less critical than missing metadata.
"""
params = Params()
all_keys = {k.decode('utf-8') for k in params.all_keys()}
with open(METADATA_PATH) as f:
metadata = json.load(f)
extra_keys = [key for key in metadata.keys() if key not in all_keys]
if extra_keys:
print(f"Warning: The following keys in metadata do not exist in Params: {extra_keys}")
def test_no_default_titles():
"""
Test that no parameter has a title that is identical to its key.
Why:
The default behavior of the update script is to set the title equal to the key.
We want to force developers to provide human-readable, descriptive titles for all parameters.
Expected:
No parameter metadata should have 'title' == 'key'.
"""
with open(METADATA_PATH) as f:
metadata = json.load(f)
default_title_keys = [key for key, meta in metadata.items() if meta.get("title") == key]
if default_title_keys:
pytest.fail(
f"The following parameters have default titles (title == key): {default_title_keys}. "
+ "Please update 'params_metadata.json' with descriptive titles."
)
def test_options_structure():
"""
Test that the 'options' field in metadata follows the correct structure.
Why:
The UI expects 'options' to be a list of objects with 'value' and 'label' keys.
Incorrect structure will break the UI rendering for dropdowns/toggles.
Expected:
If 'options' is present, it must be a list of dicts, and each dict must have 'value' and 'label'.
"""
with open(METADATA_PATH) as f:
metadata = json.load(f)
for key, meta in metadata.items():
if "options" in meta:
options = meta["options"]
assert isinstance(options, list), f"Options for {key} must be a list"
for option in options:
assert isinstance(option, dict), f"Option in {key} must be a dictionary"
assert "value" in option, f"Option in {key} must have a 'value' key"
assert "label" in option, f"Option in {key} must have a 'label' key"
def test_numeric_constraints():
"""
Test that numeric parameters have valid 'min', 'max', and 'step' constraints.
Why:
The UI uses these constraints to validate user input and render sliders/steppers.
Missing or invalid constraints can lead to UI bugs or invalid parameter values.
Expected:
If any of min/max/step is present, ALL of them must be present.
They must be numbers (int/float), and min must be less than max.
"""
with open(METADATA_PATH) as f:
metadata = json.load(f)
for key, meta in metadata.items():
if "min" in meta or "max" in meta or "step" in meta:
assert "min" in meta, f"Numeric param {key} must have 'min'"
assert "max" in meta, f"Numeric param {key} must have 'max'"
assert "step" in meta, f"Numeric param {key} must have 'step'"
assert isinstance(meta["min"], (int, float)), f"Min for {key} must be number"
assert isinstance(meta["max"], (int, float)), f"Max for {key} must be number"
assert isinstance(meta["step"], (int, float)), f"Step for {key} must be number"
assert meta["min"] < meta["max"], f"Min must be less than max for {key}"
def test_known_params_metadata():
"""
Test specific known parameters to ensure they have the expected rich metadata.
Why:
This acts as a spot check to ensure that our rich metadata population logic is working correctly
and that critical parameters (like LongitudinalPersonality) have their options and constraints preserved.
Expected:
'LongitudinalPersonality' should have 3 options (Aggressive, Standard, Relaxed).
'CustomAccLongPressIncrement' should have min=1, max=10, step=1.
"""
with open(METADATA_PATH) as f:
metadata = json.load(f)
# Check an enum-like param
lp = metadata.get("LongitudinalPersonality")
assert lp is not None
assert "options" in lp
assert len(lp["options"]) == 3
assert lp["options"][0]["label"] == "Aggressive"
assert lp["options"][0]["value"] == 0
# Check a numeric param
acc_long = metadata.get("CustomAccLongPressIncrement")
assert acc_long is not None
assert acc_long["min"] == 1
assert acc_long["max"] == 10
assert acc_long["step"] == 1
def test_torque_control_tune_versions_in_sync():
"""
Test that TorqueControlTune options in params_metadata.json match versions in latcontrol_torque_versions.json.
Why:
The TorqueControlTune dropdown in the UI should always reflect the available torque tune versions.
If versions are added/removed from latcontrol_torque_versions.json, the metadata must be updated accordingly.
