sain — Write safe Python code like Rust

a dependency-free which implements Rust's core crates, purely, in Python. Offering core Rust items such as Vec<T>, Result<T, E>, Option<T> and more. See the built-in types section below.

Install

Python 3.10 or higher is required.

pip install sain

Overview

sain is built completely on-top of Python's stdlib, inspired by The Rust Standard Library, designed to bring modern, safe, and ergonomic data structures and utilities to Python.

The goal is to provide developers with high-level abstractions and low-level control similar to what Rust offers, making it easier to write robust, and maintainable code.

Example

In this example we are writing a simple app that takes advantage of Rust's types which're implemented in Python.

This example provides a simple library that contains a Vec of books. Here, we take advantage of the Vec type and its idiomatic methods.

from __future__ import annotations
from sain import Result, Ok, Err  # used for safe error handling.
from sain import Option  # used for absense of a value, similar to `T | None`.
from sain import Vec  # A replacement for `list` type.

from dataclasses import dataclass, field

@dataclass
class Book:
    name: str
    author: str
    tags: set[str]
    pages = field(default_factory=set[str])

@dataclass
class Library:
    # just like any other mutable sequences.
    # it needs a default factory.
    books = field(default_factory=Vec[Book])

    # add a books to this library.
    def add(self, *books: Book):
        self.books.extend(books)

    # finds the first book that contains a specific tag.
    def find(self, pattern: str) -> Option[Book]:
        return self.books.iter().find(lambda book: pattern in book.tags)

    # finds the first book that contains a specific tag,
    # mapping Option[Book] to Result[Book, str] where `str` is the context
    # of the error.
    def find_or(self, pattern: str) -> Result[Book, str]:
        return self.find(pattern).ok_or(f"book with pattern {pattern} not found.")

    # We simply filter books that matches `author` and collect
    # them into a list[Book].
    def books_for(self, author: str):
        return self.books.iter().filter(lambda book: book.author == author).collect()


lib = Library()
lib.add(
    Book("Twilight", "Stephenie Meyer", {"Vampire", "Romance"}),
    Book("Silo", "Hugh Howey", {"Dystopian", "Sci-Fi"}),
    Book("The Eternaut", "Héctor Germán", {"Invasion", "Mystery"}),
)

# find the first vampire book then maps the book to its pages.
# maps Option[Book] -> Option[set[str]]
for page in lib.find("Vampire").map(lambda book: book.pages).unwrap():
    print(page)  # {page1, page2, ...}

match lib.find_or("Sci-Fi"):
    case Ok(book):
        print(book)  # Book("Silo", ...)
    case Err(err):
        print(err)  # book with pattern Sci-Fi not found.

# show Hugh's books.
print(lib.books_for("Hugh Howey"))  # [Book("Silo", ...)]

How to use this library

• Option, Result, Iterator, */collections, Default - The are core routines which are used almost everywhere within the library. You can easily opt these in your projects.
• sync, convert, error, macros, cfg - These are also core impls, but not widely used as the ones above.
• maybe_uninit, futures, boxed - You will probably see those used once or less anywhere, they're designed for lower-level, general use.

built-in types

name in Rust	name in Python	note	restrictions
Option<T>, Some(T), None	Option[T], Some(T), Some(None)	Some(None) has the same layout as None in Rust
Result<T, E>, Ok(T), Err(E)	Result[T, E], Ok(T), Err(E)	Basically a better and more verbose try/except
Vec<T>	Vec[T]	Same layout as list[T]
HashMap<K, V>	HashMap[K, V]	Same layout as dict[K, V]
bytes::Bytes	Bytes
bytes::BytesMut	BytesMut
LazyLock<T>	Lazy[T]
OnceLock<T>	Once[T]
Box<T>	Box[T]	this isn't a heap box
MaybeUninit<T>	MaybeUninit[T]	they serve the same purpose, but slightly different
impl Default	Default[T]
&dyn Error	Error
impl Iterator<T>	Iterator[T]
Iter<'a, T>	Iter[T]	collections called by .iter() are built from this type
iter::once::<T>()	iter.once[T]
iter::empty::<T>()	iter.empty[T]
iter::repeat::<T>()	iter.repeat[T]
cfg!()	cfg()	runtime cfg, not all predictions are supported
#[cfg_attr]	@cfg_attr()	runtime cfg, not all predictions are supported
#[doc]	@doc()	the docs get generated at runtime
todo!()	todo()
#[deprecated]	@deprecated()	will get removed when it get stabilized in warnings in Python 3.13
unimplemented!()	@unimplemented()

Remaining work

This is still early days for sain, it is no where near as stable as Python's stdlib.

This project mainly started as a fun / learning experience but turned into something more inspiring.

The release cycles were breaking due to poor decision making at first, but it should be stable enough now.