inept 0.1.0

INteractive Editable oPTions

inept

An "interactive editable options" library for handling complex option hierarchies from multiple sources (script, CLI, config files, ...).

Installation

Install inept with pip

  pip install inept

Quick start

An option hierarchy is build by subclassing inept.Config and using the with statement. Inside of the class body, the special variable _ is made avalaible and provides different context managers that can customize the option nesting:

• a with _.options block will list optional values
• a with _.group block will list mandatory values
• a with _.switch block will list mutually exclusive values

import inept

class RestaurantMenu(inept.Config):
    with _.options:
        coffe: bool = False
        drink: str
        with _.switch:
            with _.group as simple_menu:
                dish: str = "steak and fries"
                with _.switch:
                    starter: str
                    dessert: str = "apple pie"
            with _.group as full_menu:
                starter: str = "salad"
                dish: str = "lasagna"
                dessert: str = "compote"

An instance of this class will store option values while ensuring consistency with the declared hierachy. Each option has a unique key, a string made of the nested option names, assembled with '.'.

order = RestaurantMenu()
print(list(order))  # list of the option names
# Outputs:
# ['coffe',
#  'drink',
#  'simple_menu',
#  'simple_menu.plat',
#  'simple_menu.starter',
#  'simple_menu.dessert',
#  'full_menu',
#  'full_menu.starter',
#  'full_menu.dish',
#  'full_menu.dessert']

A configuration object acts like a dict for the set and get operations.

order['coffe'] = True
order['full_menu.dessert'] = "creme brulee"
print(order['full_menu.starter'])  # default values are automatically set
# Outputs:
# 'salad'

The method to_dict() provides the full state of the configuration.

print(order.to_dict())
# Outputs:
# {'coffe': True,
#  'full_menu': True,
#  'full_menu.starter': 'salad',
#  'full_menu.dish': 'lasagna',
#  'full_menu.dessert': 'creme brulee'}

According to the options/group/switch used during declaration, the options are optional, mandatory or mutually exclusive. The configuration object ensures the consistency of its own state.

order['simple_menu'] = True  # this erases 'full_menu' because 'simple_menu' and 'full_menu' are in a switch block
print(order.to_dict())
# Outputs:
# {'coffe': True,
#  'simple_menu': True,
#  'simple_menu.dish': 'steak and fries',
#  'simple_menu.dessert': 'apple pie'}

A nested dict can also be produced.

print(order.to_nested_dict())
# Outputs:
# {'coffe': True,
#  'simple_menu': {'plat': 'steak and fries', 'dessert': 'apple pie'}}

A configuration can be filled from a dict.

new_order = RestaurantMenu()
some_dict = {'full_menu': True, 'drink': 'water'}
some_nested_dict = {'simple_menu': {'starter': 'salad'}}
new_order.update(some_dict)
print(new_order.to_dict())
# Outputs:
# {'coffe': False,
#  'drink': 'water',
#  'full_menu': True,
#  'full_menu.starter': 'salad',
#  'full_menu.dish': 'lasagna',
#  'full_menu.dessert': 'compote'}
new_order.update(inept.flatten(some_nested_dict))
print(new_order.to_dict())
# Outputs:
# {'coffe': False,
#  'drink': 'water',
#  'simple_menu': True,
#  'simple_menu.plat': 'steak and fries',
#  'simple_menu.starter': 'salad'}

The command line arguments can be loaded at any time with

# $ python myscript.py --coffe true --full_menu.dessert jelly
order.load_cli()

Roadmap

Inept is still in its early development.
Here is a non-exhaustive list of the upcoming features:

• documentation
• handling of complexe types like file/path or enum
• composing tree declaration
• extracting sub configuration
• serialization to/from configuration file formats (yaml, json, INI, ...)
• improving command line interface

Running Tests

To run tests, run the following command

  pytest

Used By

This project is used by the following softwares:

• https://github.com/BRGM/ComPASS

License

GPLv3