File size: 16,994 Bytes
93065ec | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 | """React: short-horizon contact-rich window dataset over pre-sliced segments.
This is the `mode2_v1/` companion to `react_window_dataset.py`. The
difference is purely architectural:
- `ReactWindowDataset` operates on `processed/mode1_v1/<task>/<date>/episode_*.pt`,
which include bad intervals (LED flicker, pose teleports, OT track loss),
and uses `bad_frames.json` to skip windows that overlap a flagged span.
- `ReactSegmentDataset` (this file) operates on
`processed/mode2_v1/<task>/<date>/episode_*.segment_*.pt`, where every
`.pt` is already a contiguous clean span. No `bad_frames.json` lookup is
needed; the data is *constructively* clean.
Filters that remain (these are about what kind of window you want, not
whether the data is good):
1. `respect_active_sensors` — ignore inactive sensors per `tasks.json`
2. `min_contact_fraction` — drop windows below the contact threshold
3. `require_motion` — drop "operator paused" windows
The per-segment `_contact_meta.source_h5_frame_range` lets you map any
window back to its position in the original H5 recording, e.g. for
inspection in `twm.visualize` or for cross-referencing with the original
H5 archive.
Usage
-----
```python
from react_segment_dataset import ReactSegmentDataset
from torch.utils.data import DataLoader
ds = ReactSegmentDataset(
segments_root = "processed/mode2_v1/motherboard",
tasks_json_path= "tasks.json",
window_length = 16,
stride = 1,
window_step = 8,
# contact filter
contact_metric = "mixed",
tactile_threshold = 0.4,
min_contact_fraction = 0.5,
which_sensors = "both",
# motion filter (recommended for dynamics learning)
require_motion = True,
min_motion_mps = 0.01,
min_motion_fraction = 0.25,
which_sensors_must_move = "all_active",
)
loader = DataLoader(ds, batch_size=8, shuffle=True, num_workers=2)
```
Each sample is the same dict as `ReactWindowDataset`, plus the segment
provenance (`source_episode`, `source_segment_idx`, `source_h5_frame_range`).
"""
from __future__ import annotations
import json
from pathlib import Path
from typing import Iterable
import numpy as np
import torch
from torch.utils.data import Dataset
CONTACT_METRICS = ("intensity", "area", "mixed")
def _per_frame_speed_mps(pose7: np.ndarray, fps: float = 30.0) -> np.ndarray:
if pose7.shape[0] < 2:
return np.zeros(pose7.shape[0], dtype=np.float64)
d = np.linalg.norm(np.diff(pose7[:, :3], axis=0), axis=1) * fps
out = np.empty(pose7.shape[0], dtype=np.float64)
out[0] = d[0]
out[1:] = d
return out
class ReactSegmentDataset(Dataset):
"""Per-window dataset over the pre-sliced React segments.
Parameters
----------
segments_root : path
Directory containing `<task>/<date>/episode_*.segment_*.pt`
(searched recursively). e.g. `processed/mode2_v1/motherboard`.
tasks_json_path : optional path
Path to `tasks.json`. Used for the `respect_active_sensors` mode.
window_length, stride, window_step : window enumeration
contact_metric, tactile_threshold, min_contact_fraction, which_sensors :
contact filter parameters
respect_active_sensors : bool, default True
require_motion, min_motion_mps, min_motion_fraction,
which_sensors_must_move : motion filter parameters
tasks, dates : optional iterables of str
Restrict to specific task / date strings.
fps : float, default 30
"""
def __init__(
self,
segments_root: str | Path,
tasks_json_path: str | Path | None = None,
*,
window_length: int = 16,
stride: int = 1,
window_step: int | None = None,
contact_metric: str = "mixed",
tactile_threshold: float = 0.4,
min_contact_fraction: float = 0.5,
which_sensors: str = "both",
tasks: Iterable[str] | None = None,
dates: Iterable[str] | None = None,
respect_active_sensors: bool = True,
require_motion: bool = False,
min_motion_mps: float = 0.01,
min_motion_fraction: float = 0.25,
which_sensors_must_move: str = "all_active",
fps: float = 30.0,
):
if contact_metric not in CONTACT_METRICS:
raise ValueError(f"contact_metric must be one of {CONTACT_METRICS}")
if which_sensors not in ("any", "both", "left", "right"):
raise ValueError("which_sensors must be 'any' | 'both' | 'left' | 'right'")
if which_sensors_must_move not in ("any", "all_active"):
raise ValueError("which_sensors_must_move must be 'any' | 'all_active'")
if window_length < 1 or stride < 1:
raise ValueError("window_length and stride must be ≥ 1")
self.segments_root = Path(segments_root)
self.window_length = int(window_length)
self.stride = int(stride)
self.window_step = int(window_step) if window_step is not None else max(1, window_length // 2)
self.contact_metric = contact_metric
self.tactile_threshold = float(tactile_threshold)
self.min_contact_fraction = float(min_contact_fraction)
self.which_sensors = which_sensors
self.respect_active_sensors = bool(respect_active_sensors)
self.require_motion = bool(require_motion)
self.min_motion_mps = float(min_motion_mps)
self.min_motion_fraction = float(min_motion_fraction)
self.which_sensors_must_move = which_sensors_must_move
self.fps = float(fps)
# Per-date active-sensor info
self.per_date = {}
if tasks_json_path is not None and Path(tasks_json_path).is_file():
tj = json.loads(Path(tasks_json_path).read_text())
for tk, td in tj.get("tasks", {}).items():
for d, info in td.get("per_date_notes", {}).items():
self.per_date[d] = info
# Discover segments
pt_files = sorted(self.segments_root.rglob("episode_*.segment_*.pt"))
if not pt_files:
raise RuntimeError(f"No segment .pt files under {self.segments_root}")
tasks_set = set(tasks) if tasks is not None else None
dates_set = set(dates) if dates is not None else None
self.segments: list[dict] = [] # cached .pt dicts
self.segment_paths: list[Path] = []
self.segment_meta: list[dict] = [] # source_episode, segment_idx, etc.
