precompute_curves

flipbook.precompute_curves(model_fn: Callable[[ndarray[tuple[Any, ...], dtype[float64]], ndarray[tuple[Any, ...], dtype[float64]]], ndarray[tuple[Any, ...], dtype[float64]]], t: _Buffer | _SupportsArray[dtype[Any]] | _NestedSequence[_SupportsArray[dtype[Any]]] | complex | bytes | str | _NestedSequence[complex | bytes | str], chain: _Buffer | _SupportsArray[dtype[Any]] | _NestedSequence[_SupportsArray[dtype[Any]]] | complex | bytes | str | _NestedSequence[complex | bytes | str] | Chain, *, steps: slice | tuple[int | None, int | None] | Sequence[int] | ndarray[tuple[Any, ...], dtype[integer[Any]]] | None = None, walkers: str | int | Sequence[int] | ndarray[tuple[Any, ...], dtype[integer[Any]]] = 'all', log_prob: _Buffer | _SupportsArray[dtype[Any]] | _NestedSequence[_SupportsArray[dtype[Any]]] | complex | bytes | str | _NestedSequence[complex | bytes | str] | None = None, param_transform: Callable[[ndarray[tuple[Any, ...], dtype[float64]]], ndarray[tuple[Any, ...], dtype[float64]]] | None = None, vectorized: bool = False, n_jobs: int = 1, chunk_size: int = 32, topk_by_logp: int | None = None, max_curves_per_frame: int | None = None) → Iterator[dict[str, object]]

Pre-compute model curves for later use.

Parameters:

model_fncallable

Callable implementing f(theta, t) -> y.

tarray_like

One-dimensional array of time samples.

chainarray_like or Chain

Chain data or Chain instance.

stepsslice, tuple, sequence of int, optional

Steps to pre-compute. When None all steps are considered.

walkers{‘all’, int, sequence of int}, optional

Walker selection specification.

log_probarray_like, optional

Log-probability values for the chain.

param_transformcallable, optional

Optional transformation applied prior to model evaluation.

vectorizedbool, optional

Indicates that model_fn supports vectorized evaluation.

n_jobsint, optional

Number of worker threads for non-vectorized evaluation.

chunk_sizeint, optional

Batch size for threaded evaluation.

topk_by_logpint, optional

If provided, restricts walkers to the top-K by log probability per step.

max_curves_per_frameint, optional

Maximum number of curves retained per step.

Returns:

generator of dict

Generator yielding dictionaries with keys 'step_index', 'walker_indices', 'curves', and 'log_prob'.