Module ai 

AI inference module, containing model registry, provider trait, and backends. Backend-agnostic AI inference types: registry, provider trait, cache, call log, adapters, and backends. Inference runs on a Ring 1 worker, never Ring 0.

Re-exports

pub use adapter::parse_response;

pub use adapter::AdapterError;

pub use cache::content_hash;

pub use cache::params_version;

pub use cache::AiCacheKey;

pub use cache::AiResultCache;

pub use cache::AiResultCacheConfig;

pub use cache::CachedOutput;

pub use call_log::AiCallLog;

pub use call_log::AiCallRecord;

pub use call_log::CallOutcome;

pub use provider::InferenceOutputs;

pub use provider::InferenceParams;

pub use provider::InferenceProvider;

pub use provider::InferenceRequest;

pub use provider::InferenceResponse;

pub use provider::ProviderError;

pub use provider::Usage;

pub use registry::BackendKind;

pub use registry::ModelBackend;

pub use registry::ModelEntry;

pub use registry::ModelRegistry;

pub use registry::RegistryError;

pub use registry::Task;

pub use runtime::AiRuntime;

pub use runtime::AiRuntimeError;

pub use runtime::ResolvedModel;

Modules

adapter

Adapt a raw provider response to a task’s per-row output. Stays Arrow-free; the operator builds the column from the returned InferenceOutputs.

backends

Concrete inference backends, feature-gated so the default build has no HTTP or ML weight. Transport only: task framing lives in the adapter, not here.

cache

Per-row inference result cache, keyed (content_hash, model_id, params_version).

call_log

Bounded in-memory log of inference calls, surfaced as laminar.ai_calls. One record per batch; written from Ring 1, never Ring 0. Oldest record is dropped when full; a monotonic counter tracks lifetime call count.

provider

InferenceProvider trait and request/response types.

registry

Model registry: maps SQL-referenced model names to backends and tasks. Built once at startup; AI functions resolve against it at plan time.

runtime

Assembled AI subsystem: registry, provider clients, result cache, and call log. Built once at startup; AiRuntime::resolve returns everything the inference operator needs to run a named model.