self.segment_active: list[list[str]] = []
self.windows: list[tuple[int, int]] = [] # (seg_idx, t_start)
span = (self.window_length - 1) * self.stride + 1
n_total_candidates = 0
n_drop_contact = 0
n_drop_motion = 0
for pt in pt_files:
rel = pt.relative_to(self.segments_root)
# rel.parts == (<task>, <date>, "episode_NNN.segment_MM.pt") OR (<date>, ...)
if len(rel.parts) == 3:
task, date, _ = rel.parts
elif len(rel.parts) == 2:
task, date = None, rel.parts[0]
else:
task, date = None, None
if tasks_set is not None and task not in tasks_set:
continue
if dates_set is not None and date not in dates_set:
continue
d = torch.load(pt, weights_only=False, map_location="cpu")
meta = d.get("_contact_meta", {})
src_ep = meta.get("source_episode") or rel.stem.split(".")[0]
seg_idx = int(meta.get("source_segment_idx", 0))
active = ["left", "right"]
if self.respect_active_sensors and date in self.per_date:
active = list(self.per_date[date].get("active_sensors", active))
mL = d[f"tactile_left_{self.contact_metric}"].numpy()
mR = d[f"tactile_right_{self.contact_metric}"].numpy()
T = mL.shape[0]
if T < span:
continue # segment too short to host any window
# Contact predicate
cL = mL > self.tactile_threshold
cR = mR > self.tactile_threshold
if "left" not in active: cL[:] = False
if "right" not in active: cR[:] = False
req = self.which_sensors
if req == "any": contact_frame = cL | cR
elif req == "both": contact_frame = cL & cR
elif req == "left": contact_frame = cL
else: contact_frame = cR
# Per-frame motion mask
if self.require_motion:
speed_L = _per_frame_speed_mps(d["sensor_left_pose"].numpy(), self.fps)
speed_R = _per_frame_speed_mps(d["sensor_right_pose"].numpy(), self.fps)
moving_L = speed_L >= self.min_motion_mps
moving_R = speed_R >= self.min_motion_mps
else:
moving_L = moving_R = None
seg_id = len(self.segments)
self.segments.append(d)
self.segment_paths.append(pt)
self.segment_meta.append({
"source_episode": src_ep,
"source_segment_idx": seg_idx,
"source_h5_frame_range": meta.get("source_h5_frame_range"),
"source_pt_frame_range": meta.get("source_pt_frame_range"),
})
self.segment_active.append(active)
kept = 0
for t_start in range(0, T - span + 1, self.window_step):
n_total_candidates += 1
t_end = t_start + span - 1
frame_idx = np.arange(t_start, t_start + span, self.stride)
if contact_frame[frame_idx].mean() < self.min_contact_fraction:
n_drop_contact += 1
continue
if self.require_motion:
passed = []
for side, mov in [("left", moving_L), ("right", moving_R)]:
if side not in active:
continue
passed.append(mov[frame_idx].mean() >= self.min_motion_fraction)
if self.which_sensors_must_move == "all_active":
ok = bool(passed) and all(passed)
else:
ok = any(passed)
if not ok:
n_drop_motion += 1
continue
self.windows.append((seg_id, t_start))
kept += 1
# Compact per-segment print so noisy episodes don't drown the summary
if kept > 0 or T >= span:
print(f"[ReactSegmentDataset] {src_ep}/seg{seg_idx:02d} "
f"T={T:>5d} kept={kept:>3d} windows")
self.stats = {
"n_source_episodes": len({m["source_episode"] for m in self.segment_meta}),
"n_segments_loaded": len(self.segments),
"n_candidates": n_total_candidates,
"n_dropped_contact": n_drop_contact,
"n_dropped_motion": n_drop_motion,
"n_contact_rich_windows": len(self.windows),
"window_length": self.window_length,
"stride": self.stride,
"window_step": self.window_step,
"min_contact_fraction": self.min_contact_fraction,
"tactile_threshold": self.tactile_threshold,
"contact_metric": self.contact_metric,
"which_sensors": self.which_sensors,
"require_motion": self.require_motion,
"min_motion_mps": self.min_motion_mps,
"min_motion_fraction": self.min_motion_fraction,
"which_sensors_must_move":self.which_sensors_must_move,
}
pct = 100.0 * len(self.windows) / max(1, n_total_candidates)
print()
print(f"[ReactSegmentDataset] =================================")
print(f"[ReactSegmentDataset] Contact-rich windows sampled: {len(self.windows):,}")
print(f"[ReactSegmentDataset] ({pct:.1f}% of {n_total_candidates:,} sliding-window candidates)")
print(f"[ReactSegmentDataset] =================================")
print(f"[ReactSegmentDataset] From {self.stats['n_segments_loaded']} segments across "
f"{self.stats['n_source_episodes']} source episodes.")
