Constraints

qcs.Constraint

Bases: BaseModel

Defines a general constraint for portfolio assets. It is not specific to any given snapshot. Used as input for snapshot.build_portfolio_constraints()

Parameters:

Name	Type	Description	Default
type	Literal['limit_min', 'limit_max', 'target']	The constraint type. type="target" means that both limit_min and limit_max are set to the same value.	required
value	float	The constraint value.	required
filter	Callable[[Asset], bool]	Function to filter applicable assets. Receives an Asset object as argument and returns a boolean.	required
exposure	Callable[[Asset], float] | float = 1	Function to calculate the exposure of an asset, or a fixed exposure value. Default is 1. If given a function: This function receives an Asset object as argument and should return a float or None. If it returns None, then exposure=0 will be used for that position	required
asset_exposure_type	Literal['weight', 'abs_weight', 'threshold']	The type of asset exposure. Defaults to "weight".	required
label	str	Constraint label, default is "".	required

Examples:

>>> Constraint(
...     type="limit_min",
...     value=0.1,
...     filter=lambda asset: asset.type == "EQT",
...     exposure=lambda asset: asset.othernum
... )

qcs.OptimizerConstraint

Bases: BaseModel

Defines a global constraint for the optimizer.

Parameters:

Name	Type	Description	Default
constraint_type	Literal['std', 'std=', 'std.bm', 'var', 'cvar', 'mdd', 'cmdd', 'turnover', 'min_return', 'scr']	Specifies the type of constraint. Can be one of the following: - std - Volatility - std= - Volatility target - std.bm - Tracking error - var - Value at risk (requires probability) - cvar - Conditional value at risk (requires probability) - mdd - Maximum drawdown (requires probability) - cmdd - Conditional maximum drawdown (requires probability) - turnover - Total portfolio turnover - min_return - Minimum return - scr - SCR	required
target	float	The target value for the constraint. Required for all constraint types.	required
probability	float | None	The probability level for the constraint. Only applicable for var, cvar, mdd, and cmdd constraint types.	required

Examples:

>>> OptimizerConstraint(constraint_type="std", target=0.2)
OptimizerConstraint(constraint_type='std', target=0.2, probability=None)

>>> OptimizerConstraint(constraint_type="var", target=0.05, probability=0.95)
OptimizerConstraint(constraint_type='var', target=0.05, probability=0.95)

qcs.PortfolioConstraint

Bases: BaseModel

Specific portfolio constraint definition. Must be in line with a given snapshot.

Parameters:

Name	Type	Description	Default
exposure_coefficients	List[float]	A list of exposure coefficients, where each coefficient must be a float. Must have the same length as positions. Typically exposure is 1 or 0, but can be any number, including nan for subportfolio constraints.	required
limit_min	float = 0.0	The minimum limit. Defaults to 0.0	required
limit_max	float = 1.0	The maximum limit. Defaults to 1.0	required
label	str = ""	Constraint label. Not used in calculations. Defaults to "".	required
asset_exposure_type	Literal["weight", "abs_weight", "threshold"] = "weight"	The type of asset exposure. Defaults to "weight". weight: check if weights ⋅ exposure_coefficients are in [limit_min, limit_max] abs_weight: check if abs(weights) ⋅ exposure_coefficients are in [limit_min, limit_max] threshold: check if weights[weights > limit_min] ⋅ exposure_coefficients are < limit_max	required

Examples:

>>> positions = [
...     {"ticker": "MSFT", "allocation": 200.0},
...     {"ticker": "AAPL", "allocation": 100.0},
... ]
>>> snapshot = AllocationSnapshot(currency="EUR", positions=positions)
>>> portfolio_constraint = PortfolioConstraint(
...     limit_min=0.5, limit_max=0.7, exposure_coefficients=[1.0, 0.0]
... )

qcs.PositionConstraint

Bases: BaseModel

Defines a position level constraint for a portfolio. Used as input for snapshot.apply_position_constraints()

Parameters:

Name	Type	Description	Default
type		The constraint type.	required
filter	Callable[[Asset], bool]	Function to filter applicable assets. Receives an Asset object as argument and returns a boolean.	required
value	float | None	The constraint value. Used and required only for type="turnover"	required

Examples:

>>> # Locks all EQT positions
>>> PositionConstraint(type="lock", filter=lambda asset: asset.type == "EQT")
>>> # All positions can be long/short
>>> PositionConstraint(type="long_short", filter=lambda _asset: True)
>>> # Sets the max turnover to 10% for each USA asset
>>> PositionConstraint(
...     type="turnover", value=0.1, filter=lambda asset: asset.country == "USA"
... )