Nice and clean iterator patterns for python
Project description
FluentIter
Nicer iterator patterns for Python! Chain map, filter, zip, unzip, cycle, skip and scan like there is no tomorrow.
Install with
pip install fluentiter
TLDR:
With fluentiter you can do this
lines = iterator(haystack_csv.split("\n"))
header = [x for x in lines.next().split(",") if x.strip(" ")]
needle = (
lines.map(lambda line: line.split(","))
.map(lambda values: {k: v for k, v in zip(header, values)})
.filter(lambda value_dct: value_dct["material"] != "hay")
.find(lambda value_dct: value_dct["type"] == "needle")
)
instead of this
lines = iter(some_csv.split("\n"))
header = [x for x in next(lines).split(",") if x.strip()]
needle = next(
filter(
lambda value_dct: value_dct["material"] != "hay",
filter(
lambda value_dct: value_dct["type"] == "needle",
map(
lambda values: {k: v for k, v in zip(header, values)},
map(lambda line: line.split(","), lines),
),
),
)
)
See the this example for a more complete version.
Motivation
To me Pythons iterator functions (i.e. map
, filter
, zip
, etc.) always seemed backward.
While generator expressions and list/dict/set comprehensions give you nice left-to-right readability,
map
and filter
force you to read "inside-out":
[func(x) for x in something if x == y]
map(func, filter(lambda x: x == y, something))
This is not an issue for short an simple statements, as they can be written concisely using expressions, but chaining multiple operations, will result in either many intermediate variable assignements or serious spaghettification.
How to use
Any Iterable
can be turned into a FluentIterator
by just passing it to the iterator
function:
from fluentiter import iterator
bugs = ["john", "paul", "ringo", "george"]
fluent = iterator(bugs)
# you can now do
# fluent.map(...).filter(...).cycle(...).scan(...) and so on
The FluentIterator
provides a rich set of methods you can call and chain together to your liking:
Of course there are the classic map
, filter
, and reduce
functions, but also some more really useful features like cycle
to repeat an iterator forever, find
to find an element or even partition
to turn your iterator into two.
In total FluentIterator
provides 37 methods to compose beautiful and easy to follow iteration
patterns.
Check the API docs to see them all.
Features
- 37 cool fresh iterator methods
- 100% Type annotated e.g.
iterator(["foo", "bar"]).map(len).to_list()
gets correctly inferred aslist[int]
- 100% Test coverage
- 0 dependencies outside the Python standard library[^0]
Contributing
The simplest way to contribute is to open an issue. If you would like to see some feature implemented or found a bug, head over to the issues section.
Developing
- Clone the repository
- Install Poetry
- Install development dependencies with
poetry install
- Enter the poetry shell with
poetry shell
Guidelines
Your PR should include relevant tests for the changes you are contributing and be fully type annotated. If you are unsure or stuck, please open the PR anyway and we can work it out together.
By opening a PR you agree to your code becoming part of the fluentiter package and being published under fluentiters license.
Running formatting, tests, and linting
- Formatting:
poe format
- Linting:
poe lint
- Tests:
poe test
All of the above: poe all
Viewing the coverage report
Running coverage html
will create a file htmlcov/index.html
you can open to view the test coverage report
Changelog
- 1.0.0
- initial release
- 1.0.1
- fixed errors int readme
- 1.1.0
- add
into(...)
method
- add
- 1.1.1
- add
py.typed
marker file
- add
- 1.2.0
- add
tumbling_window
function - fix mypy inference on
.flatten()
and.flat_map
- fix
more_itertools
being installed automatically
- add
Special Thanks
Thank you to all Rust maintainers for creating the Iterator trait, which served as the main source of inspiration.
[^0]: Some functions may require the more
extra which has more-itertools as a dependency
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for fluentiter-1.2.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | a623eda3a4a7974072e0761cf894b6124126237a8dd51e4e553f40ac6be6f4d8 |
|
MD5 | 836df2dac7223558f652d53cc8b41a03 |
|
BLAKE2b-256 | 9d9fac985c4960e3e77b3395177519fbf6bf79372787747066f9064d8fb9e373 |