Expected:
- TorqueControlTune should have a 'Default' option with empty string value
- All versions from latcontrol_torque_versions.json should be present in the options
- The version values and labels should match between both files
"""
from openpilot.common.basedir import BASEDIR
versions_json_path = os.path.join(BASEDIR, "sunnypilot", "selfdrive", "controls", "lib", "latcontrol_torque_versions.json")
sync_script_path = "python3 sunnypilot/sunnylink/tools/sync_torque_versions.py"
# Load both files
with open(METADATA_PATH) as f:
metadata = json.load(f)
with open(versions_json_path) as f:
versions = json.load(f)
# Get TorqueControlTune metadata
torque_tune = metadata.get("TorqueControlTune")
if torque_tune is None:
pytest.fail(f"TorqueControlTune not found in params_metadata.json. Please run '{sync_script_path}' to sync.")
if "options" not in torque_tune:
pytest.fail(f"TorqueControlTune must have options. Please run '{sync_script_path}' to sync.")
options = torque_tune["options"]
if not isinstance(options, list):
pytest.fail(f"TorqueControlTune options must be a list. Please run '{sync_script_path}' to sync.")
if len(options) == 0:
pytest.fail(f"TorqueControlTune must have at least one option. Please run '{sync_script_path}' to sync.")
# Check that Default option exists
default_option = next((opt for opt in options if opt.get("value") == ""), None)
if default_option is None:
pytest.fail(f"TorqueControlTune must have a 'Default' option with empty string value. Please run '{sync_script_path}' to sync.")
if default_option.get("label") != "Default":
pytest.fail(f"Default option must have label 'Default'. Please run '{sync_script_path}' to sync.")
# Build expected options from versions.json
expected_version_keys = set(versions.keys())
actual_version_keys = set()
for option in options:
if option.get("value") == "":
continue # Skip the default option
label = option.get("label")
value = option.get("value")
# Check that this option corresponds to a version
if label not in versions:
pytest.fail(f"Option label '{label}' not found in latcontrol_torque_versions.json. Please run '{sync_script_path}' to sync.")
# Check that the value matches the version number
expected_value = float(versions[label]["version"])
if value != expected_value:
pytest.fail(f"Option '{label}' has value {value}, expected {expected_value}. Please run '{sync_script_path}' to sync.")
actual_version_keys.add(label)
# Check that all versions are represented
missing_versions = expected_version_keys - actual_version_keys
if missing_versions:
pytest.fail(f"The following versions are missing from TorqueControlTune options: {missing_versions}. " +
f"Please run '{sync_script_path}' to sync.")
extra_versions = actual_version_keys - expected_version_keys
if extra_versions:
pytest.fail("The following versions in TorqueControlTune options are not in latcontrol_torque_versions.json: " +
f"{extra_versions}. Please run '{sync_script_path}' to sync.")
@@ -1,133 +0,0 @@
#!/usr/bin/env python3
"""
Copyright (c) 2021-, Haibin Wen, sunnypilot, and a number of other contributors.
This file is part of sunnypilot and is licensed under the MIT License.
See the LICENSE.md file in the root directory for more details.
