src.config

🦉 AthenaAI is an AI-powered ensemble machine-learning trading bot for PocketOption, built on BinaryOptionsToolsV2 → Get it on GitLab
💬 Need help? Join the Chipa Discord · ⚡ Go no-code with ChipaEditor

Source: src/config.py. Defines a single dataclass, BotConfig, holding all tuneable knobs for the live bot. It is constructed in main.py (build_config()) with a subset of fields overridden from environment variables — the rest keep the defaults below. See the environment reference for env-settable fields and the config field reference for the code-only fields.

`@dataclass BotConfig`

All fields with their defaults, units, and where they are consumed:

Connection

Field	Type	Default	Meaning / consumed in
`ssid`	`str`	`""`	PocketOption session ID. Passed to `PocketOptionAsync(ssid=...)` in `src/bot.py`
`asset`	`str`	`"EURUSD"`	Trading pair, used for candle subscription and `buy()`/`sell()`
`timeframe`	`int`	`60`	Candle period in seconds

AI expiry selection

Field	Type	Default	Meaning
`expiry_options`	`tuple`	`(120, 300, 600)`	Candidate expiries in seconds fed to `ExpirySelector` (60s/180s dropped in v6 as bad performers)
`default_expiry`	`int`	`300`	Seconds; used as the label horizon when pre-training from a dataset in `src/bot.py` `_load_dataset()`

Money management

Field	Type	Default	Meaning
`base_stake`	`float`	`25.0`	Minimum trade size ($), floor in `MoneyManager.compute_stake()`
`max_stake`	`float`	`100.0`	Hard stake ceiling ($)
`kelly_fraction`	`float`	`0.50`	Fraction of full Kelly used for sizing
`max_daily_loss`	`float`	`300.0`	Daily stop-loss ($); `MoneyManager.can_trade()` returns `False` once daily P&L reaches −this
`max_concurrent_trades`	`int`	`1`	Max simultaneously open trades (gate in `_trade_loop`)

ML and signals

Field	Type	Default	Meaning
`warmup_candles`	`int`	`60`	Candles required before the first trade (the bot enforces a floor of 60 regardless — features need ≥50 candles for SMA50)
`min_confidence`	`float`	`0.63`	Minimum ensemble confidence (fraction) to trade
`max_confidence`	`float`	`0.85`	Signals above this are rejected as likely overfit
`retrain_every`	`int`	`10`	`partial_fit()` after N new live samples
`lookback`	`int`	`200`	Max candle history kept (`deque(maxlen=lookback)`)
`feature_window`	`int`	`20`	Rolling window (candles) passed to `FeatureEngine.compute()`

Signal readiness

Field	Type	Default	Meaning
`signal_confirmations`	`int`	`1`	Consecutive candles that must agree on direction (1 = instant)
`require_indicator_alignment`	`bool`	`True`	Require 2-of-3 of RSI/MACD/SMA-cross to agree with the ML direction
`skip_volatile_regime`	`bool`	`False`	Skip trades when regime is `VOLATILE` (off by default — ML decides)
`min_wait_between_trades`	`int`	`60`	Seconds between trade placements

Trading schedule

Field	Type	Default	Meaning
`trading_hours`	`tuple`	`(0, 1, 2, 3, 4, 16, 17, 19, 20, 21, 22, 23)`	UTC hours during which trading is allowed. Per the source comments, based on a 546-trade analysis showing 08–15 UTC loses money. An empty tuple disables the filter

Dataset, risk, persistence, misc

Field	Type	Default	Meaning
`dataset_path`	`str`	`""`	CSV for pre-training (optional; `main.py` overrides with `"EURUSD_M1.csv"`)
`max_consec_losses`	`int`	`3`	Consecutive losses triggering a cooldown
`cooldown_seconds`	`int`	`300`	Cooldown duration in seconds
`regime_window`	`int`	`30`	Candles used by `RegimeDetector.detect()`
`db_path`	`str`	`"trade_journal.db"`	SQLite journal file
`brain_path`	`str`	`"athena_po_brain.pkl"`	Saved-model pickle path (`main.py` overrides with `"athena_brain.pkl"` via `PO_BRAIN_PATH`)
`poll_interval`	`float`	`2.0`	Seconds between trade-loop iterations

There are no methods and no validation — the dataclass is a plain container. Invalid combinations (e.g. base_stake > max_stake) are not rejected here.

Usage example

from src.config import BotConfig
from src.bot import AITradingBot
import asyncio

cfg = BotConfig(ssid="your-session-id", asset="EURUSD",
                base_stake=10.0, max_stake=50.0, min_confidence=0.65)
asyncio.run(AITradingBot(cfg).start())

Risk note

base_stake, max_stake, kelly_fraction, and max_daily_loss directly control money at risk. Binary options trading can lose money; past performance does not guarantee future results. See the risk management guide.