ebinexpy 0.1.0

Async Python library for the Ebinex traderoom

ebinexpy

ebinexpy is an asynchronous Python library for controlling the operational capabilities of the Ebinex trading room from bots, workers, and other services. It is not an HTTP API, does not execute strategies, and does not depend on a browser at runtime.

The current contract is pre-alpha and supports accounts, profiles, balances, live and historical market data, and the OPTION order lifecycle. The TEST account is always the default, and orders on a REAL account are blocked without explicit opt-in.

Installation

pip install ebinexpy

For development:

python -m venv .venv
.venv/bin/pip install -e '.[dev]'

Client lifecycle

import asyncio
import os

from ebinexpy import EbinexClient


async def main() -> None:
    async with EbinexClient(os.environ["EBINEX_EMAIL"], os.environ["EBINEX_PASSWORD"]) as client:
        balance = await client.get_balance()
        assets = await client.list_assets()
        print(balance.amount, [asset.symbol for asset in assets if asset.tradable])


asyncio.run(main())

The constructor does not open connections. connect() authenticates, selects TEST, subscribes to essential topics, and only then marks the client as ready. disconnect() is idempotent and preserves the session; logout() disconnects and removes only the current identity's session. The context manager closes the HTTP and WebSocket connections and event handlers.

Sessions

The default is MemorySessionStore. To restore sessions across processes, use a private file:

from pathlib import Path
from ebinexpy import ClientConfig, EbinexClient

config = ClientConfig.with_file_sessions(Path.home() / ".local/state/ebinexpy")
client = EbinexClient("email", "password", config)

The store uses an identity-derived key, a 0700 directory, a 0600 file, and atomic replacement. The library does not read .env; load secrets into the process with your service's configuration tool.

Market data and streams

Assets and payouts always come from configModes.OPTION; the library does not maintain a static list of tradable assets. Candle dates must be timezone-aware.

from ebinexpy import Timeframe

stream = await client.stream_candles("IDXUSDT", Timeframe.M1)
async with stream:
    async for event in stream:
        print(event.candle.close, event.snapshot)

Candle, ticker, and order book streams have bounded queues and share subscriptions. Superseded snapshots may be coalesced to prevent a slow consumer from blocking the socket. Handlers registered with client.events also run without blocking the receive loop.

OPTION orders

from decimal import Decimal
from ebinexpy import Direction, OrderRequest, Timeframe

# Run only on a deliberately selected TEST account.
request = OrderRequest(
    symbol="IDXUSDT",
    direction=Direction.CALL,
    investment=Decimal("1"),
    timeframe=Timeframe.M1,
    price=Decimal("2998.92"),  # current price observed in the feed
)
order = await client.place_order(request)
settlement = await client.wait_order(order.id)

Only the window between submission and receipt of the broker ID is serialized; accepted orders are tracked independently. An ambiguous submission failure raises OrderSubmissionUnknownError and is never retried. A settlement timeout raises SettlementTimeoutError.last_order; a timeout or unknown result is never converted into a loss.

To enable a REAL account, explicitly construct ClientConfig(environment=AccountEnvironment.REAL, allow_real_trading=True). The same guard also protects client.raw.send() on the execution destination.

Raw access and stability

client.raw.request, subscribe, and send reuse authentication, TLS, readiness, redaction, and account selection. The raw REST/STOMP format is deliberately unstable and may change between pre-alpha versions. Safe HTTP requests may reauthenticate once after a 401 response; order execution is never automatically replayed.

See examples/read_market.py and the DEMO order example with an explicit gate in examples/demo_order.py. The documentation overview and public method examples are available in docs/index.md.