help message of sub model in basesetting do not show the proper default value if the top field has a default

[braindevices]

class SubModel(BaseModel):
    v1: str = "default"
    v2: bytes = b"hello"
    v3: int

class Settings(BaseSettings):
    model_config = SettingsConfigDict(
        cli_prog_name="prog1",
        cli_parse_args=True,
        env_prefix="MYTEST_")

    v0: str = "ok"
    sub_model: SubModel = SubModel(v1="top default", v3=3)

sys.argv = [
        'example.py',
        "-h"
    ]
Settings()

It show help:

...
  --sub_model.v1 str    (default: default)
  --sub_model.v2 bytes  (default: b'hello')
  --sub_model.v3 int    (required)

This is apparently wrong, because the default values:

sys.argv = [
        'example.py',
    ]
print(Settings())
#> v0='ok' sub_model=SubModel(v1='top default', v2=b'hello', v3=3)

The help message should use the actual top level default values instead of the sub model's own default values.

We expect:

...
  --sub_model.v1 str    (default: top default)
  --sub_model.v2 bytes  (default: b'hello')
  --sub_model.v3 int    (default: 3)

[hramezani]

Thanks @braindevices for reporting this issue.

I think the help extracts default values from the model definition.
@kschwab would it be too complex to consider defaults in sub field definition?

[kschwab]

@hramezani, no, I've opened the PR. This is the fix I was discussing in #347, where the CLI will now properly display the defaults:

options:
  -h, --help            show this help message and exit
  --v0 str              (default: ok)

sub_model options:
  --sub_model JSON      set sub_model from JSON string
  --sub_model.v1 str    (default: top default)
  --sub_model.v2 bytes  (default: b'hello')
  --sub_model.v3 int    (default: 3)

However, without fixing #347, it will not function properly:

sys.argv = [
        'example.py',
        '--sub_model.v1=oops',
    ]
Settings()
"""
pydantic_core._pydantic_core.ValidationError: 1 validation error for Settings
sub_model.v3
  Field required [type=missing, input_value={'v1': 'oops'}, input_type=dict]
    For further information visit https://errors.pydantic.dev/2.7/v/missing
"""