Struct LlmConfig 

pub struct LlmConfig {
    pub model: String,
    pub endpoint: String,
    pub dry_run: bool,
    pub target_grade: f64,
    pub max_refinement_attempts: usize,
    pub timeout_secs: u64,
    pub cache_disabled: bool,
    pub cache_dir: Option<PathBuf>,
}

Configuration for the LLM plugin.

Fields

model: String

Model name (e.g., "llama3", "mistral").

endpoint: String

Ollama API endpoint.

dry_run: bool

If true, print generated text but don’t write files.

target_grade: f64

Target Flesch-Kincaid Grade Level (default: 8.0).

max_refinement_attempts: usize

Max refinement attempts if readability exceeds target (default: 1).

timeout_secs: u64

Per-call HTTP timeout for the local LLM endpoint, in seconds (default: 120). Set via llm.timeout_secs in ssg.toml. Exceeding this budget returns SsgError::LlmTimeout — no zombie subprocess is left behind because the call goes through ureq, not curl (issue #520).

cache_disabled: bool

When true, skip the deterministic content-hash cache and always perform a live inference (issue #528). Driven by the --no-llm-cache CLI flag and the SSG_NO_LLM_CACHE env var so users debugging non-determinism can rule the cache out without nuking it on disk.

cache_dir: Option<PathBuf>

Optional override for the on-disk cache root. None resolves to LlmCache::default_cache_dir at call time — the platform-correct path (XDG / Library / %LOCALAPPDATA%) chosen by LlmCache. Tests and operators wanting a project-local cache override this directly.

Trait Implementations

impl Clone for LlmConfig

fn clone(&self) -> LlmConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for LlmConfig

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

Formats the value using the given formatter.

impl Default for LlmConfig

fn default() -> Self

Returns the “default value” for a type.