Files
sentiment-engine/MALKHUT/malkhut/actions.py
Codex d24d9bc6bd malkhut(wire): OrderType as three orthogonal dimensions — Fable's corrections
CRITICAL REFACTOR based on Fable's review (S9 roadmap item):

Before: flat enum conflating order types with TIF/instructions
  OrderType had MARKET, LIMIT, IOC, FOK, POST_ONLY, REDUCE_ONLY, etc.

After: three orthogonal dimensions (FIX-aligned):
  1. OrderType (Tag 40): what the order IS
     LIMIT, MARKET, STOP_MARKET, STOP_LIMIT, TRIGGER_MARKET, TRIGGER_LIMIT,
     TRAILING_STOP, OCO, TP_SL
  2. TimeInForce (Tag 59): how long it LIVES
     GTC, IOC, FOK, GTD
  3. Instructions (Tag 18): behavioral modifiers
     POST_ONLY, REDUCE_ONLY, HIDDEN, ICEBERG

Key corrections:
- POST_ONLY is an instruction on a LIMIT order, not a standalone type
- IOC/FOK are TimeInForce values, not order types
- BingX trailing_stop -> native TRAILING_STOP_MARKET (not TRIGGER_MARKET)
- FulfilmentAction.time_in_force: new field, default GTC

Exchange mappings restructured:
  EXCHANGE_ORDER_TYPE_MAP: OrderType -> exchange native 'type' param
  EXCHANGE_TIF_MAP: TimeInForce -> exchange native 'timeInForce' param
  EXCHANGE_INSTRUCTION_MAP: Instruction -> exchange encoding

21 files changed. 380+ tests pass. Backward compatible.
2026-07-14 14:46:44 +02:00

76 lines
2.1 KiB
Python

"""
Action model — compact menu for simultaneous-move tree search.
Bad: enumerate every price tick x every quantity x every TTL x every TIF.
Good: 8-24 meaningful actions per player, 3-12 per counterparty role.
"""
from __future__ import annotations
from dataclasses import dataclass, field
from typing import Any, Mapping, Optional, Tuple
from malkhut.state import ActionKind, AgentRole, OrderType, Side
# Lazy import to avoid circular dependency (training -> actions -> training)
_TimeInForce = None
def _get_TimeInForce():
global _TimeInForce
if _TimeInForce is None:
from malkhut.training.order_types import TimeInForce
_TimeInForce = TimeInForce
return _TimeInForce
@dataclass(frozen=True, slots=True)
class FulfilmentAction:
"""One atomic action candidate."""
kind: ActionKind
side: Optional[Side]
order_type: Optional[OrderType]
price_ticks_from_best: int
qty_fraction: float
ttl_ms: int
cancel_order_id: Optional[str] = None
reduce_only: bool = False
post_only: bool = False
time_in_force: str = "GTC"
metadata: Mapping[str, Any] = field(default_factory=dict)
@property
def time_in_force_enum(self):
TIF = _get_TimeInForce()
return TIF(self.time_in_force)
@dataclass(frozen=True, slots=True)
class CounterpartyAction:
"""Adversarially useful aggregate actions that alter book/fill outcomes."""
role: AgentRole
kind: ActionKind
side: Optional[Side]
price_ticks_from_best: int
qty_fraction_of_top: float
toxicity: float = 0.0
metadata: Mapping[str, Any] = field(default_factory=dict)
JointAction = Tuple[Any, ...] # (our_action, cp_action_1, cp_action_2, ...)
@dataclass(frozen=True, slots=True)
class PlannedPolicy:
"""Output of the planner: distribution over actions + selected action."""
actions: Tuple[FulfilmentAction, ...]
probabilities: Tuple[float, ...]
selected_action: FulfilmentAction
diagnostics: Mapping[str, Any]
@dataclass(frozen=True, slots=True)
class RiskDecision:
approved: bool
action: Optional[FulfilmentAction]
reason: str
adjusted: bool = False