"""
import json
import os
from openpilot.common.basedir import BASEDIR
from openpilot.common.params import Params
from openpilot.sunnypilot.system.params_migration import ONROAD_BRIGHTNESS_TIMER_VALUES
METADATA_PATH = os.path.join(os.path.dirname(__file__), "../params_metadata.json")
TORQUE_VERSIONS_JSON = os.path.join(BASEDIR, "sunnypilot", "selfdrive", "controls", "lib", "latcontrol_torque_versions.json")
def main():
params = Params()
all_keys = params.all_keys()
if os.path.exists(METADATA_PATH):
with open(METADATA_PATH) as f:
try:
data = json.load(f)
except json.JSONDecodeError:
data = {}
else:
data = {}
# Add new keys
for key in all_keys:
key_str = key.decode("utf-8")
if key_str not in data:
print(f"Adding new key: {key_str}")
data[key_str] = {
"title": key_str,
"description": "",
}
# Remove deleted keys
# keys_to_remove = [k for k in data.keys() if k.encode("utf-8") not in all_keys]
# for k in keys_to_remove:
# print(f"Removing deleted key: {k}")
# del data[k]
# Sort keys
sorted_data = dict(sorted(data.items()))
with open(METADATA_PATH, "w") as f:
json.dump(sorted_data, f, indent=2)
f.write("\n")
print(f"Updated {METADATA_PATH}")
# update onroad screen brightness params
update_onroad_brightness_param()
# update onroad screen brightness timer params
update_onroad_brightness_timer_param()
# update torque versions param
update_torque_versions_param()
def update_onroad_brightness_param():
try:
with open(METADATA_PATH) as f:
params_metadata = json.load(f)
if "OnroadScreenOffBrightness" in params_metadata:
options = [
{"value": 0, "label": "Auto (Default)"},
{"value": 1, "label": "Auto (Dark)"},
{"value": 2, "label": "Screen Off"},
]
for i in range(3, 23):
options.append({"value": i, "label": f"{(i - 2) * 5} %"})
params_metadata["OnroadScreenOffBrightness"]["options"] = options
with open(METADATA_PATH, 'w') as f:
json.dump(params_metadata, f, indent=2)
f.write('\n')
print(f"Updated OnroadScreenOffBrightness options in params_metadata.json with {len(options)} options.")
except Exception as e:
print(f"Failed to update OnroadScreenOffBrightness versions in params_metadata.json: {e}")
def update_onroad_brightness_timer_param():
try:
with open(METADATA_PATH) as f:
params_metadata = json.load(f)
if "OnroadScreenOffTimer" in params_metadata:
options = []
for _index, seconds in sorted(ONROAD_BRIGHTNESS_TIMER_VALUES.items()):
label = f"{seconds}s" if seconds < 60 else f"{seconds // 60}m"
options.append({"value": seconds, "label": label})
params_metadata["OnroadScreenOffTimer"]["options"] = options
with open(METADATA_PATH, 'w') as f:
json.dump(params_metadata, f, indent=2)
f.write('\n')
print(f"Updated OnroadScreenOffTimer options in params_metadata.json with {len(options)} options.")
except Exception as e:
print(f"Failed to update OnroadScreenOffTimer options in params_metadata.json: {e}")
def update_torque_versions_param():
with open(TORQUE_VERSIONS_JSON) as f:
current_versions = json.load(f)
try:
with open(METADATA_PATH) as f:
params_metadata = json.load(f)
options = [{"value": "", "label": "Default"}]
for version_key, version_data in current_versions.items():
version_value = float(version_data["version"])
options.append({"value": version_value, "label": str(version_key)})
