Configuration

aef_bng.config

Configuration for AEF-BNG processing.

AEFBNGConfig(years, bounds=BNG_BOUNDS, chunk_size=CHUNK_SIZE, output_path='./aef_bng_output', resampling='nearest', max_workers=4, table_name=None, boundary_path=None, boundary_query=None, boundary_buffer_m=0.0) dataclass

Configuration for AEF to BNG processing pipeline.

Attributes:

Name	Type	Description
years	list[int]	List of years to process (e.g. [2020, 2021, 2022]).
bounds	tuple[int, int, int, int]	BNG bounding box (minx, miny, maxx, maxy) in EPSG:27700 metres.
chunk_size	int	Size of processing chunks in metres (must align with BNG grid).
output_path	str	Path for GeoParquet output directory.
resampling	str	Resampling method for reprojection.
max_workers	int	Maximum concurrent workers for local processing.
table_name	str | None	Unity Catalog table name (e.g. "my_catalog.schema.aef_bng"). When set, Spark writes to this table instead of output_path.
boundary_path	str | None	Optional OGR-readable vector file (GeoJSON, GPKG, ...) used to spatially filter ingestion: only pixels whose 10m cell intersects the boundary are kept. None (default) disables filtering entirely.
boundary_query	str | None	Optional pandas-style attribute filter applied to the boundary file, e.g. "CTRY25NM in ['England', 'Scotland', 'Wales']".
boundary_buffer_m	float	Optional outward buffer in metres applied to the dissolved boundary geometry.

__post_init__()

Validate configuration values.

Source code in aef_bng/config.py

def __post_init__(self) -> None:
    """Validate configuration values."""
    if not self.years:
        raise ValueError("At least one year must be specified")
    if self.chunk_size <= 0 or self.chunk_size % 1000 != 0:
        raise ValueError("chunk_size must be a positive multiple of 1000")
    if len(self.bounds) != 4:
        raise ValueError("bounds must be a 4-tuple (minx, miny, maxx, maxy)")
    minx, miny, maxx, maxy = self.bounds
    if minx >= maxx or miny >= maxy:
        raise ValueError("Invalid bounds: min must be less than max")
    if not math.isfinite(self.boundary_buffer_m) or self.boundary_buffer_m < 0:
        raise ValueError("boundary_buffer_m must be finite and >= 0")
    if self.boundary_path is None and (self.boundary_query or self.boundary_buffer_m > 0):
        raise ValueError("boundary_query/boundary_buffer_m require boundary_path")