File size: 12,940 Bytes
ba3985e | 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 | from pathlib import Path
import argparse
import json
import pickle
import sys
from typing import Dict, List, Optional, Sequence, Tuple
import pandas as pd
PROJECT_ROOT = Path(__file__).resolve().parents[1]
if str(PROJECT_ROOT) not in sys.path:
sys.path.insert(0, str(PROJECT_ROOT))
from rr_label_study.oven_study import (
_aggregate_summary,
_annotate_phase_columns,
_derive_templates,
_json_safe,
_load_demo,
)
from scripts.recompute_oven_episode_parallel import (
_chunk_frame_indices,
_collect_debug_rows,
_collect_rows,
_launch_xvfb,
_spawn_frame_batch_job,
_stop_process,
)
def _merge_new_columns(
base_df: pd.DataFrame, probe_df: pd.DataFrame
) -> Tuple[pd.DataFrame, List[str]]:
new_columns = [
column for column in probe_df.columns if column != "frame_index" and column not in base_df.columns
]
if not new_columns:
return base_df.copy(), []
merged = base_df.merge(
probe_df[["frame_index", *new_columns]],
on="frame_index",
how="left",
sort=False,
)
return merged, new_columns
def _verification_record(
base_df: pd.DataFrame,
merged_df: pd.DataFrame,
base_key_df: Optional[pd.DataFrame],
merged_key_df: Optional[pd.DataFrame],
base_metrics: Optional[Dict[str, object]],
output_metrics: Optional[Dict[str, object]],
debug_rows: List[Dict[str, object]],
) -> Dict[str, object]:
dense_equal = base_df.equals(merged_df[base_df.columns])
key_equal = True
if base_key_df is not None and merged_key_df is not None:
key_equal = base_key_df.equals(merged_key_df[base_key_df.columns])
metrics_equal = True
if base_metrics is not None and output_metrics is not None:
metrics_equal = base_metrics == output_metrics
dense_pose_consistent = True
best_pose_consistent = True
num_goal_consistent = True
debug_by_frame = {
int(row["frame_index"]): row for row in debug_rows
}
for _, row in merged_df.iterrows():
frame_index = int(row["frame_index"])
debug = debug_by_frame.get(frame_index)
if debug is None:
dense_pose_consistent = False
best_pose_consistent = False
num_goal_consistent = False
continue
state = debug["state"]
if "left_arm_pose_x" in merged_df.columns:
expected_pose = state.get("left_arm_pose", [])
actual_pose = [
float(row["left_arm_pose_x"]),
float(row["left_arm_pose_y"]),
float(row["left_arm_pose_z"]),
float(row["left_arm_pose_qx"]),
float(row["left_arm_pose_qy"]),
float(row["left_arm_pose_qz"]),
float(row["left_arm_pose_qw"]),
]
if any(abs(a - b) > 1e-9 for a, b in zip(actual_pose, expected_pose)):
dense_pose_consistent = False
if "p_pre_num_goal_poses" in merged_df.columns:
if int(round(float(row["p_pre_num_goal_poses"]))) != int(
debug["p_pre"].get("num_goal_poses", 0)
):
num_goal_consistent = False
if "p_pre_best_target_pose_x" in merged_df.columns:
expected_best = debug["p_pre"].get("best_goal_pose", [])
if expected_best:
actual_best = [
float(row["p_pre_best_target_pose_x"]),
float(row["p_pre_best_target_pose_y"]),
float(row["p_pre_best_target_pose_z"]),
float(row["p_pre_best_target_pose_qx"]),
float(row["p_pre_best_target_pose_qy"]),
float(row["p_pre_best_target_pose_qz"]),
float(row["p_pre_best_target_pose_qw"]),
]
if any(abs(a - b) > 1e-9 for a, b in zip(actual_best, expected_best)):
best_pose_consistent = False
else:
actual_best = [
row["p_pre_best_target_pose_x"],
row["p_pre_best_target_pose_y"],
row["p_pre_best_target_pose_z"],
row["p_pre_best_target_pose_qx"],
row["p_pre_best_target_pose_qy"],
row["p_pre_best_target_pose_qz"],