if "TorqueControlTune" in params_metadata:
params_metadata["TorqueControlTune"]["options"] = options
with open(METADATA_PATH, 'w') as f:
json.dump(params_metadata, f, indent=2)
f.write('\n')
print(f"Updated TorqueControlTune options in params_metadata.json with {len(options)} options: \n{options}")
except Exception as e:
print(f"Failed to update TorqueControlTune versions in params_metadata.json: {e}")
if __name__ == "__main__":
main()
Generated
+22 -22
View File
@@ -1369,15 +1369,15 @@ wheels = [
[[package]]
name = "sentry-sdk"
version = "2.61.1"
version = "2.62.0"
source = { registry = "https://pypi.org/simple" }
dependencies = [
{ name = "certifi" },
{ name = "urllib3" },
]
sdist = { url = "https://files.pythonhosted.org/packages/63/3b/4bc6b348bbd331daa14d4babe9f2b99bc854f4da41560eefb9488d78481d/sentry_sdk-2.61.1.tar.gz", hash = "sha256:9c6adccb3feefa9ba032c8d295ca477575c2f11896046a2b0ad686c47c4af555", size = 459429, upload-time = "2026-06-01T07:24:18.875Z" }
sdist = { url = "https://files.pythonhosted.org/packages/f6/5d/a343201726150e05f2036eeb6e493e2e2f8bf8a66f5aa70f2f4ac96f9ca3/sentry_sdk-2.62.0.tar.gz", hash = "sha256:3c870b9f50d9fd15b58c817dbde1c7cfaa9fe3f05df0a4c6edd5571cb82f5491", size = 463986, upload-time = "2026-06-08T13:23:49.223Z" }
wheels = [
{ url = "https://files.pythonhosted.org/packages/df/54/c9218db183846e08efaf68534889ef42e499dde432778881104a42f7071b/sentry_sdk-2.61.1-py3-none-any.whl", hash = "sha256:fa36eaf4b8ad708f718500d4bdcc1532637526a22beb874d88cbc0a46458b5ae", size = 483735, upload-time = "2026-06-01T07:24:17.027Z" },
{ url = "https://files.pythonhosted.org/packages/3d/07/05440381627877aae223fd68f330df9b9fc6641d08bf65328b55235617a2/sentry_sdk-2.62.0-py3-none-any.whl", hash = "sha256:27f61d13a86c3c1648dec666dd5a64f79772dd6a84b446f11866601ecab24f6f", size = 490586, upload-time = "2026-06-08T13:23:47.486Z" },
]
[[package]]
@@ -1491,27 +1491,27 @@ wheels = [
[[package]]
name = "ty"
version = "0.0.44"
version = "0.0.46"
source = { registry = "https://pypi.org/simple" }
sdist = { url = "https://files.pythonhosted.org/packages/13/f4/fbb120226e4f239652525a664bad976a23fea58c646d1323f2296fee8a61/ty-0.0.44.tar.gz", hash = "sha256:5886229830ab77022842a1c55d2ef57405621a91fc465969fa6d538661898173", size = 5803665, upload-time = "2026-06-05T03:33:48.612Z" }
sdist = { url = "https://files.pythonhosted.org/packages/5a/7d/d95b5a9dea83472006be3ce5e480028c44b34138d84d0172e910f287fb69/ty-0.0.46.tar.gz", hash = "sha256:c6c2d7105b5633b49950b4c3a90d1ed2613eb9d794ad582bbbf6c4ffcb93accf", size = 5832380, upload-time = "2026-06-09T03:28:05.056Z" }
wheels = [
{ url = "https://files.pythonhosted.org/packages/e8/c6/b5b8c4762efb4d85401652658786506867553ecfc2beac3bcf361a15937f/ty-0.0.44-py3-none-linux_armv6l.whl", hash = "sha256:272d31e7ad49b1dc5e8465a9fe700354e14c755b40d9c75f08f031d786903df3", size = 11607267, upload-time = "2026-06-05T03:33:27.154Z" },
{ url = "https://files.pythonhosted.org/packages/1c/5c/f4b405570737f44ab0fc4214117fe43353f8f0825a1823d9e99e9c8e57be/ty-0.0.44-py3-none-macosx_10_12_x86_64.whl", hash = "sha256:b92c4ddd7a3daf2049715edec9dc70cf6fd31a5a318ee647258f90dd75495eed", size = 11382826, upload-time = "2026-06-05T03:33:54.374Z" },
