chunk_worker

sqllocks_spindle.engine.chunk_worker

Subprocess worker function for multi-process chunk generation.

Functions:

generate_chunk(schema_path, seed, offset, count)

Generate a chunk of synthetic data for all tables in a schema.

Designed to run inside a ProcessPoolExecutor worker. Must be importable at the top level (no closure captures) and return plain Python lists so the result pickles cleanly across process boundaries.

Parameters:

Name	Type	Description	Default
schema_path	str	Path to a .spindle.json file.	required
seed	int	Random seed for this chunk. Each chunk should use a unique seed derived from the base seed and chunk index so that chunks are independent but individually reproducible.	required
offset	int	Row offset for PK sequence columns. A chunk with offset=N and count=M will produce sequence IDs starting at start + N * step (e.g. offset=100, start=1, step=1 → IDs 101-200).	required
count	int	Number of rows to generate per table.	required

Returns:

Type	Description
dict[str, dict[str, list]]	{table_name: {column_name: [values...]}} — plain Python lists,
dict[str, dict[str, list]]	NOT numpy arrays.