from __future__ import annotations
import builtins
import itertools
import os
import threading
from enum import Enum
from typing import TYPE_CHECKING, Any
import gams.transfer as gt
import pandas as pd
from gams.core.gdx import GMS_DT_EQU
from gams.transfer._internals import (
EQU_TYPE,
TRANSFER_TO_GAMS_EQUATION_SUBTYPES,
)
import gamspy as gp
import gamspy._algebra.condition as condition
import gamspy._algebra.expression as expression
import gamspy._algebra.operable as operable
import gamspy._symbols.implicits as implicits
import gamspy._validation as validation
import gamspy.utils as utils
from gamspy._symbols.symbol import Symbol
from gamspy.exceptions import ValidationError
if TYPE_CHECKING:
from collections.abc import Sequence
from types import EllipsisType
from gamspy import Alias, Container, Set, Variable
from gamspy._algebra.expression import Expression
from gamspy._algebra.operation import Operation
EQ_TYPES = ["=e=", "=l=", "=g=", "=n=", "=x=", "=b="]
IRREGULAR_EQ_MAP = {
"nonbinding": "=n=",
"external": "=x=",
"boolean": "=b=",
}
[docs]
class EquationType(Enum):
"""
Enumeration of available equation types.
"""
REGULAR = "regular"
"""Regular equations with =, >= and <= sign."""
NONBINDING = "nonbinding"
"""No relationship implied between left-hand side and right-hand side. This equation type is ideally suited for use in MCP models and in variational inequalities."""
EXTERNAL = "external"
"""Equation is defined by external programs."""
BOOLEAN = "boolean"
"""Boolean equations."""
[docs]
@classmethod
def values(cls):
"""Convenience function to return all values of enum"""
return list(cls._value2member_map_.keys())
def __str__(self) -> str:
return self.value
[docs]
class Equation(gt.Equation, Symbol):
"""
Represents an Equation symbol in GAMS.
Equations represent the constraints or relationships in a model. They can be
defined using equality (==) or inequality (<=, >=) operators.
See https://gamspy.readthedocs.io/en/latest/user/basics/equation.html
Parameters
----------
container : Container
The Container object that this equation belongs to.
name : str, optional
Name of the equation. If not provided, a unique name is generated automatically.
type : str, optional
Type of the equation. Options: "regular", "nonbinding", "external", "boolean".
Default is "regular".
domain : Sequence[Set | Alias | str] | Set | Alias | str, optional
The domain of the equation. Can be a list of Sets/Aliases, a single Set/Alias,
or strings representing set names. Use "*" for the universe set. Default is [] (scalar).
definition : Variable | Operation | Expression, optional
The mathematical definition of the equation. Can be set later via assignment.
records : Any, optional
Initial records to populate the equation.
domain_forwarding : bool | list[bool], optional
If True, adding records to this equation will implicitly add new elements to the
domain sets (if they are dynamic). Default is False.
description : str, optional
A human-readable description of the equation.
uels_on_axes : bool, optional
If True, implies that the Unique Element Labels (UELs) for the domain are
contained in the axes (index/columns) of the provided `records` object.
is_miro_output : bool, optional
If True, flags this equation as an output symbol for GAMS MIRO. Default is False.