{ url = "https://files.pythonhosted.org/packages/9d/aa/fb9835aa492b148d7754cb4c3db07f31a7e2e09f0d8e0e8e297f01125dd2/ty-0.0.44-py3-none-macosx_11_0_arm64.whl", hash = "sha256:4d42cfd84a690f6654b2a4f0515027c21b692cf2512d32e6433f754893a95609", size = 10809741, upload-time = "2026-06-05T03:33:33.22Z" },
{ url = "https://files.pythonhosted.org/packages/47/f5/0b20ba6b66837a5a37bab7f74ac0732c66e766b5f0b2d55b30816b15f348/ty-0.0.44-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5dc47ae87e4cb7db2a9166bb23b78a905c3626e523296ec5bccf36b5e89bda6b", size = 11318153, upload-time = "2026-06-05T03:34:09.403Z" },
{ url = "https://files.pythonhosted.org/packages/ca/bb/b82ea730774a4f950f06d355fbc120d51eac7da23b57fc79ef6ff7c79cbb/ty-0.0.44-py3-none-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:46d867e80f16f421ac72c9a85240dbf050d62d9b3fbd10a8b5b082fb21679e0b", size = 11403108, upload-time = "2026-06-05T03:33:57.745Z" },
{ url = "https://files.pythonhosted.org/packages/8b/41/e2c83856165291049c702eda4e2ef3d3ebd875e8a0a77b8cc4ef3156aa1c/ty-0.0.44-py3-none-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:411f5de0f96a4e4e5cccc3e0d55954c41f6a99ee6ca1fe5a7226cbc68406e053", size = 11944815, upload-time = "2026-06-05T03:34:15.793Z" },
{ url = "https://files.pythonhosted.org/packages/66/95/1fa6a101eb9d5bec042b87e5ca9c8fc349b75961beca6306f95af5cd5539/ty-0.0.44-py3-none-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:4b15f01ecb4e2b46c05a1769293f9d32c3d4a1e4e7dfccf37c604d705dc3e3f4", size = 12476121, upload-time = "2026-06-05T03:33:51.529Z" },
{ url = "https://files.pythonhosted.org/packages/72/6a/da4b45b1229d39207c6140681c2aaf4f5691bcb1dc830b84450ca25c8f57/ty-0.0.44-py3-none-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:edd32b7467af509c99c0244c2226a4e4c03400699003ec33373282ab931654d9", size = 12091340, upload-time = "2026-06-05T03:33:36.289Z" },
{ url = "https://files.pythonhosted.org/packages/16/c7/e1c9260ea5188195962ff1214ace418b5d69187e8fa7c0a1ec4994b8071b/ty-0.0.44-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:503a585f4007387c3afc58bae23a7ca1b9f236cbdb1a881dc36110655ceb1937", size = 11986201, upload-time = "2026-06-05T03:34:00.624Z" },
{ url = "https://files.pythonhosted.org/packages/92/f9/312bb112da9b1a7da295bb0426be85e72ad48da4e4266c36d77256b4058d/ty-0.0.44-py3-none-manylinux_2_31_riscv64.whl", hash = "sha256:2d28bcfa83243d77c2316944e8cf197f73597bf17d1ddc047d0b10a762531252", size = 12168475, upload-time = "2026-06-05T03:33:30.386Z" },
{ url = "https://files.pythonhosted.org/packages/02/de/64978d603f6c3e5dd7cb97eca2214567d8ad0c85fa4a7435b7852ae4b779/ty-0.0.44-py3-none-musllinux_1_2_aarch64.whl", hash = "sha256:56fd2dd0192def189715b25f5338f6222fb827884dc34111e50aa1c4e061cee5", size = 11292937, upload-time = "2026-06-05T03:34:06.448Z" },
{ url = "https://files.pythonhosted.org/packages/64/63/a625d8a3c71dcaa01988d330f849c465fe72ead4b0bbab44fe4bd6e672b5/ty-0.0.44-py3-none-musllinux_1_2_armv7l.whl", hash = "sha256:7f8d990489032de1984e73c159f3e760d754cf83a602b874827d943821f63595", size = 11421560, upload-time = "2026-06-05T03:33:23.995Z" },
{ url = "https://files.pythonhosted.org/packages/99/96/61aeba0e629b0c91bd316ff94d00e38817ec493ae4316f39508988daa287/ty-0.0.44-py3-none-musllinux_1_2_i686.whl", hash = "sha256:f61ffe72996a755432922fe90b28db593f572eb5cbf48e3ef4e67b282533d1b0", size = 11580282, upload-time = "2026-06-05T03:34:03.308Z" },
