studio/mistralai-client-rs
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update to latest Mistral AI API (v1.0.0)
Some checks failed
Test / Test Documentation (push) Has been cancelled
Details
Test / Test Examples (push) Has been cancelled
Details
Test / Test (push) Has been cancelled
Details

Browse Source 
- Replace closed Model enum with flexible string-based Model type
  with constructor methods for all current models (Mistral Large 3,
  Small 4, Magistral, Codestral, Devstral, Pixtral, Voxtral, etc.)
- Add new API endpoints: FIM completions, Files, Fine-tuning, Batch
  jobs, OCR, Audio transcription, Moderations/Classifications, and
  Agent completions (sync + async for all)
- Add new chat fields: frequency_penalty, presence_penalty, stop,
  n, parallel_tool_calls, reasoning_effort, min_tokens, json_schema
  response format
- Add embedding fields: output_dimension, output_dtype
- Tool parameters now accept raw JSON Schema (serde_json::Value)
  instead of limited enum types
- Add tool call IDs and Required tool choice variant
- Add DELETE HTTP method support and multipart file upload
- Bump thiserror to v2, add reqwest multipart feature
- Remove strum dependency (no longer needed)
- Update all tests and examples for new API
This commit is contained in:
Sienna Meridian Satterwhite
2026-03-20 17:16:26 +00:00
parent 9ad6a1dc84
commit 79bc40bb15
33 changed files with 1977 additions and 622 deletions
Download Patch File Download Diff File Expand all files Collapse all files

