Struct ModelConfig 

pub struct ModelConfig {
    pub provider: Provider,
    pub model: String,
    pub endpoint: Option<String>,
    pub api_key: Option<String>,
    pub max_output_tokens: u32,
    pub timeout_secs: u64,
    pub max_retries: u32,
}

Fields

provider: Provider

model: String

e.g. “gpt-4o” or “llama3.1:8b”

endpoint: Option<String>

Ollama base URL override; default http://localhost:11434

api_key: Option<String>

OpenAI; read from OPENAI_API_KEY if None.

max_output_tokens: u32

Default 4096.

timeout_secs: u64

How long to wait for one response. There is no “off”: a request that waits forever is the most common way a program hangs.

max_retries: u32

How many times to retry a request that failed for a reason that may not repeat: a refused connection, a timeout, a 429, a 5xx.

There is no “off” for the same reason http retries a GET: a provider under load is the ordinary case, not the exceptional one, and a program that gives up on the first 429 is a program that gives up several times an hour.

Trait Implementations

impl Clone for ModelConfig

fn clone(&self) -> ModelConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ModelConfig

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.