aimx/inference/
provider.rs

1//! Inference provider configuration for external AI APIs used by AIMX.
2//! Defines serializable types for selecting APIs, endpoints, models,
3//! and runtime limits so hosts can configure inference behavior.
4
5use crate::inference::Capability;
6use serde::{Deserialize, Serialize};
7use std::fmt;
8
9/// AI inference APIs supported by AIMX configuration.
10#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
11pub enum Api {
12    /// Ollama API for local model inference.
13    Ollama,
14    /// OpenAI-compatible chat completions API.
15    Openai,
16}
17
18impl fmt::Display for Api {
19    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
20        match self {
21            Api::Ollama => write!(f, "ollama"),
22            Api::Openai => write!(f, "openai"),
23        }
24    }
25}
26
27impl Api {
28    /// Construct from identifier: "ollama" => `Ollama`, other => `Openai`.
29    pub fn new(api: &str) -> Self {
30        match api {
31            "ollama" => Api::Ollama,
32            _ => Api::Openai,
33        }
34    }
35}
36
37/// Configuration for an AI inference provider used by AIMX.
38#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
39pub struct Provider {
40    /// API implementation.
41    pub api: Api,
42    /// Base URL for the API.
43    pub url: String,
44    /// API key or token (empty if not required).
45    pub key: String,
46    /// Model identifier for the provider.
47    pub model: String,
48    /// Abstract capability class of the model.
49    pub capability: Capability,
50    /// Sampling temperature.
51    pub temperature: f64,
52    /// Maximum tokens to generate.
53    pub max_tokens: u32,
54    /// Context window size in tokens.
55    pub context_length: u32,
56    /// Connection timeout in milliseconds.
57    #[serde(default = "default_connection_timeout")]
58    pub connection_timeout_ms: u64,
59    /// Request timeout in milliseconds.
60    #[serde(default = "default_request_timeout")]
61    pub request_timeout_ms: u64,
62}
63
64/// Default connection timeout (milliseconds).
65fn default_connection_timeout() -> u64 {
66    30000
67}
68
69/// Default request timeout (milliseconds).
70fn default_request_timeout() -> u64 {
71    120000
72}
aimx/inference/provider.rs

aimx/inference/
provider.rs
