Loader for excluded_nodes.csv.
The MATLAB workflow uses readmatrix on the file
excluded_nodes_<N>.csv (see reduction_test.m line 18).
readmatrix happily ignores a leading text header (ExcludedNodes
in the expected-output sample) and returns the numeric column. We
emulate that here.
load_excluded_nodes(path: str | Path) -> np.ndarray
Read an excluded_nodes.csv file into a 1-D integer array.
Accepts files that either have a single ExcludedNodes header row
or no header at all.
Source code in src/simplenet/io/csv_loader.py
| def load_excluded_nodes(path: str | Path) -> np.ndarray:
"""Read an ``excluded_nodes.csv`` file into a 1-D integer array.
Accepts files that either have a single ``ExcludedNodes`` header row
or no header at all.
"""
p = Path(path)
first_line = p.read_text(encoding="utf-8", errors="replace").splitlines()[0] if p.exists() else ""
try:
float(first_line.strip().split(",")[0])
header: int | None = None
except (ValueError, IndexError):
header = 0
df = pd.read_csv(p, header=header)
if df.shape[1] == 0:
return np.zeros(0, dtype=np.int64)
col = df.iloc[:, 0].dropna()
arr = np.asarray(col, dtype=float)
int_arr = arr.astype(np.int64)
if not np.all(int_arr == arr):
return arr
return int_arr
|