Examples
--------
>>> import gamspy as gp
>>> m = gp.Container()
>>> i = gp.Set(m, "i", records=['i1','i2'])
>>> a = gp.Parameter(m, "a", [i], records=[['i1',1],['i2',2]])
>>> v = gp.Variable(m, "v", domain=[i])
>>> e = gp.Equation(m, "e", domain=[i])
>>> e[i] = a[i] <= v[i]
"""
@classmethod
def _constructor_bypass(
cls,
container: Container,
name: str,
type: str | EquationType = "regular",
domain: Sequence[Set | Alias | str] | Set | Alias | str | None = None,
records: Any | None = None,
description: str = "",
):
if domain is None:
domain = []
if isinstance(domain, (gp.Set, gp.Alias, str)):
domain = [domain]
if isinstance(domain, gp.math.Dim):
domain = gp.math._generate_dims(container, domain.dims)
# create new symbol object
obj = Equation.__new__(
cls,
container,
name,
type,
domain,
records=records,
description=description,
)
# set private properties directly
type = cast_type(type)
obj.type = EQU_TYPE[type]
obj._assignment = None
obj._gams_type = GMS_DT_EQU
obj._gams_subtype = TRANSFER_TO_GAMS_EQUATION_SUBTYPES[type]
obj._requires_state_check = False
obj._container = container
container._requires_state_check = True
obj._name = name
obj._domain = domain
obj._domain_forwarding = False
obj._description = description
obj._records = records
obj._modified = True
obj._domain_violations = None
# add to container
container.data.update({name: obj})
# gamspy attributes
obj._definition = None
obj.where = condition.Condition(obj)
obj._latex_name = name.replace("_", r"\_")
obj.container._add_statement(obj)
obj._synchronize = True
obj._metadata = {}
obj._winner = "python"
# create attributes
obj._l, obj._m, obj._lo, obj._up, obj._s = obj._init_attributes()
obj._stage = obj._create_attr("stage")
obj._range = obj._create_attr("range")
obj._slacklo = obj._create_attr("slacklo")
obj._slackup = obj._create_attr("slackup")
obj._slack = obj._create_attr("slack")
obj._infeas = obj._create_attr("infeas")
# miro support
obj._is_miro_output = False
return obj
def __new__(
cls,
container: Container | None = None,
name: str | None = None,
type: str | EquationType = "regular",
domain: Sequence[Set | Alias | str] | Set | Alias | str | None = None,
definition: Variable | Operation | Expression | None = None,
records: Any | None = None,
domain_forwarding: bool | list[bool] = False,
description: str = "",
uels_on_axes: bool = False,
is_miro_output: bool = False,
definition_domain: list | None = None,
):
if container is not None and not isinstance(container, gp.Container):
raise TypeError(f"Container must of type `Container` but found {container}")
if name is None:
return object.__new__(cls)
else:
if not isinstance(name, str):
raise TypeError(
f"Name must of type `str` but found {builtins.type(name)}"
)
try:
if not container:
container = gp._ctx_managers[
(os.getpid(), threading.get_native_id())
]
symbol = container[name]
if isinstance(symbol, cls):
return symbol
raise TypeError(
f"Cannot overwrite symbol `{name}` in container"
" because it is not an Equation object)"
)
except KeyError:
return object.__new__(cls)
def __init__(
self,
container: Container | None = None,
name: str | None = None,
type: str | EquationType = "regular",
domain: Sequence[Set | Alias] | Set | Alias | None = None,
definition: Variable | Operation | Expression | None = None,
records: Any | None = None,
domain_forwarding: bool | list[bool] = False,
description: str = "",
uels_on_axes: bool = False,
is_miro_output: bool = False,
definition_domain: list | None = None,
):
self._metadata: dict[str, Any] = {}
self._assignment: Expression | None = None
if is_miro_output and name is None:
raise ValidationError("Please specify a name for miro symbols.")
# miro support
self._is_miro_output = is_miro_output
self._domain_violations = None
self._synchronize = True
self._winner = "python"
# domain handling
if domain is None:
domain = []
if isinstance(domain, (gp.Set, gp.Alias)):
domain = [domain]
if isinstance(domain, gp.math.Dim):
domain = gp.math._generate_dims(container, domain.dims) # type: ignore
# does symbol exist
has_symbol = False
if isinstance(getattr(self, "container", None), gp.Container):
has_symbol = True
if has_symbol:
type = cast_type(type)
if self.type != type.casefold():
raise TypeError(
"Cannot overwrite symbol in container unless equation"
f" types are equal: `{self.type}` !="
f" `{type.casefold()}`"
)
if any(d1 != d2 for d1, d2 in itertools.zip_longest(self._domain, domain)):
raise ValueError(
"Cannot overwrite symbol in container unless symbol"
" domains are equal"
)
if self._domain_forwarding != domain_forwarding:
raise ValueError(
"Cannot overwrite symbol in container unless"
" 'domain_forwarding' is left unchanged"
)
# reset some properties
self._requires_state_check = True
self.container._requires_state_check = True
if description != "":
self.description = description
previous_state = self.container._options.miro_protect
self.container._options.miro_protect = False
self._records = None
self._modified = True
self._init_definition(definition)
# only set records if records are provided
if records is not None:
self.setRecords(records, uels_on_axes=uels_on_axes)
self.container._options.miro_protect = previous_state
else:
if container is None:
try:
container = gp._ctx_managers[