187
src/v1/chat.rs
View File

@@ -11,13 +11,31 @@ pub struct ChatMessage {
pub content: String,
#[serde(skip_serializing_if = "Option::is_none")]
pub tool_calls: Option<Vec<tool::ToolCall>>,
/// Tool call ID, required when role is Tool.
#[serde(skip_serializing_if = "Option::is_none")]
pub tool_call_id: Option<String>,
/// Function name, used when role is Tool.
#[serde(skip_serializing_if = "Option::is_none")]
pub name: Option<String>,
}
impl ChatMessage {
pub fn new_system_message(content: &str) -> Self {
Self {
role: ChatMessageRole::System,
content: content.to_string(),
tool_calls: None,
tool_call_id: None,
name: None,
}
}
pub fn new_assistant_message(content: &str, tool_calls: Option<Vec<tool::ToolCall>>) -> Self {
Self {
role: ChatMessageRole::Assistant,
content: content.to_string(),
tool_calls,
tool_call_id: None,
name: None,
}
}
@@ -26,6 +44,18 @@ impl ChatMessage {
role: ChatMessageRole::User,
content: content.to_string(),
tool_calls: None,
tool_call_id: None,
name: None,
}
}
pub fn new_tool_message(content: &str, tool_call_id: &str, name: Option<&str>) -> Self {
Self {
role: ChatMessageRole::Tool,
content: content.to_string(),
tool_calls: None,
tool_call_id: Some(tool_call_id.to_string()),
name: name.map(|n| n.to_string()),
}
}
}
@@ -44,17 +74,32 @@ pub enum ChatMessageRole {
}
/// The format that the model must output.
///
/// See the [API documentation](https://docs.mistral.ai/api/#operation/createChatCompletion) for more information.
#[derive(Clone, Debug, Serialize, Deserialize)]
pub struct ResponseFormat {
#[serde(rename = "type")]
pub type_: String,
#[serde(skip_serializing_if = "Option::is_none")]
pub json_schema: Option<serde_json::Value>,
}
impl ResponseFormat {
pub fn text() -> Self {
Self {
type_: "text".to_string(),
json_schema: None,
}
}
pub fn json_object() -> Self {
Self {
type_: "json_object".to_string(),
json_schema: None,
}
}
pub fn json_schema(schema: serde_json::Value) -> Self {
Self {
type_: "json_schema".to_string(),
json_schema: Some(schema),
}
}
}
@@ -63,91 +108,83 @@ impl ResponseFormat {
// Request
/// The parameters for the chat request.
///
/// See the [API documentation](https://docs.mistral.ai/api/#operation/createChatCompletion) for more information.
#[derive(Clone, Debug)]
pub struct ChatParams {
/// The maximum number of tokens to generate in the completion.
///
/// Defaults to `None`.
pub max_tokens: Option<u32>,
/// The seed to use for random sampling. If set, different calls will generate deterministic results.
///
/// Defaults to `None`.
pub min_tokens: Option<u32>,
pub random_seed: Option<u32>,
/// The format that the model must output.
///
/// Defaults to `None`.
pub response_format: Option<ResponseFormat>,
/// Whether to inject a safety prompt before all conversations.
///
/// Defaults to `false`.
pub safe_prompt: bool,
/// What sampling temperature to use, between `Some(0.0)` and `Some(1.0)`.
///
/// Defaults to `0.7`.
pub temperature: f32,
/// Specifies if/how functions are called.
///
/// Defaults to `None`.
pub temperature: Option<f32>,
pub tool_choice: Option<tool::ToolChoice>,
/// A list of available tools for the model.
///
/// Defaults to `None`.
pub tools: Option<Vec<tool::Tool>>,
/// Nucleus sampling, where the model considers the results of the tokens with `top_p` probability mass.
///
/// Defaults to `1.0`.
pub top_p: f32,
pub top_p: Option<f32>,
pub stop: Option<Vec<String>>,
pub n: Option<u32>,
pub frequency_penalty: Option<f32>,
pub presence_penalty: Option<f32>,
pub parallel_tool_calls: Option<bool>,
/// For reasoning models (Magistral). "high" or "none".
pub reasoning_effort: Option<String>,
}
impl Default for ChatParams {
fn default() -> Self {
Self {
max_tokens: None,
min_tokens: None,
random_seed: None,
safe_prompt: false,
response_format: None,
temperature: 0.7,
temperature: None,
tool_choice: None,
tools: None,
top_p: 1.0,
}
}
}
impl ChatParams {
pub fn json_default() -> Self {
Self {
max_tokens: None,
random_seed: None,
safe_prompt: false,
response_format: None,
temperature: 0.7,
tool_choice: None,
tools: None,
top_p: 1.0,
top_p: None,
stop: None,
n: None,
frequency_penalty: None,
presence_penalty: None,
parallel_tool_calls: None,
reasoning_effort: None,
}
}
}
#[derive(Debug, Serialize, Deserialize)]
pub struct ChatRequest {
pub messages: Vec<ChatMessage>,
pub model: constants::Model,
pub messages: Vec<ChatMessage>,
pub stream: bool,
#[serde(skip_serializing_if = "Option::is_none")]
pub max_tokens: Option<u32>,
#[serde(skip_serializing_if = "Option::is_none")]
pub min_tokens: Option<u32>,
#[serde(skip_serializing_if = "Option::is_none")]
pub random_seed: Option<u32>,
#[serde(skip_serializing_if = "Option::is_none")]
pub response_format: Option<ResponseFormat>,
pub safe_prompt: bool,
pub stream: bool,
pub temperature: f32,
#[serde(skip_serializing_if = "Option::is_none")]
pub safe_prompt: Option<bool>,
#[serde(skip_serializing_if = "Option::is_none")]
pub temperature: Option<f32>,
#[serde(skip_serializing_if = "Option::is_none")]
pub tool_choice: Option<tool::ToolChoice>,
#[serde(skip_serializing_if = "Option::is_none")]
pub tools: Option<Vec<tool::Tool>>,
pub top_p: f32,
#[serde(skip_serializing_if = "Option::is_none")]
pub top_p: Option<f32>,
#[serde(skip_serializing_if = "Option::is_none")]
pub stop: Option<Vec<String>>,
#[serde(skip_serializing_if = "Option::is_none")]
pub n: Option<u32>,
#[serde(skip_serializing_if = "Option::is_none")]
pub frequency_penalty: Option<f32>,
#[serde(skip_serializing_if = "Option::is_none")]
pub presence_penalty: Option<f32>,
#[serde(skip_serializing_if = "Option::is_none")]
pub parallel_tool_calls: Option<bool>,
#[serde(skip_serializing_if = "Option::is_none")]
pub reasoning_effort: Option<String>,
}
impl ChatRequest {
pub fn new(
@@ -156,30 +193,28 @@ impl ChatRequest {
stream: bool,
options: Option<ChatParams>,
) -> Self {
let ChatParams {
max_tokens,
random_seed,
safe_prompt,
temperature,
tool_choice,
tools,
top_p,
response_format,
} = options.unwrap_or_default();
let opts = options.unwrap_or_default();
let safe_prompt = if opts.safe_prompt { Some(true) } else { None };
Self {
messages,
model,
max_tokens,
random_seed,
safe_prompt,
messages,
stream,
temperature,
tool_choice,
tools,
top_p,
response_format,
max_tokens: opts.max_tokens,
min_tokens: opts.min_tokens,
random_seed: opts.random_seed,
safe_prompt,
temperature: opts.temperature,
tool_choice: opts.tool_choice,
tools: opts.tools,
top_p: opts.top_p,
response_format: opts.response_format,
stop: opts.stop,
n: opts.n,
frequency_penalty: opts.frequency_penalty,
presence_penalty: opts.presence_penalty,
parallel_tool_calls: opts.parallel_tool_calls,
reasoning_effort: opts.reasoning_effort,
}
}
}
@@ -192,7 +227,7 @@ pub struct ChatResponse {
pub id: String,
pub object: String,
/// Unix timestamp (in seconds).
pub created: u32,
pub created: u64,
pub model: constants::Model,
pub choices: Vec<ChatResponseChoice>,
pub usage: common::ResponseUsage,
@@ -203,14 +238,18 @@ pub struct ChatResponseChoice {
pub index: u32,
pub message: ChatMessage,
pub finish_reason: ChatResponseChoiceFinishReason,
// TODO Check this prop (seen in API responses but undocumented).
// pub logprobs: ???
}
#[derive(Clone, Debug, Eq, PartialEq, Deserialize, Serialize)]
pub enum ChatResponseChoiceFinishReason {
#[serde(rename = "stop")]
Stop,
#[serde(rename = "length")]
Length,
#[serde(rename = "tool_calls")]
ToolCalls,
#[serde(rename = "model_length")]
ModelLength,
#[serde(rename = "error")]
Error,
}