lash-core 0.1.0-alpha.39

Sans-IO turn machine and runtime kernel for the lash agent runtime.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42

//! Provider components for pluggable LLM backends.
//!
//! A provider is split into narrow capabilities: configured state,
//! request transport, failure classification, and model policy.
//! [`ProviderHandle`] owns those components and is the executable handle
//! installed by the host for a running session.
//!
//! [`ProviderSpec`] is only a host/config-file data shape. Runtime
//! persistence records provider identity separately and never rebuilds a
//! [`ProviderHandle`] from disk.

mod factory;
mod handle;
mod model_policy;
mod options;
mod rate_limit;
mod resolver;
mod spec;
mod support;
#[cfg(test)]
mod tests;
mod traits;

pub use factory::ProviderFactory;
pub use handle::{ProviderComponents, ProviderHandle, UnconfiguredProvider};
pub use model_policy::StaticModelPolicy;
pub use options::{
    CacheRetention, DEFAULT_CHUNK_TIMEOUT_MS, DEFAULT_REQUEST_TIMEOUT_MS, LlmTimeouts,
    ProviderOptions, ProviderRateLimitPolicy, ProviderReliability, ProviderReliabilityBuilder,
    ProviderRetryPolicy, ProviderThinkingPolicy, ProviderTimeoutPolicy, RequestTimeout,
    ResolvedGenerationPolicy, resolve_generation_policy,
};
pub use rate_limit::{ProviderRateLimitPermit, ProviderRateLimiter};
pub use resolver::{
    EmptyProviderResolver, MapProviderResolver, ProviderBinding, ProviderResolutionError,
    RuntimeProviderResolver, SingleProviderResolver,
};
pub use spec::ProviderSpec;
pub use traits::{
    DefaultProviderFailureClassifier, Provider, ProviderFailureClassifier, ProviderModelPolicy,
    is_context_overflow_text,
};