(os.getpid(), threading.get_native_id())
]
except KeyError as e:
raise ValidationError("Equation requires a container.") from e
assert container is not None
type = cast_type(type)
if name is not None:
name = validation.validate_name(name)
if is_miro_output:
name = name.lower() # type: ignore
else:
name = container._get_symbol_name(prefix="e")
previous_state = container._options.miro_protect
container._options.miro_protect = False
super().__init__(
container,
name,
type,
domain,
domain_forwarding=domain_forwarding,
description=description,
uels_on_axes=uels_on_axes,
)
self._latex_name = self.name.replace("_", r"\_")
if is_miro_output:
container._miro_output_symbols.append(self.name)
validation.validate_container(self, self._domain)
self.where = condition.Condition(self)
self.container._add_statement(self)
self._definition: Expression | None = None
self._definition_domain = definition_domain
self._init_definition(definition)
# create attributes
(
self._l,
self._m,
self._lo,
self._up,
self._s,
) = self._init_attributes()
self._stage = self._create_attr("stage")
self._range = self._create_attr("range")
self._slacklo = self._create_attr("slacklo")
self._slackup = self._create_attr("slackup")
self._slack = self._create_attr("slack")
self._infeas = self._create_attr("infeas")
if records is not None:
self.setRecords(records, uels_on_axes=uels_on_axes)
else:
if not self._is_miro_output:
self._modified = False
self.container._synch_with_gams()
container._options.miro_protect = previous_state
def _serialize(self) -> dict:
info = {
"_domain_forwarding": self._domain_forwarding,
"_is_miro_output": self._is_miro_output,
"_metadata": self._metadata,
"_synchronize": self._synchronize,
}
if self._assignment is not None:
info["_assignment"] = self._assignment.getDeclaration()
if self._definition is not None:
info["_definition"] = self._definition.getDeclaration()
return info
def _deserialize(self, info: dict) -> None:
for key, value in info.items():
if key == "_assignment":
left, right = value.split(" = ")
value = expression.Expression(left, "=", right[:-1])
elif key == "_definition":
left, right = value.split(" .. ")
value = expression.Expression(left, "..", right[:-1])
setattr(self, key, value)
# Relink domain symbols
new_domain = []
for elem in self._domain:
if elem == "*":
new_domain.append(elem)
continue
new_domain.append(self.container[elem])
self.domain = new_domain
def __getitem__(self, indices: EllipsisType | slice | tuple | str):
domain = validation.validate_domain(self, indices)
return implicits.ImplicitEquation(
self,
name=self.name,
type=self.type,
domain=domain, # type: ignore
)
def __setitem__(
self,
indices: EllipsisType | slice | tuple | str | implicits.ImplicitSet,
rhs: Expression,
):
# self[domain] = rhs
domain = validation.validate_domain(self, indices)
self._set_definition(domain, rhs)
self.container._synch_with_gams(gams_to_gamspy=True)
self._winner = "gams"
def __repr__(self) -> str:
return f"Equation(name='{self.name}', type='{self.type}', domain={self.domain})"
def _init_attributes(self) -> tuple:
level = self._create_attr("l")
marginal = self._create_attr("m")
lower = self._create_attr("lo")
upper = self._create_attr("up")
scale = self._create_attr("scale")
return level, marginal, lower, upper, scale
def _create_attr(self, attr_name):
return implicits.ImplicitParameter(
self,
name=f"{self.name}.{attr_name}",
records=self.records,
domain=self.domain,
)
def _update_attr_domains(self):
self._l.__init__(
self,
name=f"{self.name}.l",
records=self.records,
domain=self.domain,
)
self._m.__init__(
self,
name=f"{self.name}.m",
records=self.records,
domain=self.domain,
)
self._lo.__init__(
self,
name=f"{self.name}.lo",
records=self.records,
domain=self.domain,
)
self._up.__init__(
self,
name=f"{self.name}.up",
records=self.records,
domain=self.domain,
)
self._s.__init__(
self,
name=f"{self.name}.scale",
records=self.records,
domain=self.domain,
)
self._stage.__init__(
self,
name=f"{self.name}.stage",
records=self.records,
domain=self.domain,
)
self._range.__init__(
self,
name=f"{self.name}.range",
records=self.records,
domain=self.domain,
)
self._slackup.__init__(
self,
name=f"{self.name}.slackup",
records=self.records,
domain=self.domain,
)
self._slacklo.__init__(
self,
name=f"{self.name}.slacklo",
records=self.records,
domain=self.domain,
)
self._slack.__init__(
self,
name=f"{self.name}.slack",
records=self.records,
domain=self.domain,
)
self._infeas.__init__(
self,
name=f"{self.name}.infeas",
records=self.records,
domain=self.domain,
)
def _init_definition(
self,
assignment: Variable | Operation | Expression | None = None,
):
if assignment is None:
return None
domain = self.domain
if self._definition_domain is not None:
domain = validation.validate_domain(self, self._definition_domain)
self._set_definition(domain, assignment)
def _set_definition(self, domain, rhs):
# self[domain] = rhs
rhs_repr = rhs.gamsRepr()
if self.type == "nonbinding" and not any(
eq_type in rhs_repr for eq_type in EQ_TYPES
):
# x - c -> x - c == 0
rhs = rhs == gp.Number(0)
rhs_repr = rhs.gamsRepr()
if not any(eq_type in rhs_repr for eq_type in EQ_TYPES):
raise ValidationError(
"Equation definition must contain at least one equality sign such as ==, <= or >=."
