Data model rules#

Best practice

  • Every field needs a schema—use typed Pydantic fields with defaults.

  • Prefer Pydantic BaseModel subclasses over plain dict for structured data with a predictable shape.

  • Use EntityHandle for files and folders.

  • Keep field names stable—they become part of the migration contract.

Well-typed step model with diverse field types#
from enum import Enum
from pydantic import BaseModel
from ansys.bdm.api import EntityHandle, NO_ENTITY, RecursiveDictionaryOfEntityHandles
from ansys.saf.glow.solution import StepModel


class SolverType(Enum):
    """Available solver types."""

    DIRECT = "direct"
    ITERATIVE = "iterative"


class MeshSettings(BaseModel):
    """Mesh configuration — structured data with a predictable shape."""

    element_size: float = 1.0
    refinement_level: int = 2


class BeamBendingStep(StepModel):
    """Step model with representative field types."""

    # Primitives with defaults
    length: float = 1000.0  # mm
    diameter: float = 50.0  # mm
    load: float = 10e3  # N
    num_elements: int = 4
    is_nonlinear: bool = False

    # Enum
    solver: SolverType = SolverType.DIRECT

    # Nested Pydantic model (preferred over plain dict)
    mesh: MeshSettings = MeshSettings()

    # Collections
    deflection: list[float] = []
    material_properties: dict[str, float] = {}

    # BDM handles for files and directories
    result_file: EntityHandle = NO_ENTITY
    report_files: list[EntityHandle] = []
    output_directory: RecursiveDictionaryOfEntityHandles = RecursiveDictionaryOfEntityHandles()

    # Optional
    notes: str | None = None

Supported field types#

  • Primitives: int, float, str, bool, bytes

  • Collections: list[T], dict[str, T], tuple[...]

  • Pydantic models: Any BaseModel subclass

  • Special SAF types: EntityHandle, RecursiveDictionaryOfEntityHandles

  • Enums: Any enum.Enum subclass

  • Optional: T | None

On this page