mermaid-model 0.25.0

Model-layer core for Mermaid: provider adapters, wire types, capability catalog, and the shared utilities they need.
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

//! Provider adapters module
//!
//! Contains implementations of the Model trait for Ollama, the
//! OpenAI-compatible long tail, Anthropic Claude, and Google Gemini.
//!
//! Each owns its wire format and nothing else: [`driver`] holds the read
//! loop they all used to carry a copy of.

pub mod anthropic;
/// One test suite driven over recorded response bodies, one per provider.
/// In-crate rather than under `tests/` so it can reach the protocol
/// structs, which are wire-format detail and not public API.
#[cfg(test)]
mod conformance;
pub mod driver;
pub mod gemini;
pub mod meta;
pub mod ollama;
pub mod ollama_sizing;
pub mod openai_compat;
pub mod output_budget;

/// A model's token limits as reported by its provider's models endpoint.
/// `None` means the provider didn't expose that limit — never a guess.
#[derive(Debug, Clone, Copy, Default, PartialEq, Eq)]
pub struct ModelLimits {
    pub max_context_tokens: Option<usize>,
    pub max_output_tokens: Option<usize>,
}