)
if self.type in IRREGULAR_EQ_MAP and "=e=" in rhs_repr:
rhs.operator = IRREGULAR_EQ_MAP[self.type]
if self.type == "external" and "=e=" not in rhs.gamsRepr():
raise ValidationError("External equations must contain ==")
statement = expression.Expression(
implicits.ImplicitEquation(
self,
name=self.name,
type=self.type,
domain=domain,
),
"..",
rhs,
)
statement._validate_definition(utils._unpack(domain))
self.container._add_statement(statement)
self._definition = statement
def _check_ambiguity(self) -> None:
"""Ambiguity check for MCP, EMP, MPEC models. See #610"""
# Looks for =e=, =l= and =g= in an equation definition
# with a stack based inorder traversal algorithm (O(N)).
stack = []
assert self._definition is not None
node = self._definition.right
while True:
if node is not None:
stack.append(node)
node = getattr(node, "left", None) # type: ignore
elif stack:
node = stack.pop()
if (
isinstance(node, expression.Expression)
and node.operator in {"=e=", "=l=", "=g=", "=x=", "=n="}
and not isinstance(node.right, operable.Operable)
):
raise ValidationError(
f"Definition of `{self.name}` is ambigiuous. Please "
"use gp.Number for numeric values or disable ambiguity "
"check via gp.set_options({'ALLOW_AMBIGUOUS_EQUATIONS': 'no'}). "
"Using numeric values in equations without gp.Number can result in "
f"different order than expected. Print `{self.name}.getDefinition()` to "
"make sure that the equation definition is as expected."
)
node = getattr(node, "right", None)
else:
break # pragma: no cover
@property
def l(self):
"""
The Level of the equation (its current value).
This corresponds to the `.l` suffix in GAMS. After a solve, this represents the
value of the equation.
Returns
-------
ImplicitParameter
Examples
--------
>>> from gamspy import Container, Parameter, Variable, Equation, Model
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq[...] = 2*x1 + 3*x2 <= 10
>>> solved_model = Model(m, "my_model", equations=[eq], objective=10*x1 + 6*x2, sense="MAX").solve()
>>> repr = Parameter(m, "repr")
>>> repr[...] = eq.l
>>> repr.toValue()
np.float64(10.0)
"""
return self._l
@l.setter
def l(self, value: int | float | Expression):
self._l[...] = value
@property
def m(self):
"""
The Marginal (dual) value of the equation.
This corresponds to the `.m` suffix in GAMS. It represents the shadow price
or dual variable associated with the constraint.
Returns
-------
ImplicitParameter
Examples
--------
>>> from gamspy import Container, Parameter, Variable, Equation, Model
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq[...] = 2*x1 + 3*x2 <= 10
>>> solved_model = Model(m, "my_model", equations=[eq], objective=10*x1 + 6*x2, sense="MAX").solve()
>>> repr = Parameter(m, "repr")
>>> repr[...] = eq.m
>>> repr.toValue()
np.float64(5.0)
"""
return self._m
@m.setter
def m(self, value: int | float | Expression):
self._m[...] = value
@property
def lo(self):
"""
The Lower Bound of the equation.
