src/xstream_piflow_add_uuid_stop.py
1"""PiFlow AddUUIDStop: prepend a UUID column to each record."""
2import uuid
3import pyarrow as pa
4from dacp_piflow_support import params
5
6
7def process(ctx):
8 p = params(ctx, ['column'])
9 column = p['column']
10 source = ctx.input('source')
11 schema = source.schema
12 if not isinstance(column, str) or not column or column in schema.names:
13 raise ValueError('column must be a nonempty new name')
14 target = pa.schema([pa.field(column, pa.string(), nullable=False), *schema], metadata=schema.metadata)
15 out = ctx.records_output('result', target)
16 for batch in source.records():
17 ids = pa.array([uuid.uuid4().hex for _ in range(batch.num_rows)])
18 out.write(pa.RecordBatch.from_arrays([ids, *batch.columns], schema=target))
19
20
21def infer_outputs(inputs, params):
22 import pyarrow as pa
23 from xstream_operator import descriptor, result, records_schema
24 expected = {'source'}
25 if set(inputs) != expected:
26 raise ValueError('unexpected input ports')
27 outputs = {'result': {'representations': ['records'], 'formats': ['arrow'], 'certainty': 'unknown'}}
28 schema = records_schema(inputs['source']) if 'source' in inputs else None
29 column = params.get('column')
30 if not isinstance(column, str) or not column or (schema is not None and column in schema.names):
31 raise ValueError('column must be a nonempty new name')
32 if schema is not None:
33 schema = pa.schema([pa.field(column, pa.string(), nullable=False), *schema], metadata=schema.metadata)
34 if schema is not None:
35 outputs = {p: descriptor(['records'], ['arrow'], schema, certainty='known') for p in outputs}
36 return result(outputs)
37
38
39def preview_outputs(inputs, params, limits):
40 from xstream_operator import preview_process
41 return preview_process(process, inputs, params, limits)
42
43
44if __name__ == '__main__':
45 from xstream_operator import run_operator
46 run_operator(process, infer_outputs, preview_outputs=preview_outputs)