shore.io.geo

Xall geo format reader and writer.

File format

The geo format is a plain ASCII file containing a single structured block:

ni nj nk
x(0,0,0)  x(1,0,0)  x(2,0,0)  ...    (all x values, i-major Fortran order)
y(0,0,0)  y(1,0,0)  ...
z(0,0,0)  ...

Coordinate ordering is Fortran column-major: the i-index is fastest (innermost), then j, then k. This matches the memory layout of Fortran arrays declared as real :: x(ni, nj, nk).

Six values are written per line for readability. Coordinates are written with 10 significant digits in scientific notation (1.0000000000e+00).

Example for a 2×2×2 grid:

2 2 2
1.0000000000e+00  2.0000000000e+00  1.0000000000e+00  2.0000000000e+00
1.0000000000e+00  2.0000000000e+00  1.0000000000e+00  2.0000000000e+00
0.0000000000e+00  0.0000000000e+00  1.0000000000e+00  1.0000000000e+00
0.0000000000e+00  0.0000000000e+00  0.0000000000e+00  0.0000000000e+00
0.0000000000e+00  0.0000000000e+00  0.0000000000e+00  0.0000000000e+00
1.0000000000e+00  1.0000000000e+00  1.0000000000e+00  1.0000000000e+00

Functions

write_geo

def write_geo(path: str | Path, grid: np.ndarray) -> None

Write a single structured block to an ASCII geo file.

Parameters

Name	Type	Description
path	str | Path	Output file path (created or overwritten)
grid	np.ndarray (ni, nj, nk, 3)	Coordinate array; last axis is (x, y, z)

Example

from shore.io.geo import write_geo
import numpy as np

# Create a trivial 4×5×6 grid
ni, nj, nk = 4, 5, 6
i = np.arange(ni, dtype=float)
j = np.arange(nj, dtype=float)
k = np.arange(nk, dtype=float)
ig, jg, kg = np.meshgrid(i, j, k, indexing="ij")
grid = np.stack([ig, jg, kg], axis=-1)

write_geo("grid.geo", grid)

---

read_geo

def read_geo(path: str | Path) -> np.ndarray

Read a single structured block from an ASCII geo file.

Parameters

Name	Type	Description
path	str | Path	Path to the .geo file

Returns

numpy.ndarray of shape (ni, nj, nk, 3).

Raises

Exception	Condition
FileNotFoundError	path does not exist
ValueError	The number of values in the file does not match 3 × ni × nj × nk

Example

from shore.io.geo import read_geo

grid = read_geo("sphere.geo")
ni, nj, nk, _ = grid.shape
print(f"Grid: {ni}×{nj}×{nk}")

Round-trip guarantee

write_geo followed by read_geo on the same path is guaranteed to reproduce the original array within floating-point precision (numpy.testing.assert_allclose with default tolerances). This is verified by tests/test_geo.py::test_round_trip.