Returns
-------
ImplicitParameter
Examples
--------
>>> from gamspy import Container, Parameter, Variable, Equation, Model
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq[...] = 2*x1 + 3*x2 <= 10
>>> solved_model = Model(m, "my_model", equations=[eq], objective=10*x1 + 6*x2, sense="MAX").solve()
>>> repr = Parameter(m, "repr")
>>> repr[...] = eq.lo
>>> repr.toValue()
np.float64(-inf)
"""
return self._lo
@lo.setter
def lo(self, value: int | float | Expression):
self._lo[...] = value
@property
def up(self):
"""
The Upper Bound of the equation.
Returns
-------
ImplicitParameter
Examples
--------
>>> from gamspy import Container, Parameter, Variable, Equation, Model
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq[...] = 2*x1 + 3*x2 <= 10
>>> solved_model = Model(m, "my_model", equations=[eq], objective=10*x1 + 6*x2, sense="MAX").solve()
>>> repr = Parameter(m, "repr")
>>> repr[...] = eq.up
>>> repr.toValue()
np.float64(10.0)
"""
return self._up
@up.setter
def up(self, value: int | float | Expression):
self._up[...] = value
@property
def scale(self):
"""
The Scale factor of the equation.
This corresponds to the `.scale` suffix in GAMS, used for scaling the equation
to improve numerical stability during solving.
Returns
-------
ImplicitParameter
Examples
--------
>>> from gamspy import Container, Parameter, Variable, Equation, Model
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq[...] = 2*x1 + 3*x2 <= 10
>>> solved_model = Model(m, "my_model", equations=[eq], objective=10*x1 + 6*x2, sense="MAX").solve()
>>> repr = Parameter(m, "repr")
>>> repr[...] = eq.scale
>>> repr.toValue()
np.float64(1.0)
"""
return self._s
@scale.setter
def scale(self, value: int | float | Expression):
self._s[...] = value
@property
def stage(self):
"""
The Stage of the equation.
This corresponds to the `.stage` suffix in GAMS, often used in stochastic programming
or model translation contexts.
Returns
-------
ImplicitParameter
Examples
--------
>>> from gamspy import Container, Parameter, Variable, Equation, Model
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq[...] = 2*x1 + 3*x2 <= 10
>>> solved_model = Model(m, "my_model", equations=[eq], objective=10*x1 + 6*x2, sense="MAX").solve()
>>> repr = Parameter(m, "repr")
>>> repr[...] = eq.stage
>>> repr.toValue()
np.float64(1.0)
"""
return self._stage
@stage.setter
def stage(self, value: int | float | Expression):
self._stage[...] = value
@property
def range(self):
"""
The Range of the equation.
This corresponds to the `.range` suffix in GAMS. It is used to define the sensitivity
range for range constraints.
Returns
-------
ImplicitParameter
Examples
--------
>>> from gamspy import Container, Parameter, Variable, Equation, Model
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq[...] = 2*x1 + 3*x2 <= 10
>>> solved_model = Model(m, "my_model", equations=[eq], objective=10*x1 + 6*x2, sense="MAX").solve()
>>> repr = Parameter(m, "repr")
>>> repr[...] = eq.range
>>> repr.toValue()
np.float64(inf)
"""
return self._range
@property
def slacklo(self):
"""
The lower bound slack of the equation.
Returns
-------
ImplicitParameter
Examples
--------
>>> from gamspy import Container, Parameter, Variable, Equation, Model
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq[...] = 2*x1 + 3*x2 <= 10
>>> solved_model = Model(m, "my_model", equations=[eq], objective=10*x1 + 6*x2, sense="MAX").solve()
>>> repr = Parameter(m, "repr")
>>> repr[...] = eq.slacklo
>>> repr.toValue()
np.float64(inf)
"""
return self._slacklo
@property
def slackup(self):
"""
The upper bound slack of the equation.
Returns
-------
ImplicitParameter
Examples
--------
>>> from gamspy import Container, Parameter, Variable, Equation, Model
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq[...] = 2*x1 + 3*x2 <= 10
>>> solved_model = Model(m, "my_model", equations=[eq], objective=10*x1 + 6*x2, sense="MAX").solve()
>>> repr = Parameter(m, "repr")
>>> repr[...] = eq.slackup
>>> repr.toValue()
np.float64(0.0)
"""
return self._slackup
@property
def slack(self):
"""
The Slack of the equation.