{ url = "https://files.pythonhosted.org/packages/fa/f7/256e1538ce21cab67b381201444c42454de69d310059c4929d92a0ee9c48/ty-0.0.44-py3-none-musllinux_1_2_x86_64.whl", hash = "sha256:2b237a143bac4f30cec9257d45f01e72da97030a80a09a2b69cfef065f09c37f", size = 12085723, upload-time = "2026-06-05T03:33:45.953Z" },
{ url = "https://files.pythonhosted.org/packages/d3/76/ec3957c10872643a98db7a7895101ad89c5b7cba4bc6c4aebbbfc91756cc/ty-0.0.44-py3-none-win32.whl", hash = "sha256:6a24586c65419223ac5bab4822d49ab493a5d19ea2a897514284c232b9d6166a", size = 10892978, upload-time = "2026-06-05T03:34:12.603Z" },
{ url = "https://files.pythonhosted.org/packages/a5/7d/ba24050432196e7d7f03945e5c379951593c48e04e5c5d5275cfc4624791/ty-0.0.44-py3-none-win_amd64.whl", hash = "sha256:8cccb27e348c89a9733fbad1b2efadfbad79b107c7e52adb52dfd8a70156a38d", size = 11987058, upload-time = "2026-06-05T03:33:42.692Z" },
{ url = "https://files.pythonhosted.org/packages/71/34/16ec3f1fec75292d9c56a8b5fef037ceaba85a5c30562206c1a245a00a67/ty-0.0.44-py3-none-win_arm64.whl", hash = "sha256:58049504e7a12bf1957f24a5384a332c94d5590127083a80db5e5a1bed34190b", size = 11329961, upload-time = "2026-06-05T03:33:39.427Z" },
{ url = "https://files.pythonhosted.org/packages/0e/24/f9f7533c391610521f4164e6b8e37ef72d0c1ee8651bc0d9ce9e658b953b/ty-0.0.46-py3-none-linux_armv6l.whl", hash = "sha256:5e716337994699cbc1a1a7b7a3e6622306f2574c710330f9d9691c2c3d8391b0", size = 11756264, upload-time = "2026-06-09T03:28:20.112Z" },
{ url = "https://files.pythonhosted.org/packages/66/49/ff3d13655b9b5cc8176f4c3446bf7ec2df43c8ad9e5272d4adc5d952fa45/ty-0.0.46-py3-none-macosx_10_12_x86_64.whl", hash = "sha256:51d618dec5403635690d0e3e298cd0ad3d84ebc6a576652939ef30ce96fce4b2", size = 11492723, upload-time = "2026-06-09T03:28:13.23Z" },
{ url = "https://files.pythonhosted.org/packages/82/4a/e7e3209e353c5835c7756339bbcdfda10852407b80fbb9ed46c17241873a/ty-0.0.46-py3-none-macosx_11_0_arm64.whl", hash = "sha256:acbafd6a2351b07a6cf4c945b0b1d47f6d2826faac2526a351dfa74d3a3cc664", size = 10892822, upload-time = "2026-06-09T03:27:51.179Z" },
{ url = "https://files.pythonhosted.org/packages/6c/20/4390c90434a9ddefcecb65e8df00e4c2700e9739dc0baf58bed36d25f713/ty-0.0.46-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:de5df602ffd760612ae36602bbad69b0123ff6cffd92e62aa92b7709317d69e3", size = 11408745, upload-time = "2026-06-09T03:27:58.049Z" },
{ url = "https://files.pythonhosted.org/packages/75/0c/f13a1bf9c6798530c773667095a6cf8f73ec9721db359423e7249bff7fbc/ty-0.0.46-py3-none-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:7abf5a10b30d8641faad90f6a19989daec941bb90261159e05cfeb04d2012046", size = 11544432, upload-time = "2026-06-09T03:27:53.519Z" },
{ url = "https://files.pythonhosted.org/packages/56/69/eb3710c13dff846a0362df04fadd8a39b64ccc244c0d02ce5285ede8eae5/ty-0.0.46-py3-none-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:8770404139c6ccee2ce2fc226478cfa4100915133c876c257e52197b8b92051d", size = 12031228, upload-time = "2026-06-09T03:28:29.816Z" },
