mlsimkit.common.cli.options_from_schema_shorthand

mlsimkit.common.cli.options_from_schema_shorthand(*param_decls, cls: type[Option] | None = None, model: Type[BaseModel], delimiter=',', quotechar='|', **attrs)

Decorator to generate a Click option from a Pydantic model using shorthand notation.

This decorator generates a single Click option that accepts a shorthand string representation of the Pydantic model’s fields. The shorthand string should be in the format "key1=value1, key2=value2, ..., keyN=valueN".

Example:

from pydantic import BaseModel

class MySettings(BaseModel):
    level: int
    mode: bool

@click.command()
@options_from_schema_shorthand('--settings', model=MySettings)
def cmd(settings: MySettings):
    print(settings)

From the command line:

$ python script.py --settings "level=10, mode=true"
level=10 mode=True

Standard click.option parameters are supported like short names (-o, --option, …) and multiple=True for accepting more than one shorthand strings:

@click.command()
@options_from_schema_shorthand(
    '-s', '--setting', 'settings',
    model=MySettings, multiple=True
)
def cmd(settings: MySettings):
    for s in settings:
        print(s)

From the command line:

$ python script.py -s "level=1,mode=true" -s "level=2,mode=false"
level=1 mode=True
level=2 mode=False

Pydantic List fields and non-string types are supported:

class MyModel(BaseModel):
    names: List[str]
    ages: List[int]

@click.command()
@options_from_schema_shorthand("--option", model=MyModel)
def cmd(option: MyModel):
    print(option)

From the command line:

$ python script.py --option "names=Alice Bob Charlie, ages=25 30 35"
names=['Alice', 'Bob', 'Charlie'] ages=[25, 30, 35]

A shorthand option can be loaded configuration file just like options_from_schema(), which is convenient when using program() --config functionality:

# config.yaml
option:
  names: [Alice, Bob, Charlie]
  ages: [25, 30, 35]

With multiple=True, the configuration supports lists:

# config.yaml
option:
  - names: [Alice, Bob, Charlie]
    ages: [25, 30, 35]
  - names: [Apple, Orange, Banana]
    ages: [1, 2, 3]
  - names: [Blue, Red, Green]
    ages: [0, 0, 0]

And the corresponding program code:

@mlsimkit.cli.program(name="MyProgram", version="0.1", use_config_file=True)
def program(ctx, config_file):
    ctx.obj["config_file"] = config_file

@program.command()
@options_from_schema_shorthand("--option", model=MyModel, multiple=True)
def cmd(ctx, option: MyModel):
    for opt in option:
        print(opt)

Parameters:

• *param_decls (str) – Parameter declarations for the Click option.

• cls (type[click.Option], optional) – The Click option class to use. Defaults to click.Option.

• model (Type[BaseModel]) – The Pydantic model class to generate the option for.

• **attrs – Additional keyword arguments to pass to the underlying click.Option.

Returns:

A decorator function that takes a command function and returns a

decorated version with the specified options.

Return type:

Callable

Raises:

ValueError – If type or callback are specified in **attrs, as they are handled automatically by this decorator.