Result Pattern

A library designed to implement the "Errors as Values" pattern in Python 3.14+. It bridges the gap between pure functional safety and the pragmatic realities of the exception-heavy Python ecosystem.

Installation

# Package name is 'result-pattern', but you import 'result'
$ pip install result-pattern

Help

See documentation for more details.

Core Concept: Result (Sum Type)

A Result is a Sum Type representing a state of being either completely successful (Ok) or completely failed (Err).

1. Basic Usage

from result import Ok, Err, Result, is_ok


def divide(a: int, b: int) -> Result[float, str]:
    if b == 0:
        return Err("Cannot divide by zero")
    return Ok(a / b)


res = divide(10, 2)
if is_ok(res):
    # res is narrowed to Ok[float]
    print(f"Success: {res.ok()}")

2. Modern Pattern Matching

Leverage Python 3.10+ native matching with optimized __match_args__:

match divide(10, 0):
    case Ok(val):
        print(f"Value: {val}")
    case Err(msg):
        print(f"Error: {msg}")

3. Functional Chaining

Avoid nested if statements with linear pipelines:

Ok(10).map(lambda x: x * 2).tap(print).and_then(lambda x: Ok(x + 1))

Interoperability: The Catch System

Bridge the procedural world of exceptions to the functional world of Results.

1. `@catch` Decorator

Lift exception-throwing code into Result containers with zero boilerplate.

from result import catch


@catch(ValueError)
def parse_int(s: str) -> int:
    return int(s)


parse_int("10")  # Ok(10)
parse_int("abc")  # Err(ValueError(...))

2. Exception Mapping (`map_to` & Enums)

Immediately sanitize wild exceptions into strictly typed Domain Enums.

from enum import StrEnum
from result import catch


class ErrorCode(StrEnum):
    INVALID = "invalid_input"


# Single mapping
@catch(ValueError, map_to=ErrorCode.INVALID)
def risky_op(s: str): ...


# Multiple mapping dictionary
@catch({ValueError: ErrorCode.INVALID, KeyError: "missing"})
def complex_op(x): ...

3. `catch_call` Inline

Execute standard library functions inline without opening context blocks.

from result import catch_call
import json

res = catch_call(json.JSONDecodeError, json.loads, '{"key": "value"}')

4. `catch` Context Manager

Lasso an entire block of wild Python code and capture the result safely.

from result import catch
import json

with catch(json.JSONDecodeError) as safe_block:
    data = json.loads(payload)
    # Perform complex logic...
    safe_block.set(data["nested"]["key"])

# result is narrowed to Result[T, JSONDecodeError]
print(safe_block.result)

Monadic Orchestration: Do-Notation

Write procedural-looking code that automatically handles short-circuiting logic.

from result import do_notation, Do


@do_notation
def compile_pipeline(source: str) -> Do[str, Exception]:
    tokens = yield tokenize(source)  # Returns list[Token] or short-circuits Err
    ast = yield parse(tokens)  # Returns AST or short-circuits Err
    code = yield generate(ast)
    return code  # Automatically wrapped in Ok

Combinators: Advanced Utilities

Utility	Description
`validate()`	Applicative: Accumulates all errors instead of failing fast.
`traverse()`	Maps a fallible function over an iterable, failing fast.
`flow()`	A sequential pipeline macro for piped data transformations.
`succeeds()`	Filters a collection of Results, returning only the success values.
`partition_exceptions()`	Splits a mixed list of `[Value, Exception]` into `[Ok, Err]`.

Outcome: Product Type (Go/Odin Style)

While Result is mathematically pure, real-world systems often require fault tolerance. Outcome[T, E] is a Product Type that holds both a partial success value and diagnostic baggage simultaneously.

Advantages:

Native Unpacking: Inherits from NamedTuple for val, err = do_work() syntax.
Fault Tolerance: Retain your AST or data payload even if warnings/errors occurred.
Odin Style: Use .to_result() inside a @do_notation block to simulate Odin's or_return.

Usage:

from result import Outcome


def parse_with_diagnostics(source: str) -> Outcome[AST, list[str]]:
    # build AST and accumulate errors...
    return Outcome(ast, accumulated_errors)


# 1. Procedural Unpacking (Go/Odin style)
ast, errors = parse_with_diagnostics(src)
if errors:
    print(f"Warnings: {errors}")

# 2. Accumulation
new_outcome = Outcome(node, "e1").push_err("e2").merge(other_outcome)


# 3. Transition to Strict (or_return)
@do_notation
def strict_flow():
    # .to_result() halts execution if any errors exist
    ast = yield parse_with_diagnostics(src).to_result()
    return emit(ast)

API Philosophy

Zero-Escape Safety

Panics are isolated in the .unsafe namespace. Direct .unwrap() access is disabled at the root level to encourage safe functional patterns.

res = Err("fail")
res.unsafe.unwrap()  # Explicit panic

True Covariance

Full support for inheritance subtyping (e.g., Ok[Dog] is assignable to Result[Animal, Any]).