row["p_pre_best_target_pose_qw"],
]
if any(pd.notna(value) for value in actual_best):
best_pose_consistent = False
return {
"dense_existing_columns_unchanged": bool(dense_equal),
"keyframe_existing_columns_unchanged": bool(key_equal),
"metrics_json_preserved": bool(metrics_equal),
"debug_row_count": int(len(debug_rows)),
"dense_row_count": int(len(merged_df)),
"dense_pose_columns_match_debug_state": bool(dense_pose_consistent),
"best_target_pose_columns_match_debug": bool(best_pose_consistent),
"num_goal_pose_columns_match_debug": bool(num_goal_consistent),
}
def main() -> int:
parser = argparse.ArgumentParser()
parser.add_argument("--dataset-root", required=True)
parser.add_argument("--episode-dir", required=True)
parser.add_argument("--base-dense-csv", required=True)
parser.add_argument("--output-dir", required=True)
parser.add_argument("--checkpoint-stride", type=int, default=16)
parser.add_argument("--num-workers", type=int, default=8)
parser.add_argument("--base-display", type=int, default=700)
parser.add_argument("--template-episode-dir")
parser.add_argument("--stagger-seconds", type=float, default=0.15)
parser.add_argument("--base-keyframes-csv")
parser.add_argument("--base-metrics-json")
parser.add_argument("--base-summary-json")
parser.add_argument("--keep-frame-json", action="store_true")
args = parser.parse_args()
dataset_root = Path(args.dataset_root)
episode_dir = Path(args.episode_dir)
output_dir = Path(args.output_dir)
output_dir.mkdir(parents=True, exist_ok=True)
base_dense_csv = Path(args.base_dense_csv)
base_df = pd.read_csv(base_dense_csv)
base_key_df = None
if args.base_keyframes_csv:
base_key_df = pd.read_csv(args.base_keyframes_csv)
base_metrics = None
if args.base_metrics_json:
base_metrics = json.loads(Path(args.base_metrics_json).read_text())
base_summary = None
if args.base_summary_json:
base_summary = json.loads(Path(args.base_summary_json).read_text())
demo = _load_demo(episode_dir)
num_frames = len(demo)
if len(base_df) != num_frames:
raise ValueError(
f"base dense rows {len(base_df)} do not match demo length {num_frames} for {episode_dir}"
)
template_episode_dir = (
Path(args.template_episode_dir) if args.template_episode_dir else episode_dir
)
templates, template_frames = _derive_templates(dataset_root, template_episode_dir)
templates_pkl = output_dir.joinpath("templates.pkl")
with templates_pkl.open("wb") as handle:
pickle.dump(templates, handle)
with output_dir.joinpath("templates.json").open("w", encoding="utf-8") as handle:
json.dump(
{
"template_mode": "per_episode",
"template_episode": template_episode_dir.name,
"template_frames": template_frames,
"templates": templates.to_json(),
"preserve_base_dense_csv": str(base_dense_csv),
},
handle,
indent=2,
)
frame_json_dir = output_dir.joinpath("frame_rows")
frame_json_dir.mkdir(parents=True, exist_ok=True)
frame_indices = list(range(num_frames))
frame_chunks = _chunk_frame_indices(frame_indices, args.num_workers)
displays = [args.base_display + index for index in range(len(frame_chunks))]
xvfb_procs = []
active: Dict[int, Tuple[Sequence[int], object]] = {}
try:
for display_num in displays:
xvfb_procs.append(
_launch_xvfb(display_num, output_dir.joinpath(f"xvfb_{display_num}.log"))
)
for display_num, frame_chunk in zip(displays, frame_chunks):
process = _spawn_frame_batch_job(
display_num=display_num,
episode_dir=episode_dir,
templates_pkl=templates_pkl,
frame_indices=frame_chunk,
checkpoint_stride=args.checkpoint_stride,
