cyantic 0.1.5

Use Pydantic to build complex types from simple parameters.

The cyanotype process, which gave birth to the blueprint, uses iron salts and UV light to transform ordinary paper into deep Prussian blue prints.

Features

Cyantic lets you build complex types from simple blueprints, using Pydantic.

• Build complex objects using intermediate Pydantic models.
• Reference other values using @value:x.y.z
• Import objects using @import:x.y.z
• Define custom @reftag handlers (see tests)

Installation

• uv add cyantic

Usage

from cyantic import Blueprint, blueprint, CyanticModel
from torch import Tensor
import torch
import yaml

# 1. Create and register some useful parameterisations
#       (or soon install from PyPi, i.e. `rye add cyantic-torch`)

@blueprint(Tensor)
class NormalTensor(Blueprint[Tensor]):

    mean: float
    std: float
    size: tuple[int, ...]

    def build(self) -> Tensor:
        return torch.normal(self.mean, self.std, size=self.size)

@blueprint(Tensor)
class UniformTensor(Blueprint[Tensor]):
    low: float
    high: float
    size: tuple[int, ...]

    def build(self) -> Tensor:
      return torch.empty(self.size).uniform_(self.low, self.high)


# 2. Write pydantic models using `CyanticModel` base class

class MyModel(CyanticModel):
    normal_tensor: Tensor
    uniform_tensor: Tensor


# 3. Validate from YAML files that specify the parameterisation

some_yaml = """common:
    size: [3, 5]
normal_tensor:
    mean: 0.0
    std: 0.1
    size: @value:common.size
uniform_tensor:
    low: -1.0
    std: 1.0
    size: @value:common.size
"""

# 4. Receive objects built from the parameterisations.

my_model = MyModel.model_validate(yaml.safe_load(some_yaml))
assert isinstance(my_model.normal_tensor, Tensor)
assert isinstance(my_model.uniform_tensor, Tensor)

Development

• git clone https://github.com/flywhl/cyantic.git
• cd cyantic
• uv sync
• just test

Flywheel

Science needs humble software tools. Flywheel is an open source collective building simple tools to preserve scientific momentum, inspired by devtools and devops culture.