Struct Dataset 

pub struct Dataset { /* private fields */ }

Lazy event dataset combining a source, read plan, and row transformations.

Implementations

impl Dataset

pub fn new<S>(source: S) -> Self

where S: EventSource + 'static,

Creates a dataset from an event source.

pub fn from_arc(source: Arc<dyn EventSource>) -> Self

Creates a dataset from a shared dynamically dispatched source.

pub fn from_batch(batch: EventBatch) -> Self

Creates an in-memory dataset from one batch.

pub fn from_batches(batches: Vec<EventBatch>) -> LadduDataResult<Self>

Creates an in-memory dataset from schema-compatible batches.

Errors

Returns LadduDataError when batch schemas are incompatible.

pub fn from_events<I>(schema: Arc<Schema>, events: I) -> LadduDataResult<Self>

where I: IntoIterator<Item = OwnedEvent>,

Collects owned events into an in-memory dataset.

Errors

Returns LadduDataError when an event does not match schema.

pub fn schema(&self) -> LadduDataResult<Arc<Schema>>

Returns the source schema.

Errors

Returns LadduDataError when the underlying source cannot determine or load its schema.

pub fn capabilities(&self) -> SourceCapabilities

Returns source planning capabilities.

pub fn num_events(&self) -> LadduDataResult<Option<u64>>

Returns the source event count when cheaply available.

Errors

Returns an error when source metadata cannot be read.

pub fn stats(&self) -> LadduDataResult<DatasetStats>

Returns cached event-count and weight-sum statistics, computing them once if needed.

Clones of a dataset view share this cache. Failed traversals are not cached and may be retried.

Errors

Returns LadduDataError when reading or transforming the dataset fails.

pub fn read_plan(&self) -> ReadPlan

Returns the current read plan.

pub fn cache_storage(&self) -> CacheStorage

Returns the compiled-cache memory policy.

pub fn memory_policy(&self) -> MemoryPolicy

Returns the memory-first cache selection policy.

pub fn memory_budget(&self) -> MemoryBudget

Returns the dataset’s host-memory budget.

pub fn last_memory_decision(&self) -> Option<MemoryDecision>

Returns the most recent memory-derived read decision.

pub fn source_traversals(&self) -> u64

Returns the number of transformed source iterators opened by this dataset view.

pub fn with_memory_budget(self, budget: MemoryBudget) -> Self

Returns this dataset with a host-memory budget.

pub fn fastest(self) -> Self

Select the fastest strategy allowed by the active memory budget.

pub fn resident(self) -> Self

Retain compiled event-dependent values for every local event.

This strict policy is intended for repeatedly evaluating a likelihood and returns an error if the resident cache cannot fit. Dataset::fastest is the default.

pub fn streaming(self) -> Self

Re-read the source and rebuild one batch cache during each parameter evaluation.

Only fixed dataset statistics are retained. This minimizes memory use but requires a repeatable source and is expected to be slower than Dataset::resident.

pub fn chunked(self, chunk_size: usize) -> LadduDataResult<Self>

Returns this dataset with a low-level nonzero maximum event count.

Prefer Dataset::with_memory_budget for portable application code. This explicit read-plan override remains available for source debugging and reproducibility and is always capped by execution memory planning.

Errors

Returns LadduDataError::InvalidArgument when chunk_size is zero.

pub fn unchunked(self) -> Self

Returns this dataset with source-native batch sizes.

pub fn filter<F>(self, f: F) -> Self

where F: Fn(Event<'_>) -> bool + Send + Sync + 'static,

Lazily retains events satisfying f.

pub fn subsample(self, fraction: f64, seed: u64) -> LadduDataResult<Self>

Lazily retains a deterministic fraction of events.

Errors

Returns LadduDataError::InvalidArgument when fraction is outside [0, 1] or is NaN.

pub fn bootstrap(self, seed: u64) -> Self

Applies deterministic Poisson bootstrap multiplicities to event weights.

pub fn for_each_event<F>(&self, f: F) -> LadduDataResult<()>

where F: FnMut(Event<'_>),

Visits each transformed event.

