pyochain 0.13.0

Method chaining for iterables and dictionaries in Python.

pyochain ⛓️

Fluent method chaining for Python.

Inspired by Rust's Iterator, Result, Option, and DataFrame libraries like Polars, pyochain provide a set of classes with a fluent and declarative API, to work with collections, handle optional values, or manage errors.

For a quick overview of the core types and their relationships, see the core types overview page.

Key Features

• ⛓️ Declarative & fluent chaining — Replace for loops, None checks, and error handling with chainable methods.
• 🦥 Lazy-first, 🔒 explicit mutability — Iter[T] for lazy, efficient iterations; Seq and Set for immutable data; Vec and SetMut when you need to mutate.
• Memory efficient - Almost all methods from Iter[T] operate in streaming fashion, and Vec[T] provides in-place methods with more memory efficiency than standard list methods (e.g. x.extend_move(y) won't create intermediate allocations like x.extend(y) followed by y.clear()).
• 🎯 Result and Option types - Handle None and exceptions in a fluent, explicit way.
• 🔥 Blazingly fast — Core Option and Result types are written in Rust for minimal overhead, and iterators use functions from cytoolz (Cython) and the stdlib itertools for maximum efficiency.
• 🛡️ 100% type-safe — Full generic support and autocompletion in your IDE.
• 📚 Accurate Documentation — Every method is documented and tested with runnable examples. Every code example in the website (or this README) is also tested, ensuring accuracy and reliability.
• 🔄 Interoperable — Seamlessly convert to/from types with various methods like .into() and .collect(), convert Iterables to Option or Result based on their truthiness, and more.
• 🐍 Mixins traits — Extend your own classes with the methods in the mixins provided by the traits module.

Installation

uv add pyochain # or pip install pyochain

See the package page on Pypi

Quick Start

Iterations

>>> import pyochain as pc
>>> # Lazy processing with Iter
>>> res: pc.Seq[tuple[int, str]] = (
...     pc.Iter.from_count(1)
...     .filter(lambda x: x % 2 != 0)
...     .map(lambda x: x**2)
...     .take(5)
...     .enumerate()
...     .map_star(lambda idx, value: (idx, str(value)))
...     .collect()
... )
>>> res
Seq((0, '1'), (1, '9'), (2, '25'), (3, '49'), (4, '81'))

For comparison, the above can be written in pure Python as the following (note that Pylance strict will complain because itertools.starmap has not the same overload exhaustiveness as pyochain's Iter.map_star):

>>> import itertools
>>>
>>> res: tuple[tuple[int, str], ...] = tuple(
...     itertools.islice(
...         itertools.starmap(
...             lambda idx, val: (idx, str(val)),
...             enumerate(
...                 map(lambda x: x**2, filter(lambda x: x % 2 != 0, itertools.count(1)))
...             ),
...         ),
...         5,
...     )
... )
>>>
>>> res
((0, '1'), (1, '9'), (2, '25'), (3, '49'), (4, '81'))

Result and Option Types

>>> import pyochain as pc
>>> def divide(a: int, b: int) -> pc.Option[float]:
...     return pc.NONE if b == 0 else pc.Some(a / b)
>>> divide(10, 2)
Some(5.0)
>>> divide (10, 0).unwrap_or(-1.0) # Provide a default value
-1.0
>>> # Convert between Collections -> Option -> Result
>>> data = pc.Seq([1, 2, 3])
>>> data.then_some() # Convert Seq to Option
Some(Seq(1, 2, 3))
>>> data.then_some().map(lambda x: x.sum()).ok_or("No values") # Convert Option to Result
Ok(6)
>>> pc.Seq[int](()).then_some().map(lambda x: x.sum()).ok_or("No values")
Err('No values')
>>> pc.Seq[int](()).then_some().map(lambda x: x.sum()).ok_or("No values").ok() # Get the Option back
NONE

Documentation

For comprehensive guides and examples:

🔍 Types Overview — Roles, comparisons and visual relationships

🔄 Interoperability - Converting between types

📖 Examples & Cookbook — Practical patterns and recipes

📚 Full API Reference — Complete API documentation

Notice on Stability ⚠️

pyochain is currently in early development (< 1.0), and the API may undergo significant changes multiple times before reaching a stable 1.0 release.

Contributing

Want to contribute? Read our contributing guide

Key Dependencies and credits

Most of the computations are done with implementations from itertools, cytoolz and more-itertools.

pyochain acts as a unifying API layer over these powerful tools.

https://github.com/pytoolz/cytoolz

https://github.com/more-itertools/more-itertools

The stubs used for the development, made by the maintainer of pyochain, can be found here:

https://github.com/OutSquareCapital/cytoolz-stubs