mario.parse_from_parquet(

path: str,

table: str,

mode: str,

calc_all: bool = False,

year: int = None,

name: str = None,

source: str = None,

model: str = 'Database',

flat: bool = False,

matrix_layouts: dict[str, object] | None = None,

tech_assumption: str | None = None,

**kwargs,

)

Parse a database from a folder of parquet files.

Parameters:

• path (str) – directory containing either one parquet file per matrix or one flat data.parquet plus units.parquet payload, or one units.parquet plus separate long-format parquet files per matrix.

• table (str) – acceptable options are ‘IOT’ & ‘SUT’

• mode (str) – acceptable options are flows and coefficients

• flat (bool, Optional) – if True, parse the canonical long-format MARIO parquet export, either as one combined data.parquet file or as separate long-format matrix files. Otherwise parse the matrix-per-file parquet layout.

• matrix_layouts (dict, Optional) – optional per-matrix semantic layout declarations for IOT parsers.