This corresponds to the `.slack` suffix. It represents the distance from the
equation's bound (e.g., RHS - LHS for <= equations).
Returns
-------
ImplicitParameter
Examples
--------
>>> from gamspy import Container, Parameter, Variable, Equation, Model
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq[...] = 2*x1 + 3*x2 <= 10
>>> solved_model = Model(m, "my_model", equations=[eq], objective=10*x1 + 6*x2, sense="MAX").solve()
>>> repr = Parameter(m, "repr")
>>> repr[...] = eq.slack
>>> repr.toValue()
np.float64(0.0)
"""
return self._slack
@property
def infeas(self):
"""
The Infeasibility of the equation.
Returns the amount by which the equation violates its bounds.
Returns
-------
ImplicitParameter
Examples
--------
>>> from gamspy import Container, Parameter, Variable, Equation, Model
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq[...] = 2*x1 + 3*x2 <= 10
>>> solved_model = Model(m, "my_model", equations=[eq], objective=10*x1 + 6*x2, sense="MAX").solve()
>>> repr = Parameter(m, "repr")
>>> repr[...] = eq.infeas
>>> repr.toValue()
np.float64(0.0)
"""
return self._infeas
[docs]
def computeInfeasibilities(self) -> pd.DataFrame:
"""
Computes infeasibilities of the equation.
Checks if the level value `.l` violates the bounds `.lo` and `.up`
and returns a DataFrame containing the violations.
Returns
-------
pd.DataFrame
Examples
--------
>>> import gamspy as gp
>>> m = gp.Container()
>>> e = gp.Equation(m, "e")
>>> e.l[...] = -10
>>> e.lo[...] = 5
>>> e.computeInfeasibilities().values.tolist()
[[-10.0, 0.0, 5.0, inf, 1.0, 15.0]]
"""
return utils._calculate_infeasibilities(self)
[docs]
def getEquationListing(
self,
n: int | None = None,
filters: list[list[str]] | None = None,
infeasibility_threshold: float | None = None,
) -> str:
"""
Returns the equation listing (log output) from the last solve.
This requires the model to have been solved with the `equation_listing_limit` option enabled.
Parameters
----------
n : int, optional
Maximum number of equations to return.
filters : list[list[str]], optional
Filters to select specific elements for the listing.
The list size must match the equation's dimension.
infeasibility_threshold: float, optional
If set, only returns equations with infeasibility values greater than this threshold.
Returns
-------
str
The text listing of the generated equations.
Raises
------
ValidationError
If the model was not solved with `equation_listing_limit`.
ValidationError
In case the length of the filters is different than the dimension of the equation.
Examples
--------
>>> import gamspy as gp
>>> m = gp.Container()
>>> i = gp.Set(m, records=["item1", "item2"])
>>> v = gp.Variable(m, domain=i)
>>> z = gp.Variable(m)
>>> e = gp.Equation(m, domain=i)
>>> e[i] = v[i] * z >= 5
>>> model = gp.Model(m, "test", equations=[e], problem="NLP", sense="MIN", objective=z)
>>> summary = model.solve(options=gp.Options(equation_listing_limit=10))
>>> print(e.getEquationListing())
e(item1).. (0)*v(item1) + (0)*z =G= 5 ; (LHS = 0, INFES = 5 ****)
e(item2).. (0)*v(item2) + (0)*z =G= 5 ; (LHS = 0, INFES = 5 ****)
"""
if not hasattr(self, "_equation_listing"):
raise ValidationError(
"The model must be solved with `equation_listing_limit` option for this functionality to work."
)
listings = self._equation_listing if filters is None else []
if filters is not None:
for listing in self._equation_listing:
lhs, _ = listing.split("..")
# symbol(elem1, elem2)
_, domain = lhs[:-1].split("(")
sets = domain.split(",") # ["elem1", "elem2"]
if len(filters) != len(sets):
raise ValidationError(
f"Filter size {len(filters)} must be equal to the domain size {len(sets)}"
)
matches = 0
for user_filter, set in zip(filters, sets, strict=False):
if set in user_filter or user_filter == []:
matches += 1
# infeasibility = float(listing.split("INFES = ")[-1][:-6])
if matches == len(sets):
listings.append(listing)
def infeasibility_filter(listing):
infeasibility = listing.split("INFES = ")
if infeasibility_threshold is None:
return True
return (
len(infeasibility) == 2
and float(infeasibility[-1][:-6]) < infeasibility_threshold
)
return "\n".join(list(filter(infeasibility_filter, listings))[:n])
@property
def records(self):
"""
Returns the records (data) of the Equation as a DataFrame.
