Struct Client 

pub struct Client { /* private fields */ }

The main client for interacting with the modelexpress_server via gRPC

Implementations

impl Client

pub async fn new(config: Config) -> CommonResult<Self>

Create a new client with the given configuration

pub async fn new_with_cache( config: Config, cache_config: CacheConfig, ) -> CommonResult<Self>

Create a new client with cache configuration

pub fn get_cache_config(&self) -> Option<&CacheConfig>

Get cache configuration

pub fn set_cache_config(&mut self, cache_config: CacheConfig)

Set cache configuration

pub fn list_cached_models(&self) -> CommonResult<CacheStats>

List cached models

pub fn clear_cached_model(&self, model_name: &str) -> CommonResult<()>

Clear specific model from cache

pub fn clear_all_cached_models(&self) -> CommonResult<()>

Clear entire cache

pub async fn get_model_path(&self, model_name: &str) -> Result<PathBuf>

pub async fn preload_model_to_cache( &mut self, model_name: &str, provider: ModelProvider, ignore_weights: bool, ) -> CommonResult<()>

Pre-download model to cache

pub async fn health_check(&mut self) -> CommonResult<Status>

Get the server health status

pub async fn send_request<T: DeserializeOwned>( &mut self, action: impl Into<String>, payload: Option<HashMap<String, Value>>, ) -> CommonResult<T>

Send a request to the server

pub async fn request_model_with_provider_and_fallback( &mut self, model_name: impl Into<String>, provider: ModelProvider, ignore_weights: bool, ) -> CommonResult<()>

Request a model from the server with a specific provider and automatic fallback This function will first try to use the server for streaming downloads. If the server is unavailable, it will fallback to downloading directly.

pub async fn request_model_with_provider( &mut self, model_name: impl Into<String>, provider: ModelProvider, ignore_weights: bool, ) -> CommonResult<()>

Request a model from the server with a specific provider This function will wait until the model is downloaded using streaming updates

pub async fn request_model( &mut self, model_name: impl Into<String>, ignore_weights: bool, ) -> CommonResult<()>

Request a model from the server using the default provider (Hugging Face) with automatic fallback This function will first try to use the server, then fallback to direct download if needed

pub async fn request_model_server_only( &mut self, model_name: impl Into<String>, ignore_weights: bool, ) -> CommonResult<()>

Request a model from the server only (no fallback) This function will wait until the model is downloaded using streaming updates from the server

pub async fn request_model_with_smart_fallback( model_name: impl Into<String>, provider: ModelProvider, config: Config, ignore_weights: bool, ) -> CommonResult<()>

Request a model with automatic server fallback, creating client connection only if needed This function will try to download via server if possible, otherwise download directly

pub async fn download_model_directly( model_name: impl Into<String>, provider: ModelProvider, ignore_weights: bool, ) -> CommonResult<()>

Download a model directly without using the server This bypasses the server entirely and downloads the model using the specified provider