pureconfig 0.1.0

simple setting decorator framework and platform-independent path library

pureconfig

Document Specification

• System Version: 0.1.0
• Document Version: v1.0.0
• Last Modified: 2026-07-09

Installation

pip install pureconfig

Quick Start

path library

this is a function return the path of app home.

from pureconfig import get_app_home

# case 1:
# get a platform home path
app = get_app_home()
print(app)
# returns a native home path

# case 2:
# get a 
get_app_home('My Module')
# returns .../My Module

this function returns native pathlib.Path instance

Supported Platforms and Mapping

• on Windows: %LocalAppData%
• on Linux: ~/.config or custom $XDG_CONFIG_HOME
• on Mac: ~/Library/Application Support
• on AOS, IOS: app container path

for Windows, Linux, Mac:

when moduleName attribute is not None,
given modultName is extended to default mapping.

• on Windows: C:/Users/<user>/AppData/Local/My Module
• on Linux: ~/.config/My Module
• on Mac: ~/Library/Application Support/My Module

configuration decorator

this is a decorator function that construct class automatically.

from pureconfig import configurate

def your_function_here(key: str) -> Any:
    if key == 'yourSetting':
        return 'setting-value'
    elif key == 'yourPath':
        return 'some/path/here'
    elif key == 'yourOption':
        return True
    elif key == 'yourID':
        return 100

@configurate(your_function_here)
class MySetting:
    yourSetting: str
    yourPath: Path
    yourOption: bool
    yourID: int

this equivalant with:

class MySetting:
    yourSetting = str('setting-value')
    yourPath = Path('some/path/here')
    yourOption = bool(True)
    yourID = int(100)

why this useful?

this decorator is built to handle variant option(s) under setting. so, you can do like this:

from pureconfig import configurate

# to handle Path, pathlib should import.
from pathlib import Path

# setting some value provider
setting_values = {
    "name": "james",
    "email": "[e-mail]",
    "department": "Customer Service",
    "db_access": False,
    "on_work": True,
    "ID": 95,
    "pc_home": "C:/Users/james"
}

# since value provider is dictionary, so i wrapped with lambda function.
# provider function should be take 1 str, return 1 data for given key.
# the "key" argument transfer the string value same as attribute that defined under your class definition.
@configurate(lambda key: setting_values[key])
class Information:
    # this class now only has name, email, department field.
    name: str
    email: str
    department: str

@configurate(lambda key: setting_values[key])
class PCSetting:
    # this class now only has name, db_access, ID, pc_home field.
    name: str
    db_access: bool
    ID: int
    pc_home: Path
@configurate(lambda key: setting_values[key])
class AllConfig:
    # this class now have all keys from setting_values
    name: str
    email: str
    department: str
    db_access: bool
    on_work: bool
    ID: int
    pc_home: Path

you can make any combination of field name for configuration, with single value provider function.
or, you can pass any custom-provider-function to use different value.

Future Update

• frozen option
• instantize option