Errors

Returns LadduDataError when reading or transforming the source fails.

pub fn try_for_each_event<F>(&self, f: F) -> LadduDataResult<()>

where F: FnMut(Event<'_>) -> LadduDataResult<()>,

Visits each transformed event and stops at the first error.

Errors

Returns the first LadduDataError produced by the source, transformations, or callback.

pub fn try_map_events<T, F>(&self, f: F) -> LadduDataResult<Vec<T>>

where F: FnMut(Event<'_>) -> LadduDataResult<T>,

Fallibly maps transformed events into a vector.

Errors

Returns the first LadduDataError produced while reading, transforming, or mapping an event.

pub fn map_events<T, F>(&self, f: F) -> LadduDataResult<Vec<T>>

where F: FnMut(Event<'_>) -> T,

Maps transformed events into a vector.

Errors

Returns LadduDataError when reading or transforming the source fails.

pub fn try_fold_events<T, F>(&self, init: T, f: F) -> LadduDataResult<T>

where F: FnMut(T, Event<'_>) -> LadduDataResult<T>,

Fallibly folds transformed events into an owned accumulator.

Errors

Returns the first LadduDataError produced while reading, transforming, or folding an event.

pub fn fold_events<T, F>(&self, init: T, f: F) -> LadduDataResult<T>

where F: FnMut(T, Event<'_>) -> T,

Folds transformed events into an owned accumulator.

Errors

Returns LadduDataError when reading or transforming the source fails.

pub fn try_accumulate_events<T, F>(&self, acc: T, f: F) -> LadduDataResult<T>

where F: FnMut(&mut T, Event<'_>) -> LadduDataResult<()>,

Fallibly updates a mutable accumulator for every transformed event.

Errors

Returns the first LadduDataError produced while reading, transforming, or accumulating an event.

pub fn accumulate_events<T, F>(&self, acc: T, f: F) -> LadduDataResult<T>

where F: FnMut(&mut T, Event<'_>),

Updates a mutable accumulator for every transformed event.

Errors

Returns LadduDataError when reading or transforming the source fails.

pub fn sum_weights(&self) -> LadduDataResult<f64>

Sums effective event weights.

Errors

Returns LadduDataError when reading or transforming the source fails.

pub fn weighted_sum<F>(&self, f: F) -> LadduDataResult<f64>

where F: FnMut(Event<'_>) -> f64,

Sums weight * f(event) over transformed events.

Errors

Returns LadduDataError when reading or transforming the source fails.

pub fn weighted_complex_sum<F>(&self, f: F) -> LadduDataResult<Complex64>

where F: FnMut(Event<'_>) -> Complex64,

Sums complex weight * f(event) contributions.

Errors

Returns LadduDataError when reading or transforming the source fails.

pub fn batches( &self, ) -> LadduDataResult<Box<dyn Iterator<Item = LadduDataResult<EventBatch>> + Send>>

Opens an iterator of fully transformed event batches.

Errors

Returns LadduDataError when the underlying source cannot initialize a batch stream for the current plan.

pub fn try_for_each_batch<F>(&self, f: F) -> LadduDataResult<()>

where F: FnMut(EventBatch) -> LadduDataResult<()>,

Visits each transformed batch and stops at the first error.

Errors

Returns the first LadduDataError produced by the source, transformations, or callback.

pub fn map_batches<T, F>(&self, f: F) -> LadduDataResult<Vec<T>>

where F: FnMut(EventBatch) -> T,

Maps transformed batches into a vector.

Errors

Returns LadduDataError when reading or transforming a batch fails.

pub fn write_to<S: EventSink>(&self, sink: &mut S) -> LadduDataResult<()>

Streams the transformed dataset into an event sink.

Errors

Returns LadduDataError when reading, transforming, or writing a batch fails, or the sink cannot begin or finish the stream.

Trait Implementations

impl Clone for Dataset

fn clone(&self) -> Dataset

Returns a duplicate of the value.

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

Performs copy-assignment from source.