Struct xaynet_client::Client

pub struct Client<C: ApiClient> {
    pub participant: Participant,
    pub has_new_coord_pk_since_last_check: bool,
    pub global_model: Option<Model>,
    pub cached_model: Option<CachedModel>,
    pub has_new_global_model_since_last_check: bool,
    pub has_new_global_model_since_last_cache: bool,
    pub local_model: Option<Model>,
    pub scalar: f64,
    // some fields omitted
}

A client of the federated learning service

Client is responsible for communicating with the service, deserialising its messages and delegating their processing to the underlying Participant.

Fields

participant: Participant

The underlying Participant

has_new_coord_pk_since_last_check: bool

global_model: Option<Model>

cached_model: Option<CachedModel>

has_new_global_model_since_last_check: bool

has_new_global_model_since_last_cache: bool

local_model: Option<Model>

scalar: f64

Implementations

impl<C> Client<C> where
    C: ApiClient, 

pub fn new(period: u64, id: u32, api: C) -> Result<Self, ClientError<C::Error>>

Create a new Client with a given service address.

• period: time period at which to poll for service data, in seconds.
• id: an ID to assign to the Client.
• addr: service address to connect to.

Errors

Returns a ParticipantInitErr if the underlying Participant is unable to initialize.

pub async fn start<'_>(&'_ mut self) -> Result<(), ClientError<C::Error>>

Starts the Client loop, iterating indefinitely over each federated learning round.

Errors

A ClientError may be returned when the round is not able to complete successfully.

pub async fn during_round<'_>(
    &'_ mut self
) -> Result<Task, ClientError<C::Error>>

Client work flow over a federated learning round. A successfully completed round will return the Task of the client.

Errors

A ClientError may be returned when the round is not able to complete successfully.