Struct InferenceEngine 

pub struct InferenceEngine { /* private fields */ }

Manages the PIE (Proxy Inference Engine) process lifecycle.

Handles:

• Binary fetching if not installed
• Process spawning with proper daemonization
• Per-process lease tracking for orchard-rs clients
• Registration with PIE’s management plane

Implementations

impl InferenceEngine

pub async fn new() -> Result<Self>

Create a new InferenceEngine and connect to (or spawn) the engine process.

pub async fn with_options( paths: EnginePaths, startup_timeout: Option<Duration>, ) -> Result<Self>

Create with custom options.

pub fn close(&mut self) -> Result<()>

Close this engine instance.

Releases this process’s orchard-rs lease and best-effort deregisters the client with PIE. Normal disconnect never sends signals to PIE.

pub fn shutdown(timeout: Duration) -> Result<()>

Force shutdown the engine regardless of reference count.

pub fn generate_response_channel_id() -> u64

Generate a unique response channel ID for this client.

Format: (PID << 32) | random_32_bits Uses true randomness to avoid collisions between rapid successive calls.

Trait Implementations

impl Drop for InferenceEngine

fn drop(&mut self)

Executes the destructor for this type.