Struct CpuPlan 

pub struct CpuPlan { /* private fields */ }

A compiled model prepared for CPU evaluation.

Implementations

impl CpuPlan

pub fn cache_event_batch( &self, batch: &EventBatch, ) -> RuntimeResult<CpuBatchCache>

Materializes the event-dependent cache for a batch.

Errors

Returns RuntimeError when required columns are missing, expression shapes are invalid, cache construction fails, or a matrix is singular.

Panics

Panics if a node selected by the validated cache plan was not evaluated.

impl CpuPlan

pub fn evaluate_with_gradient( &self, params: &ParamValues, ) -> RuntimeResult<ValueGradient>

Evaluates an event-independent model and its free-parameter gradient.

Errors

Returns RuntimeError when parameters are incompatible, the model requires event data, differentiation or evaluation fails, or a solve is singular.

pub fn evaluate_cache( &self, params: &ParamValues, cache: &CpuBatchCache, ) -> RuntimeResult<Vec<Complex64>>

Evaluates every row in a materialized batch cache.

Errors

Returns RuntimeError when parameters or cache layout are incompatible, evaluation fails, or a matrix is singular.

pub fn evaluate_cache_with_gradient( &self, params: &ParamValues, cache: &CpuBatchCache, ) -> RuntimeResult<Vec<ValueGradient>>

Evaluates every cached row and its free-parameter gradient.

Errors

Returns RuntimeError when parameters or cache layout are incompatible, or differentiation or evaluation fails.

impl CpuPlan

pub fn cache_dataset( &self, dataset: &Dataset, ) -> RuntimeResult<CpuCachedDataset>

Materializes all event-dependent caches for a dataset.

Errors

Returns RuntimeError when the dataset cannot be read, a batch schema is incompatible, cache construction fails, or a matrix is singular.

pub fn cache_memory_estimate(&self, events: usize) -> usize

Estimates retained compiled-cache bytes for events.

pub fn prepare_dataset( &self, execution: &Execution, dataset: &Dataset, ) -> RuntimeResult<CpuPreparedDataset>

Prepares a dataset according to its cache-storage policy.

Errors

Returns RuntimeError when dataset reading or cache construction fails, or another distributed worker reports failure.

Panics

Panics if successful preparation planning fails to record its staging memory decision, which would violate the planner invariant.

impl CpuPlan

pub fn reduce( &self, execution: &Execution, params: &ParamValues, dataset: &CpuPreparedDataset, reduction: ReductionPlan, ) -> RuntimeResult<f64>

Execute a weighted reduction over a prepared dataset.

Errors

Returns RuntimeError when streaming, cache validation, evaluation, or reduction fails, or another distributed worker reports failure.

pub fn reduce_with_gradient( &self, execution: &Execution, params: &ParamValues, dataset: &CpuPreparedDataset, reduction: ReductionPlan, ) -> RuntimeResult<ReductionEvaluation>

Execute a weighted reduction and its free-parameter gradient.

Errors

Returns RuntimeError when streaming, cache validation, differentiation, evaluation, or reduction fails, or another distributed worker reports failure.

pub fn evaluate_cached_dataset( &self, params: &ParamValues, dataset: &CpuCachedDataset, ) -> RuntimeResult<Vec<Complex64>>

Evaluates every event in a fully cached dataset.

Errors

Returns RuntimeError when parameters or a cache layout are incompatible, evaluation fails, or a matrix is singular.

pub fn evaluate_cached_dataset_with_gradient( &self, params: &ParamValues, dataset: &CpuCachedDataset, ) -> RuntimeResult<Vec<ValueGradient>>

Evaluates every event and gradient in a fully cached dataset.

Errors

Returns RuntimeError when parameters or a cache layout are incompatible, or differentiation or evaluation fails.

impl CpuPlan

pub fn parameter_count(&self) -> usize

Returns the number of parameters, including fixed parameters.

pub fn free_parameter_count(&self) -> usize

Returns the number of free parameters.

pub fn cache_plan(&self) -> &CachePlan

Returns the event-cache layout required by this plan.

pub fn required_event_scalars(&self) -> &[String]

Returns the event-scalar columns required by this prepared model.

pub fn evaluate(&self, params: &ParamValues) -> RuntimeResult<Complex64>

Evaluates a model that has no event-dependent inputs.

Errors

Returns RuntimeError when parameters are incompatible, the model requires event data, evaluation fails, or a matrix is singular.

pub fn evaluate_with_event( &self, params: &ParamValues, event: &impl EventLookup, ) -> RuntimeResult<Complex64>

Evaluates the model using values supplied by an event lookup.

Errors

Returns RuntimeError when a required event value is missing, parameters are incompatible, evaluation fails, or a solve is singular.

pub fn evaluate_with_event_and_gradient( &self, params: &ParamValues, event: &impl EventLookup, ) -> RuntimeResult<ValueGradient>

Evaluates the model and gradient using values supplied by an event lookup.

Errors

Returns RuntimeError when a required event value is missing, parameters are incompatible, or differentiation or evaluation fails.

pub fn evaluate_cache_row( &self, params: &ParamValues, cache: &CpuBatchCache, row: usize, ) -> RuntimeResult<Complex64>

Evaluates one row in a materialized batch cache.

Errors

Returns RuntimeError when row is out of range, parameters or cache layout are incompatible, evaluation fails, or a matrix is singular.

impl CpuPlan

pub fn evaluate_cache_row_with_gradient( &self, params: &ParamValues, cache: &CpuBatchCache, row: usize, ) -> RuntimeResult<ValueGradient>

Evaluates one cached row and its free-parameter gradient.

Errors

Returns RuntimeError when row is out of range, parameters or cache layout are incompatible, or differentiation or evaluation fails.

pub fn evaluate_batch( &self, params: &ParamValues, batch: &EventBatch, ) -> RuntimeResult<Vec<Complex64>>

Evaluates the model for every event in a batch.

Errors

Returns RuntimeError when cache materialization or evaluation fails.

pub fn evaluate_batch_with_gradient( &self, params: &ParamValues, batch: &EventBatch, ) -> RuntimeResult<Vec<ValueGradient>>

Evaluates the model and gradient for every event in a batch.

Errors

Returns RuntimeError when cache materialization, differentiation, or evaluation fails.

Trait Implementations

impl Clone for CpuPlan

fn clone(&self) -> CpuPlan

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for CpuPlan

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.