Crate ai_batch_queue 

AI Batch Queue

Model-aware batch processing queue with ETA estimation for Tauri applications.

Key Features

• Resource-aware reordering — automatically groups jobs by resource key (e.g. model name) to minimize expensive swaps
• Size-bucketed ETA estimation — tracks processing durations by (resource, operation, size) for accurate time predictions
• Item-level status tracking — each item has its own lifecycle
• Overwrite policies — skip items that already have results
• Progressive completion with retry — failed items can be retried without re-processing successful ones
• Trace context propagation — optional stack-ids TraceCtx, AttemptId, and TrialId on each batch item for cross-crate observability and retry lineage tracking

Quick Start

1. Define your item data type
2. Implement BatchItemHandler for your processing logic
3. Create a BatchQueue and register it in Tauri state
4. Call executor::spawn() to start the background processor

Re-exports

pub use queue::BatchQueue;

pub use types::BatchCompletionSummary;

pub use types::BatchItem;

pub use types::BatchItemStatus;

pub use types::BatchJob;

pub use types::BatchJobStatus;

pub use types::EtaConfidence;

pub use types::EtaEstimate;

pub use types::ItemResult;

pub use types::OverwritePolicy;

pub use types::SchedulingConfig;

pub use types::SizeBucket;

Modules

eta

executor

queue

types

Traits

BatchItemHandler

Trait for processing individual items in a batch.

BatchStore

Trait for persisting batch queue state.

Functions

build_job

Helper to build a BatchJob from a list of items.

build_job_traced

Helper to build a BatchJob with trace context propagation.