print(f"[ReactSegmentDataset] Window spec: length={self.window_length}, "
f"stride={self.stride}, step={self.window_step} "
f"(≈{self.window_length / self.fps:.2f}s @ {self.fps:.0f} fps)")
print(f"[ReactSegmentDataset] Rejected by filter:")
print(f"[ReactSegmentDataset] contact (< {self.min_contact_fraction:.0%} of frames in tactile contact): {n_drop_contact:,}")
print(f"[ReactSegmentDataset] motion ({'enabled' if self.require_motion else 'disabled'}): {n_drop_motion:,}")
print(f"[ReactSegmentDataset] (No bad_frames filter — segments are already clean by construction.)")
def __len__(self) -> int:
return len(self.windows)
def __getitem__(self, idx: int) -> dict:
seg_id, t_start = self.windows[idx]
seg = self.segments[seg_id]
meta = self.segment_meta[seg_id]
frame_idx = torch.arange(t_start, t_start + self.window_length * self.stride, self.stride)
sample = {
"view": seg["view"][frame_idx],
"tactile_left": seg["tactile_left"][frame_idx],
"tactile_right": seg["tactile_right"][frame_idx],
"sensor_left_pose": seg["sensor_left_pose"][frame_idx],
"sensor_right_pose": seg["sensor_right_pose"][frame_idx],
"timestamps": seg["timestamps"][frame_idx],
"tactile_left_intensity": seg["tactile_left_intensity"][frame_idx],
"tactile_right_intensity": seg["tactile_right_intensity"][frame_idx],
"tactile_left_mixed": seg["tactile_left_mixed"][frame_idx],
"tactile_right_mixed": seg["tactile_right_mixed"][frame_idx],
}
sample["segment_path"] = str(self.segment_paths[seg_id])
sample["source_episode"] = meta["source_episode"]
sample["source_segment_idx"] = int(meta["source_segment_idx"])
sample["source_h5_frame_range"]= meta["source_h5_frame_range"]
sample["frame_start"] = int(t_start)
sample["frame_end"] = int(frame_idx[-1].item())
sample["active_sensors"] = list(self.segment_active[seg_id])
# H5 frame index of the first frame in this window (useful for cross-ref)
h5_range = meta["source_h5_frame_range"]
if h5_range is not None:
sample["h5_frame_start"] = int(h5_range[0]) + int(t_start)
sample["h5_frame_end"] = int(h5_range[0]) + int(frame_idx[-1].item())
return sample
if __name__ == "__main__":
import argparse
ap = argparse.ArgumentParser()
ap.add_argument("--segments_root", required=True,
help="processed/mode2_v1/motherboard (relative to dataset root)")
ap.add_argument("--tasks_json", default="tasks.json")
ap.add_argument("--window_length", type=int, default=16)
ap.add_argument("--stride", type=int, default=1)
ap.add_argument("--window_step", type=int, default=None)
ap.add_argument("--tactile_threshold", type=float, default=0.4)
ap.add_argument("--min_contact_fraction", type=float, default=0.5)
ap.add_argument("--contact_metric", default="mixed", choices=CONTACT_METRICS)
ap.add_argument("--which_sensors", default="both", choices=["any", "both", "left", "right"])
ap.add_argument("--require_motion", action="store_true")
ap.add_argument("--min_motion_mps", type=float, default=0.01)
ap.add_argument("--min_motion_fraction", type=float, default=0.25)
ap.add_argument("--which_sensors_must_move", default="all_active",
choices=["any", "all_active"])
args = ap.parse_args()
ds = ReactSegmentDataset(
segments_root=args.segments_root,
tasks_json_path=args.tasks_json,
window_length=args.window_length,
stride=args.stride,
window_step=args.window_step,
contact_metric=args.contact_metric,
tactile_threshold=args.tactile_threshold,
min_contact_fraction=args.min_contact_fraction,
which_sensors=args.which_sensors,
require_motion=args.require_motion,
min_motion_mps=args.min_motion_mps,
min_motion_fraction=args.min_motion_fraction,
which_sensors_must_move=args.which_sensors_must_move,
)
print(f"\nlen(ds) = {len(ds)}")
if len(ds):
sample = ds[0]
for k, v in sample.items():
if isinstance(v, torch.Tensor):
print(f" {k:30s} {tuple(v.shape)} {v.dtype}")
else:
print(f" {k:30s} {v!r}")
|