Surface Encoding (2.5D Elevation)

Surface Encoding (2.5D Elevation)#

Use parq_blockmodel.surface.RasterSurface and parq_blockmodel.surface.MeshSurface to evaluate single-valued elevation surfaces against a regular block model grid.

Key points#

  • Evaluation is geometry-first and vectorised against block centroids.

  • Output is a fraction in [0, 1] representing how much of each block lies below the surface.

  • Raster inputs must provide a regular X/Y grid with finite values.

  • Mesh inputs must be valid 2.5D surfaces with one Z value per (x, y) location.

  • Invalid inputs fail fast; v1 does not repair or infer missing geometry.

Create from raster or mesh#

import numpy as np
from parq_blockmodel import MeshSurface, RasterSurface

raster = RasterSurface.from_arrays(
    x_coords=np.array([0.0, 1.0, 2.0]),
    y_coords=np.array([0.0, 1.0, 2.0]),
    values=np.full((3, 3), 1.5),
    name="topography",
)

mesh = MeshSurface.from_ply("horizon.ply", name="horizon_a")

Evaluate against a block model#

# convenience method: evaluates and writes the column in one step
values = pbm.evaluate_surface(raster, column="topo_fraction")

# equivalent low-level form:
# values = raster.evaluate(pbm.geometry)
# ... map values by block_id ...
# pbm.write(..., merge=True)

Notes#

Current v1 scope excludes:

  • multi-surface stacking or precedence rules,

  • overhangs, vertical faces, or other multi-valued geometry,

  • sub-block sampling or high-resolution integration, and

  • CRS handling or reprojection workflows.

See also#