Returns
-------
DataFrame
Examples
--------
>>> from gamspy import Container, Parameter, Variable, Equation, Model
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq[...] = 2*x1 + 3*x2 <= 10
>>> solved_model = Model(m, "my_model", equations=[eq], objective=10*x1 + 6*x2, sense="MAX").solve()
>>> eq.toValue()
np.float64(10.0)
"""
return self._records
@records.setter
def records(self, records):
if records is not None and not isinstance(records, pd.DataFrame):
raise TypeError("Symbol 'records' must be type DataFrame")
# set records
self._records = records
self._requires_state_check = True
self._modified = True
self.container._requires_state_check = True
self.container.modified = True
if self._records is not None and self._domain_forwarding:
self._domainForwarding()
# reset state check flags for all symbols in the container
for symbol in self.container.data.values():
symbol._requires_state_check = True
def __hash__(self):
return id(self)
def _setRecords(self, records: Any, *, uels_on_axes: bool = False) -> None:
super().setRecords(records, uels_on_axes)
if gp.get_option("DROP_DOMAIN_VIOLATIONS"):
if self.hasDomainViolations():
self._domain_violations = self.getDomainViolations()
self.dropDomainViolations()
else:
self._domain_violations = None
[docs]
def setRecords(self, records: Any, uels_on_axes: bool = False) -> None:
"""
Sets the records (data) of the Equation.
This allows manually setting the level, marginal, bounds, and scale for an equation.
Parameters
----------
records : Any
The data to load (e.g., list, numpy array, DataFrame).
uels_on_axes : bool, optional
If True, assumes domain elements are in the axes of the DataFrame. Default is False.
Examples
--------
>>> from gamspy import Container, Variable, Equation
>>> m = Container()
>>> x1 = Variable(m, "x1", type="Positive")
>>> x2 = Variable(m, "x2", type="Positive")
>>> z = Variable(m, "z")
>>> eq = Equation(m, "eq")
>>> eq.setRecords(5)
>>> eq.toValue()
np.float64(5.0)
"""
self._setRecords(records, uels_on_axes=uels_on_axes)
self.container._synch_with_gams()
self._winner = "python"
@property
def type(self):
"""
The type of the equation.
Common types include:
- 'regular' (or 'eq', 'geq', 'leq'): Standard =e=, =g=, =l= constraints.
- 'nonbinding' ('=n='): No relationship implied.
- 'external' ('=x='): External equation.
- 'boolean' ('=b='): Boolean equation.
Returns
-------
str
The type of equation
Examples
--------
>>> import gamspy as gp
>>> m = gp.Container()
>>> e = gp.Equation(m, "e", type="regular")
>>> e.type
'eq'
"""
return self._type
@type.setter
def type(self, eq_type: str | EquationType):
given_type = cast_type(eq_type)
gt.Equation.type.fset(self, given_type)
[docs]
def gamsRepr(self) -> str:
"""
Returns the string representation of this Equation in the GAMS language.
(e.g., 'e(i)').
Returns
-------
str
Examples
--------
>>> import gamspy as gp
>>> m = gp.Container()
>>> i = gp.Set(m, "i", records=['i1','i2'])
>>> e = gp.Equation(m, "e", domain=[i])
>>> e.gamsRepr()
'e(i)'
"""
representation = self.name
if self.domain:
representation += self._get_domain_str(self._domain_forwarding)
return representation
[docs]
def latexRepr(self) -> str:
r"""
Generates a LaTeX representation of the equation definition.
Returns
-------
str
LaTeX string defining the equation.
Raises
------
ValidationError
If the equation has not been defined (assigned).
Examples
--------
>>> import gamspy as gp
>>> m = gp.Container()
>>> i = gp.Set(m, "i", records=['i1'])
>>> v = gp.Variable(m, "v", domain=[i])
>>> e = gp.Equation(m, "e", domain=[i])
>>> e[i] = v[i] <= 10
>>> print(e.latexRepr())
$
v_{i} \leq 10\hfill \forall i
$
"""
if self._definition is None:
raise ValidationError(
"Equation must be defined to get its latex representation."