{ url = "https://files.pythonhosted.org/packages/e9/68/5f5db9c84c1d44acdc67281089b372d9d818ee68123a60c59c66187095e2/ty-0.0.46-py3-none-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:f960d5a6e4860076924d2b86891d9872c4a3daa4663fb416e640b22cf3dbf68e", size = 12596073, upload-time = "2026-06-09T03:28:25.204Z" },
{ url = "https://files.pythonhosted.org/packages/14/be/cfd0bb272e6a1491f6de30c60da1f39c2b3c3524ec64a5c92b71365c9185/ty-0.0.46-py3-none-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:1d9000a4a3ed08fc37e8a2ff0b801cde06e1c2af3bc053677744bb5a1b751030", size = 12284885, upload-time = "2026-06-09T03:28:10.58Z" },
{ url = "https://files.pythonhosted.org/packages/a8/3a/2cd541f6320f5d6f70a45725c4e1016efedd5545348bb23b47ffb3e4c724/ty-0.0.46-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d1160e6dc86536109ab755f7142f36f4dda5333c8330cf230d61819494d27125", size = 12079480, upload-time = "2026-06-09T03:27:55.847Z" },
{ url = "https://files.pythonhosted.org/packages/de/91/8e0075bc6568fb477e7ef4d805c67fa6902b692cb4419e0bf5ce3c04c5bc/ty-0.0.46-py3-none-manylinux_2_31_riscv64.whl", hash = "sha256:b619c0efe007731f8221fa787701bfa4402da7a83eb26c61ae25e77b6ace6384", size = 12316547, upload-time = "2026-06-09T03:28:08.28Z" },
{ url = "https://files.pythonhosted.org/packages/00/28/b96cbfeda019a4044c6a8cd06ff84d08b631d4ba7d9a1e6dc0311df3563a/ty-0.0.46-py3-none-musllinux_1_2_aarch64.whl", hash = "sha256:ad98fccb6a8a94c4121b993761a0deee602f5826c4162e0a91f4f8118ddadd42", size = 11392846, upload-time = "2026-06-09T03:28:00.418Z" },
{ url = "https://files.pythonhosted.org/packages/3b/d0/4d77f699a95ac7a13b94ca1a58682667cfe974f91557d9e2a9fc0b808a7f/ty-0.0.46-py3-none-musllinux_1_2_armv7l.whl", hash = "sha256:74536b13c3cc3f5944408669c202d4c57c3d19ff154732df8e6145718aef9191", size = 11559017, upload-time = "2026-06-09T03:28:17.619Z" },
{ url = "https://files.pythonhosted.org/packages/88/62/1d6f6b51c2b132da8011c6a41ead0c1fd2a0b17ea72304bcf6ce084d581a/ty-0.0.46-py3-none-musllinux_1_2_i686.whl", hash = "sha256:5e50b1e96ced41b609e24ed27d9e4f508584ed7f4d0bb717ca8c8d75d2fd1b7c", size = 11666509, upload-time = "2026-06-09T03:28:22.454Z" },
{ url = "https://files.pythonhosted.org/packages/fe/9a/6643894bc12cb30c281f4c8bf37f6d30c1fbd9484ef39a12b0ea6dae3c1c/ty-0.0.46-py3-none-musllinux_1_2_x86_64.whl", hash = "sha256:0a7d9f58d26d938e5d2f607481b7a412d8c00d675a1ec72004fa9d6b3b9def99", size = 12180448, upload-time = "2026-06-09T03:28:32.329Z" },
{ url = "https://files.pythonhosted.org/packages/86/68/0f3b7bb03a7da676ef51b1c0af0bde1e500d69d5f0c807ed63b6f30b66dd/ty-0.0.46-py3-none-win32.whl", hash = "sha256:26db0ce89c573e60132d14e9688c9329a1633b1a8c26fe457025c7c406f7d5e6", size = 10960002, upload-time = "2026-06-09T03:28:02.832Z" },
{ url = "https://files.pythonhosted.org/packages/b1/f4/91ff618b2dee39d0633d23e1adac0174aa1de80df17e270acac534034dbc/ty-0.0.46-py3-none-win_amd64.whl", hash = "sha256:90e8e6d446b9cb7cb4bede9fca7b3c99fd1e2355605ecf431c131a51db2a5e93", size = 12097413, upload-time = "2026-06-09T03:28:27.495Z" },
{ url = "https://files.pythonhosted.org/packages/e5/2e/300174fca375a27a7c28dd80e990d857d7b3e3b25980c65063f980aa2f17/ty-0.0.46-py3-none-win_arm64.whl", hash = "sha256:ebd320d82605079b901a095dc4711037a0c488b4ace79a602fef4df0d3f4cf74", size = 11439595, upload-time = "2026-06-09T03:28:15.355Z" },
]
[[package]]