Skip to main content

vantage_sdk.workbench.size_preset

SDK helpers + mirror models for size presets.

Classes​

GpuSpec(*, vendor: str = 'nvidia.com/gpu', count: int = 1) -> None​

GPU request. Absent on the parent means CPU-only.

SizePreset(*, name: str, description: str | None = None, labels: dict[str, str] = <factory>, hidden: bool = False, created_by: str | None = None, sizing: vantage_sdk.workbench.size_preset.SizingSpec = <factory>, workload: str | None = None) -> None​

Mirror of the server's size preset (envelope + sizing).

SizePresetSDK()​

SDK for the /size-preset CRUD (per-workload catalogs).

Methods​

  • create(self, ctx: 'typer.Context', *, cluster_name: 'str', preset: 'dict | BaseModel') -> 'ClusterServiceResponse': Create a size preset.
  • delete(self, ctx: 'typer.Context', *, cluster_name: 'str', workload: 'str', name: 'str') -> 'ClusterServiceResponse': Delete a size preset (idempotent server-side).
  • get(self, ctx: 'typer.Context', *, cluster_name: 'str', workload: 'str', name: 'str') -> 'ClusterServiceResponse': Get a single size preset by workload + name.
  • list(self, ctx: 'typer.Context', *, cluster_name: 'str', workload: 'str | None' = None) -> 'ClusterServiceResponse': List size presets, optionally scoped to one workload catalog.
  • normalize_workload(workload: 'str') -> 'str': Lowercase and validate a workload catalog against SIZING_WORKLOADS.
  • update(self, ctx: 'typer.Context', *, cluster_name: 'str', workload: 'str', name: 'str', preset: 'dict | BaseModel') -> 'ClusterServiceResponse': Full replace of an existing size preset (PUT).

SizingSpec(*, cpu: str = '1', memory: str = '2Gi', gpu: vantage_sdk.workbench.size_preset.GpuSpec | None = None, compute_pool: str | None = None) -> None​

Shared sizing vocabulary used by every workload catalog.