)
# The LHS of an equation definition can either be an ImplicitEquation or a condition.
# e.g. e[i] = ... or e[i].where[b[i]] = ...
assert isinstance(
self._definition.left,
(implicits.ImplicitEquation, condition.Condition),
)
right_side = ""
if isinstance(self._definition.left, implicits.ImplicitEquation):
if len(self._definition.left.domain) > 0:
domain_str = ",".join(
[symbol.latexRepr() for symbol in self._definition.left.domain]
)
right_side = f"\\hfill \\forall {domain_str}"
else:
domain_str = ",".join(
[
symbol.latexRepr()
for symbol in self._definition.left.conditioning_on.domain # type: ignore
]
)
domain_str = f"\\forall {domain_str}"
assert self._definition.left.condition is not None
if hasattr(self._definition.left.condition, "latexRepr"):
constraint_str = self._definition.left.condition.latexRepr()
else:
assert isinstance(self._definition.left.condition, (int, float))
constraint_str = str(self._definition.left.condition)
right_side = f"\\hfill {domain_str} ~ | ~ {constraint_str}"
assert self._definition.right is not None
definition_str = self._definition.right.latexRepr() # type: ignore
if definition_str[0] == "(":
definition_str = definition_str[1:-1]
equation_str = "$\n" + definition_str + f"{right_side}" + "\n$"
return equation_str
[docs]
def getDeclaration(self) -> str:
"""
Returns the GAMS declaration statement for this Equation.
(e.g., 'Equation e(i);').
Returns
-------
str
Examples
--------
>>> import gamspy as gp
>>> m = gp.Container()
>>> i = gp.Set(m, "i", records=['i1','i2'])
>>> a = gp.Parameter(m, "a", [i], records=[['i1',1],['i2',2]])
>>> v = gp.Variable(m, "v", domain=[i])
>>> e = gp.Equation(m, "e", domain=[i])
>>> e.getDeclaration()
'Equation e(i) / /;'
"""
output = f"Equation {self.name}"
if self.domain:
output += self._get_domain_str(self._domain_forwarding)
if self.description:
output += ' "' + self.description + '"'
if self.records is None:
output += " / /"
output += ";"
return output
def getAssignment(self) -> str:
"""
Returns the latest GAMS assignment statement for this Equation (e.g. assigning .l).
Returns
-------
str
Raises
------
ValidationError
If the equation attributes have not been assigned a value.
Examples
--------
>>> import gamspy as gp
>>> m = gp.Container()
>>> i = gp.Set(m, "i", records=['i1','i2'])
>>> e = gp.Variable(m, "e", domain=[i])
>>> e.l[i] = 0;
>>> e.getAssignment()
'e.l(i) = 0;'
"""
if self._assignment is None:
raise ValidationError("Equation was not assigned!")
return self._assignment.getDeclaration()
[docs]
def getDefinition(self) -> str:
"""
Returns the GAMS definition statement (algebra) for this Equation.
(e.g., 'e(i) .. x(i) =l= 5;').
Returns
-------
str
Raises
------
ValidationError
If the equation algebra has not been defined.
Examples
--------
>>> import gamspy as gp
>>> m = gp.Container()
>>> i = gp.Set(m, "i", records=['i1','i2'])
>>> a = gp.Parameter(m, "a", [i], records=[['i1',1],['i2',2]])
>>> v = gp.Variable(m, "v", domain=[i])
>>> e = gp.Equation(m, "e", domain=[i])
>>> e[i] = a[i] <= v[i]
>>> e.getDefinition()
'e(i) .. a(i) =l= v(i);'
"""
if self._definition is None:
raise ValidationError("Equation is not defined!")
return self._definition.getDeclaration()
def cast_type(type: str | EquationType) -> str:
if isinstance(type, str):
type = type.lower()
if type not in (
"eq",
"geq",
"leq",
"regular",
"nonbinding",
"external",
"boolean",
):
raise ValueError(
f"Allowed equation types: {EquationType.values()} but found {type}."
)
# assign eq by default
if type == "regular":
type = "eq"
elif isinstance(type, EquationType):
# assign eq by default
type = "eq" if type == EquationType.REGULAR else str(type)
return type