output_dir=frame_json_dir,
)
active[display_num] = (frame_chunk, process)
if args.stagger_seconds > 0:
import time
time.sleep(args.stagger_seconds)
while active:
import time
time.sleep(1.0)
finished: List[int] = []
for display_num, (frame_chunk, process) in active.items():
return_code = process.poll()
if return_code is None:
continue
missing = [
frame_index
for frame_index in frame_chunk
if not frame_json_dir.joinpath(f"frame_{frame_index:04d}.json").exists()
or not frame_json_dir.joinpath(f"frame_{frame_index:04d}.debug.json").exists()
]
if return_code != 0 or missing:
raise RuntimeError(
f"display :{display_num} failed for frames {list(frame_chunk)[:3]}...; missing={missing[:8]}"
)
finished.append(display_num)
for display_num in finished:
active.pop(display_num)
finally:
for _, process in list(active.values()):
_stop_process(process)
for xvfb in xvfb_procs:
_stop_process(xvfb)
probe_df = _collect_rows(frame_json_dir, num_frames)
debug_rows = _collect_debug_rows(frame_json_dir, num_frames)
merged_df, new_columns = _merge_new_columns(base_df, probe_df)
annotated_df = _annotate_phase_columns(merged_df.copy())
phase_new_columns = [
column
for column in annotated_df.columns
if column not in merged_df.columns
]
if phase_new_columns:
merged_df = merged_df.merge(
annotated_df[["frame_index", *phase_new_columns]],
on="frame_index",
how="left",
sort=False,
)
new_columns.extend(phase_new_columns)
merged_key_df = None
if base_key_df is not None:
merged_key_df, _ = _merge_new_columns(base_key_df, probe_df)
if phase_new_columns:
merged_key_df = merged_key_df.merge(
annotated_df[["frame_index", *phase_new_columns]],
on="frame_index",
how="left",
sort=False,
)
output_metrics = base_metrics if base_metrics is not None else None
output_summary = base_summary if base_summary is not None else (
_aggregate_summary([output_metrics]) if output_metrics is not None else None
)
merged_df.to_csv(output_dir.joinpath(f"{episode_dir.name}.dense.csv"), index=False)
if merged_key_df is not None:
merged_key_df.to_csv(output_dir.joinpath(f"{episode_dir.name}.keyframes.csv"), index=False)
elif args.base_keyframes_csv:
raise RuntimeError("base keyframes csv was provided but merged keyframes are missing")
if output_metrics is not None:
with output_dir.joinpath(f"{episode_dir.name}.metrics.json").open("w", encoding="utf-8") as handle:
json.dump(output_metrics, handle, indent=2)
if output_summary is not None:
with output_dir.joinpath("summary.json").open("w", encoding="utf-8") as handle:
json.dump(output_summary, handle, indent=2)
with output_dir.joinpath(f"{episode_dir.name}.debug.jsonl").open("w", encoding="utf-8") as handle:
for row in debug_rows:
handle.write(json.dumps(_json_safe(row)))
handle.write("\n")
verification = _verification_record(
base_df=base_df,
merged_df=merged_df,
base_key_df=base_key_df,
merged_key_df=merged_key_df,
base_metrics=base_metrics,
output_metrics=output_metrics,
debug_rows=debug_rows,
)
verification["new_columns_added"] = new_columns
verification["phase_new_columns_added"] = phase_new_columns
verification["probe_mode"] = "preserve_base_metrics"
with output_dir.joinpath("verification.json").open("w", encoding="utf-8") as handle:
json.dump(_json_safe(verification), handle, indent=2)
if not args.keep_frame_json:
for row_path in frame_json_dir.glob("frame_*.json*"):
row_path.unlink()
frame_json_dir.rmdir()
print(json.dumps(_json_safe(verification), indent=2))
return 0
if __name__ == "__main__":
raise SystemExit(main())
|