unwrappy

Rust-inspired Result and Option types for Python, enabling safe, expressive error handling with errors as values.

Installation

pip install unwrappy

Quick Start

from unwrappy import Ok, Err, Result

def divide(a: int, b: int) -> Result[float, str]:
    if b == 0:
        return Err("division by zero")
    return Ok(a / b)

# Pattern matching (Python 3.10+)
match divide(10, 2):
    case Ok(value):
        print(f"Result: {value}")
    case Err(error):
        print(f"Error: {error}")

# Combinator chaining
result = (
    divide(10, 2)
    .map(lambda x: x * 2)
    .and_then(lambda x: Ok(int(x)) if x < 100 else Err("too large"))
)

Why unwrappy?

• Explicit error handling: No hidden exceptions, errors are values
• Type-safe: Full generic type support with proper inference
• Functional: Rich combinator API (map, and_then, or_else, etc.)
• Async-first: LazyResult for clean async operation chaining
• Pattern matching: Works with Python 3.10+ structural matching

Core Types

Result[T, E]

A type that represents either success (Ok) or failure (Err).

from unwrappy import Ok, Err, Result

# Success
ok: Result[int, str] = Ok(42)
ok.unwrap()      # 42
ok.is_ok()       # True

# Error
err: Result[int, str] = Err("failed")
err.unwrap_err() # "failed"
err.is_err()     # True

LazyResult[T, E]

For async operation chaining without nested awaits:

from unwrappy import LazyResult, Ok, Err

async def fetch_user(id: int) -> Result[dict, str]: ...
async def fetch_profile(user: dict) -> Result[dict, str]: ...

# Clean async chaining - no nested awaits!
result = await (
    LazyResult.from_awaitable(fetch_user(42))
    .and_then(fetch_profile)
    .map(lambda p: p["name"])
    .map(str.upper)
    .collect()
)

Option[T]

A type that represents an optional value: either Some(value) or Nothing.

from unwrappy import Some, NOTHING, Option, from_nullable

# Has value
some: Option[int] = Some(42)
some.unwrap()      # 42
some.is_some()     # True

# No value
nothing: Option[int] = NOTHING
nothing.is_nothing()  # True

# From nullable Python value
value: str | None = get_optional_value()
opt = from_nullable(value)  # Some(value) or NOTHING

LazyOption[T]

For async operation chaining on optional values:

from unwrappy import LazyOption, Some

async def fetch_config(key: str) -> Option[str]: ...
async def parse_value(s: str) -> Option[int]: ...

# Clean async chaining
result = await (
    LazyOption.from_awaitable(fetch_config("timeout"))
    .and_then(parse_value)
    .map(lambda x: x * 1000)
    .collect()
)

API Overview

Result API

Transformation

Extraction

Inspection

Utilities

Option API

Transformation

Extraction

Inspection

Utilities

Examples

Error Recovery

def get_config(key: str) -> Result[str, str]:
    return Err(f"missing: {key}")

# Recover with default
value = get_config("port").unwrap_or("8080")

# Recover with computation
value = (
    get_config("port")
    .or_else(lambda e: Ok("8080"))
    .unwrap()
)

Chaining Operations

def parse_int(s: str) -> Result[int, str]:
    try:
        return Ok(int(s))
    except ValueError:
        return Err(f"invalid number: {s}")

def validate_positive(n: int) -> Result[int, str]:
    return Ok(n) if n > 0 else Err("must be positive")

result = (
    parse_int("42")
    .and_then(validate_positive)
    .map(lambda x: x * 2)
)
# Ok(84)

Async Operations with LazyResult

async def fetch_user(id: int) -> Result[User, str]:
    # async database call
    ...

async def fetch_posts(user: User) -> Result[list[Post], str]:
    # async API call
    ...

# Build pipeline, execute once
result = await (
    LazyResult.from_awaitable(fetch_user(42))
    .and_then(fetch_posts)              # async
    .map(lambda posts: len(posts))      # sync
    .tee(lambda n: print(f"Found {n}")) # side effect
    .collect()
)

Working with Optional Values

from unwrappy import Some, NOTHING, Option, from_nullable

# Convert nullable Python values
def get_user_email(user_id: int) -> str | None:
    # May return None if user has no email
    ...

email_opt = from_nullable(get_user_email(42))

# Chain operations on optional values
display_name = (
    email_opt
    .map(lambda e: e.split("@")[0])
    .map(str.title)
    .unwrap_or("Anonymous")
)

# Filter with predicates
valid_port = (
    Some(8080)
    .filter(lambda p: 1 <= p <= 65535)
    .unwrap_or(3000)
)

# Convert to Result for error context
result = (
    from_nullable(get_user_email(42))
    .ok_or("User has no email configured")
)

Batch Processing

from unwrappy import Ok, sequence_results, traverse_results

# Collect multiple Results
results = [Ok(1), Ok(2), Ok(3)]
combined = sequence_results(results)  # Ok([1, 2, 3])

# Map and collect
items = ["1", "2", "3"]
parsed = traverse_results(items, parse_int)  # Ok([1, 2, 3])

from unwrappy import Some, NOTHING, sequence_options, traverse_options, from_nullable

# Collect multiple Options
options = [Some(1), Some(2), Some(3)]
combined = sequence_options(options)  # Some([1, 2, 3])

# Fails fast if any is Nothing
options_with_nothing = [Some(1), NOTHING, Some(3)]
combined = sequence_options(options_with_nothing)  # NOTHING

# Map nullable values and collect
items: list[int | None] = [1, 2, 3]
result = traverse_options(items, from_nullable)  # Some([1, 2, 3])

Serialization

unwrappy supports JSON serialization for integration with task queues and workflow frameworks (Celery, Temporal, DBOS, etc.).

from unwrappy import Ok, Err, Some, NOTHING, dumps, loads

# Serialize Result
encoded = dumps(Ok({"key": "value"}))
# '{"__unwrappy_type__": "Ok", "value": {"key": "value"}}'

# Serialize Option
encoded = dumps(Some(42))
# '{"__unwrappy_type__": "Some", "value": 42}'

encoded = dumps(NOTHING)
# '{"__unwrappy_type__": "Nothing"}'

# Deserialize
decoded = loads(encoded)  # Some(42), NOTHING, Ok(...), or Err(...)

For standard json module usage:

import json
from unwrappy import ResultEncoder, result_decoder

encoded = json.dumps(Ok(42), cls=ResultEncoder)
decoded = json.loads(encoded, object_hook=result_decoder)

Note: LazyResult and LazyOption cannot be serialized directly. Call .collect() first to get a concrete Result or Option.

See ARCHITECTURE.md for framework integration examples.

License

MIT