Skip to content

JoseLlarena/pypey

Repository files navigation

Pypey

Library for composing lazily-evaluated data-processing pipelines:

>>> from urllib.request import urlopen
>>> import re
>>> from pypey import pype, px
>>> # prints plain text message from example.com
>>> (pype(urlopen('http://example.com'))
...  .map(bytes.decode, str.split)
...  .flat()
...  .slice(90, 114)
...  .map(px(re.sub, r'\<\/?p\>', ''))
...  .to(' '.join))
'This domain is for use in illustrative examples in documents. You may use this domain in literature without prior coordination or asking for permission.'

Supports argument unpacking:

>>> import os
>>> from pypey import pype
>>>  # reads environment variables with "PYTHON" in them and prints their name
>>> pype.dict(os.environ).select(lambda key, val: 'PYTHON' in key).pick(0).print()
PYTHONPATH
IPYTHONENABLE
PYTHONIOENCODING
PYTHONUNBUFFERED
<pypey.pype.Pype object at 0x7ffa54006a30>

Supports parallelism:

>>> from timeit import timeit
>>> timeit("from math import log10; from pypey import pype; pype(range(1, 10_000_000)).map(log10).to(sum, print)", number=5)
65657052.08005966
65657052.08005966
65657052.08005966
65657052.08005966
65657052.08005966
10.055954356997972
>>> timeit("from math import log10; from pypey import pype; pype(range(1, 10_000_000)).map(log10, workers=8).to(sum, print)", number=5)
65657052.08005966
65657052.08005966
65657052.08005966
65657052.08005966
65657052.08005966
5.446932412014576

Installing

Install with pip:

pip install pypey

Testing

Unit tests are written with pytest. Run with:

pip install pytest

pytest

Pypey has been tested with python 3.7, 3.8 and 3.9. A few tests containing positional-only arguments will fail if they are run with 3.7 as it's a feature only available from 3.8 onwards.

Documentation

Extensive documentation at https://pypey.readthedocs.io/en/stable/index.html. Unit tests document through behaviour specification.

Changelog

Check the Changelog for fixes and enhancements of each version.

License

Copyright Jose Llarena 2021.

Distributed under the terms of the MIT license, Pypey is free and open source software.

About

lazy data pipes for python

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages