src/xstream_piflow_csv_save.py
1"""PiFlow CsvSave adapted to a CSV output stream instead of a destination path."""
2import csv
3import io
4from dacp_piflow_support import params
5
6
7def process(ctx):
8 p = params(ctx, optional=['header', 'delimiter'])
9 header, delimiter = p.get('header', True), p.get('delimiter', ',')
10 if type(header) is not bool or not isinstance(delimiter, str) or len(delimiter) != 1:
11 raise ValueError('header must be boolean; delimiter must be one character')
12 source = ctx.input('source')
13 names = source.schema.names
14 if len(names) != len(set(names)):
15 raise ValueError('unique column names required')
16 output = ctx.bytes_output('result', 'csv')
17 buffer = io.StringIO(newline='')
18 writer = csv.writer(buffer, delimiter=delimiter, lineterminator='\n')
19 def emit(values):
20 writer.writerow(values)
21 output.write(buffer.getvalue().encode('utf-8'))
22 buffer.seek(0)
23 buffer.truncate(0)
24 if header:
25 emit(names)
26 for batch in source.records():
27 for row in batch.to_pylist():
28 if any(isinstance(v, (list, dict)) for v in row.values()):
29 raise ValueError('CSV export requires scalar columns')
30 emit([row[n] for n in names])
31
32
33def infer_outputs(inputs, params):
34 import pyarrow as pa
35 from xstream_operator import descriptor, result, records_schema
36 expected = {'source'}
37 if set(inputs) != expected:
38 raise ValueError('unexpected input ports')
39 outputs = {'result': {'representations': ['bytes'], 'formats': ['csv'], 'certainty': 'known'}}
40 return result(outputs)
41
42
43def preview_outputs(inputs, params, limits):
44 from xstream_operator import preview_process
45 return preview_process(process, inputs, params, limits)
46
47
48if __name__ == '__main__':
49 from xstream_operator import run_operator
50 run_operator(process, infer_outputs, preview_outputs=preview_outputs)