bounded-models 0.0.5

A framework to check if Pydantic models are bounded

Bounded Models

[!WARNING] This project is in early development. The API may change significantly.

A framework to check if Pydantic models are "bounded" and enable uniform sampling from constrained spaces.

Installation

pip install bounded-models
# or
uv add bounded-models

Quick Start

from typing import Literal
from bounded_models import FieldHandlerRegistry
from pydantic import BaseModel, Field

class Config(BaseModel):
    mode: Literal["fast", "slow"]
    threshold: float = Field(ge=0.0, le=1.0)
    count: int = Field(ge=1, le=10)

registry = FieldHandlerRegistry.default()

registry.check_model_boundedness(Config)  # True
registry.model_dimensions(Config)  # 3

# Sample from unit hypercube [0, 1]^dim
registry.sample_model([0.5, 0.5, 0.5], Config)
# Config(mode='slow', threshold=0.5, count=5)

Supported Field Types

Type	Example
Numeric (int, float)	Field(ge=0, le=10)
Literal	Literal["a", "b", "c"]
Enum	class Color(Enum): ...
Nested BaseModel	BoundedModel subclasses

Documentation

For detailed usage, see the documentation.

License

MIT