File size: 2,218 Bytes
caa4bd5 | 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 | """Lightweight h5py reader for the sharded T4 v1 dataset."""
from __future__ import annotations
import json
from pathlib import Path
import h5py
import numpy as np
DEFAULT_DATA_DIR = Path(__file__).resolve().parents[1] / "data" / "t4_structural_dynamics_v1"
def load_index(data_dir: str | Path = DEFAULT_DATA_DIR) -> list[dict]:
path = Path(data_dir) / "index.jsonl"
return [json.loads(line) for line in path.read_text(encoding="utf-8").splitlines() if line]
def trajectory_ids(data_dir: str | Path = DEFAULT_DATA_DIR) -> tuple[str, ...]:
return tuple(item["trajectory_id"] for item in load_index(data_dir))
def load_trajectory(
trajectory_id: str,
data_dir: str | Path = DEFAULT_DATA_DIR,
*,
include_fields: bool = True,
) -> dict[str, object]:
root = Path(data_dir)
record = next(item for item in load_index(root) if item["trajectory_id"] == trajectory_id)
with h5py.File(root / record["shard"], "r") as h5:
group = h5[record["group"]]
case = h5[record["case_id"]]
arrays = {
name: np.asarray(value)
for name, value in group.items()
if isinstance(value, h5py.Dataset)
}
if include_fields:
for name in ("time_s", "displacement_m", "velocity_m_per_s", "acceleration_m_per_s2"):
arrays[f"field_{name}" if name == "time_s" else name] = np.asarray(
group[f"fields/{name}"]
)
arrays["reference_geometry_m"] = np.asarray(case["common/reference_geometry_m"])
arrays["topology"] = np.asarray(case["common/topology"])
arrays["sensor_coordinates_m"] = np.asarray(case["common/sensor_coordinates_m"])
return {
"record": record,
"configuration": json.loads(case.attrs["config_json"]),
"quality": json.loads(case.attrs["quality_json"]),
"excitation": json.loads(group.attrs["spec_json"]),
"arrays": arrays,
}
if __name__ == "__main__":
ids = trajectory_ids()
sample = load_trajectory(ids[0], include_fields=False)
print(len(ids), ids[0])
print({name: value.shape for name, value in sample["arrays"].items()})
|