Source code

Download source archive ↓

src/script_text_replace.py

1"""Replace literal text within UTF-8 lines without buffering a complete stream."""
2from script_support import parameters, lines, infer, preview
3from xstream_operator import run_operator
4
5
6def validate(params):
7    p = parameters(params, ('find',), ('replacement',))
8    if not isinstance(p['find'], str) or not p['find'] or '\n' in p['find'] or '\r' in p['find']:
9        raise ValueError('find must be nonempty text without line endings')
10    if not isinstance(p.get('replacement', ''), str):
11        raise ValueError('replacement must be text')
12    if len(p['find']) > 4096 or len(p.get('replacement', '')) > 4096:
13        raise ValueError('find and replacement are limited to 4096 characters')
14    return p
15
16
17def process(ctx):
18    p = validate(ctx.params)
19    out = ctx.bytes_output('result', 'text')
20    for line in lines(ctx.input('source')):
21        # Split/join avoids allocating an expanded line when a short match repeats.
22        for index, part in enumerate(line.split(p['find'])):
23            if index:
24                out.write(p.get('replacement', '').encode('utf-8'))
25            if part:
26                out.write(part.encode('utf-8'))
27
28
29def infer_outputs(inputs, params):
30    return infer(inputs, params, validate)
31
32
33def preview_outputs(inputs, params, limits):
34    return preview(process, inputs, params, limits)
35
36
37if __name__ == '__main__':
38    run_operator(process, infer_outputs, preview_outputs=preview_outputs)