Commit ·
0c8bca5
1
Parent(s): ef994e9
update
Browse files
scripts/__pycache__/upload_preview_to_hf.cpython-313.pyc
DELETED
|
Binary file (6.53 kB)
|
|
|
scripts/upload_preview_to_hf.py
DELETED
|
@@ -1,144 +0,0 @@
|
|
| 1 |
-
#!/usr/bin/env python3
|
| 2 |
-
|
| 3 |
-
import argparse
|
| 4 |
-
import math
|
| 5 |
-
import os
|
| 6 |
-
|
| 7 |
-
import pandas as pd
|
| 8 |
-
from datasets import Dataset, Features, Image, Sequence, Value
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
def parse_args():
|
| 12 |
-
parser = argparse.ArgumentParser()
|
| 13 |
-
parser.add_argument("--data", required=True, help="Path to preview folder")
|
| 14 |
-
parser.add_argument("--repo", required=True, help="HF dataset repo (e.g. org/name)")
|
| 15 |
-
parser.add_argument("--split", default="preview", help="Dataset split name")
|
| 16 |
-
return parser.parse_args()
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
def _coerce_dt(x: object) -> float:
|
| 20 |
-
"""Coerce merge-produced values (NaN/None/str) into a sane float offset."""
|
| 21 |
-
try:
|
| 22 |
-
v = float(x)
|
| 23 |
-
except (TypeError, ValueError):
|
| 24 |
-
return 0.0
|
| 25 |
-
if math.isnan(v) or math.isinf(v):
|
| 26 |
-
return 0.0
|
| 27 |
-
return v
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
def main():
|
| 31 |
-
args = parse_args()
|
| 32 |
-
|
| 33 |
-
data_dir = args.data
|
| 34 |
-
imu_path = os.path.join(data_dir, "imu.csv")
|
| 35 |
-
odom_path = os.path.join(data_dir, "odom.csv")
|
| 36 |
-
|
| 37 |
-
if not os.path.exists(imu_path):
|
| 38 |
-
raise FileNotFoundError(f"imu.csv not found in {data_dir}")
|
| 39 |
-
|
| 40 |
-
print("Loading IMU data...")
|
| 41 |
-
imu = pd.read_csv(imu_path)
|
| 42 |
-
|
| 43 |
-
# Normalize timestamps (filename-friendly and merge-stable).
|
| 44 |
-
imu["t"] = imu["t"].map(lambda x: f"{float(x):.6f}")
|
| 45 |
-
|
| 46 |
-
has_odom = os.path.exists(odom_path)
|
| 47 |
-
if has_odom:
|
| 48 |
-
print("Loading ODOM data...")
|
| 49 |
-
odom = pd.read_csv(odom_path)
|
| 50 |
-
odom["t"] = odom["t"].map(lambda x: f"{float(x):.6f}")
|
| 51 |
-
df = pd.merge(imu, odom, on="t", how="left")
|
| 52 |
-
else:
|
| 53 |
-
print("No odom.csv found -> filling pose with inf")
|
| 54 |
-
df = imu.copy()
|
| 55 |
-
|
| 56 |
-
print(f"Rows after merge: {len(df)}")
|
| 57 |
-
|
| 58 |
-
left_paths = []
|
| 59 |
-
right_paths = []
|
| 60 |
-
timestamps = []
|
| 61 |
-
gyro = []
|
| 62 |
-
accel = []
|
| 63 |
-
sync_dt = []
|
| 64 |
-
position = []
|
| 65 |
-
orientation = []
|
| 66 |
-
|
| 67 |
-
missing = 0
|
| 68 |
-
for _, row in df.iterrows():
|
| 69 |
-
t = row["t"]
|
| 70 |
-
|
| 71 |
-
l = os.path.join(data_dir, "left", f"{t}.png")
|
| 72 |
-
r = os.path.join(data_dir, "right", f"{t}.png")
|
| 73 |
-
|
| 74 |
-
if not os.path.exists(l) or not os.path.exists(r):
|
| 75 |
-
missing += 1
|
| 76 |
-
continue
|
| 77 |
-
|
| 78 |
-
left_paths.append(l)
|
| 79 |
-
right_paths.append(r)
|
| 80 |
-
timestamps.append(float(t))
|
| 81 |
-
|
| 82 |
-
# IMU
|
| 83 |
-
gyro.append([row["gx"], row["gy"], row["gz"]])
|
| 84 |
-
accel.append([row["ax"], row["ay"], row["az"]])
|
| 85 |
-
|
| 86 |
-
# Sync offsets (seconds) relative to the left-image timestamp.
|
| 87 |
-
dt_r = _coerce_dt(row.get("dt_right", 0.0))
|
| 88 |
-
dt_i = _coerce_dt(row.get("dt_imu", 0.0))
|
| 89 |
-
sync_dt.append([dt_r, dt_i])
|
| 90 |
-
|
| 91 |
-
# ODOM (optional)
|
| 92 |
-
if has_odom:
|
| 93 |
-
position.append([row["px"], row["py"], row["pz"]])
|
| 94 |
-
orientation.append([row["qx"], row["qy"], row["qz"], row["qw"]])
|
| 95 |
-
else:
|
| 96 |
-
position.append([math.inf, math.inf, math.inf])
|
| 97 |
-
orientation.append([math.inf, math.inf, math.inf, math.inf])
|
| 98 |
-
|
| 99 |
-
if not left_paths:
|
| 100 |
-
raise RuntimeError("No valid samples found")
|
| 101 |
-
|
| 102 |
-
if missing:
|
| 103 |
-
print(f"Skipped {missing} rows due to missing images")
|
| 104 |
-
|
| 105 |
-
print(f"Final dataset size: {len(left_paths)}")
|
| 106 |
-
|
| 107 |
-
features = Features(
|
| 108 |
-
{
|
| 109 |
-
"image_left": Image(),
|
| 110 |
-
"image_right": Image(),
|
| 111 |
-
"timestamp": Value("float64"),
|
| 112 |
-
"gyro": Sequence(Value("float32"), length=3),
|
| 113 |
-
"accel": Sequence(Value("float32"), length=3),
|
| 114 |
-
"sync_dt": Sequence(Value("float32"), length=2),
|
| 115 |
-
"position": Sequence(Value("float32"), length=3),
|
| 116 |
-
"orientation": Sequence(Value("float32"), length=4),
|
| 117 |
-
}
|
| 118 |
-
)
|
| 119 |
-
|
| 120 |
-
data = {
|
| 121 |
-
"image_left": left_paths,
|
| 122 |
-
"image_right": right_paths,
|
| 123 |
-
"timestamp": timestamps,
|
| 124 |
-
"gyro": gyro,
|
| 125 |
-
"accel": accel,
|
| 126 |
-
"sync_dt": sync_dt,
|
| 127 |
-
"position": position,
|
| 128 |
-
"orientation": orientation,
|
| 129 |
-
}
|
| 130 |
-
|
| 131 |
-
print("Creating Hugging Face dataset...")
|
| 132 |
-
ds = Dataset.from_dict(data, features=features)
|
| 133 |
-
|
| 134 |
-
print(ds)
|
| 135 |
-
|
| 136 |
-
print(f"Pushing to {args.repo} (split={args.split})...")
|
| 137 |
-
ds.push_to_hub(args.repo, split=args.split)
|
| 138 |
-
|
| 139 |
-
print("Done.")
|
| 140 |
-
|
| 141 |
-
|
| 142 |
-
if __name__ == "__main__":
|
| 143 |
-
main()
|
| 144 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|