Brokers & Datafeeds

AQE keeps execution and market data behind broker/datafeed traits so strategy code can run against the same runtime surface in backtests, paper workflows, and live operation.

Current integrations

AQE currently exposes these concrete runtime integrations:

• execution brokers: PaperBroker, Mt5Broker
• datafeeds: YahooFinanceDataFeed, Mt5DataFeed

Start with the focused integration pages:

• PaperBroker and paper execution covers simulated execution, paper-account state, Yahoo market data, and the asset fee model.
• MT5 broker and datafeed integration covers the live MetaTrader 5 bridge, Expert Advisor setup, and MT5 fee data sources.

They are composed through UnifiedBroker, which gives the strategy runtime one consistent broker-facing surface.

Broker traits

aq-engine/src/core/broker/mod.rs

pub trait Broker {
    async fn connect(&self) -> Result<bool, BrokerError>;
    async fn disconnect(&self) -> Result<bool, BrokerError>;
    fn is_connected(&self) -> bool;
    fn get_name(&self) -> String;
    fn get_account_type(&self) -> Result<AccountType, BrokerError>;
}

pub trait OrderManagementProvider: Broker {
    async fn submit_order(&self, insight: Insight) -> Result<Order, BrokerError>;
    async fn cancel_order(&self, order_id: &str) -> Result<bool, BrokerError>;
    async fn close_position(&self, order_id: &str, qty: f64, price: Option<f64>) -> Result<bool, BrokerError>;
    async fn get_account(&self) -> Result<Account, BrokerError>;
    fn drain_trade_events(&self) -> Vec<(Order, TradeUpdateEvent)>;
}

Datafeed traits

aq-engine/src/core/broker/mod.rs

pub trait DataFeed {
    async fn connect(&self) -> Result<bool, BrokerError>;
    async fn disconnect(&self) -> Result<bool, BrokerError>;
    fn is_connected(&self) -> bool;
}

pub trait DataProvider: DataFeed {
    async fn get_ticker_info(&self, symbol: &str) -> Result<Asset, BrokerError>;
    async fn get_history(
        &self,
        symbol: &str,
        start: DateTime<Utc>,
        end: DateTime<Utc>,
        time_frame: TimeFrame,
    ) -> Result<DataFrame, BrokerError>;
    async fn get_latest_quote(&self, symbol: &str) -> Result<Quote, BrokerError>;
    async fn get_latest_bar(&self, symbol: &str) -> Result<Bar, BrokerError>;
}

Combining brokers and datafeeds

let execution = PaperBroker::new(AccountType::Paper, 100_000.0, 1);
let data = YahooFinanceDataFeed::new();

let broker = UnifiedBroker::new_backtest(execution, data);

For live operation, the same pattern applies, but the runtime uses the live path instead of new_backtest(...).

What AQS consumes from live execution

When live sync is enabled, AQE publishes broker/runtime state into session-scoped tables used by AQS:

• insights
• strategy_accounts
• strategy_equity_points
• strategy_live_metrics
• strategy_events

That means broker/datafeed behaviour directly shapes what the operator sees in the desktop UI.

Related source

• aq-engine/src/core/broker/mod.rs
• aq-engine/src/core/broker/types/mod.rs
• aq-engine/src/core/broker/paper_broker.rs
• aq-engine/src/core/broker/mt5_broker.rs
• aq-engine/src/core/broker/data_feeds/yahoo.rs