Essential classes#

Summary#

Here are data structures essential for Puan toolkit.

Data types#

variable: The variable class is a central key in all Puan packages. It consists of an id, Bounds and data type (dtype).

Proposition#

class Proposition#

Bases: object

assume(fixed: Dict[str, int]) → Proposition#

Assumes something about variable’s bounds and returns a new proposition with these new bounds set. Other variables, not declared in new_variable_bounds, may also get new bounds as a consequence from the ones set in new_variable_bounds.

Parameters

new_variable_boundstyping.Dict[str, Union[int, Tuple[int, int], puan.Bounds]]: A dict of ids and either int, tuple or puan.Bounds as bounds for the variable.

Returns

outProposition

static from_json(data, class_map) → Any#: Convert from JSON into class object. Class map is a list of other classes maybe related to this class

to_json() → Dict[str, Any]#: JSON representation of class

to_short() → Tuple[str, int, object, int, Tuple[int, int]]#: Short data type has (id, sign, propositions, value, bounds)

variable#

class variable(id: str, bounds: Optional[Union[int, Tuple[int, int], Bounds]] = None, dtype: Optional[Dtype] = None)#

Bases: Proposition

The variable class is a central key in all Puan packages. It consists of an id, Bounds and data type (dtype).

Raises

ValueError: If dtype is bool and bounds are not (0, 1)

Methods

`to_json`()	Converts into JSON data type.
`to_short`()	Converts into short data type.
`support_vector_variable`()	Default support vector variable settings.
`from_strings`(*variables[, default_bounds])	Returns a list of `variable` from a list of strings (ids)
`from_mixed`(*variables[, default_bounds])	Returns a list of `variable` from a list of mixed data type.
`from_json`(data[, class_map])	Creates `variable` from json format.

to_dict

assume(fixed: Dict[str, Union[int, Tuple[int, int], Bounds]]) → Proposition#

Parameters

new_variable_boundstyping.Dict[str, Union[int, Tuple[int, int], puan.Bounds]]: A dict of ids and either int, tuple or puan.Bounds as bounds for the variable.

Returns

outProposition

bounds: puan.Bounds#

evaluate(interpretation: Dict[str, Union[int, Tuple[int, int], Bounds]]) → Bounds#

Evaluates interpretation on this proposition.

Parameters

interpretationDict[str, Union[Bounds, Tuple[int,int], int]]: the values of the variables in the model to evaluate it for

Returns

outOptional[int]

Examples

>>> variable("a", bounds=(0,1)).evaluate({"x": 1})
Bounds(lower=0, upper=1)

>>> variable("a", bounds=(0,2)).evaluate({"a": 2})
Bounds(lower=2, upper=2)

>>> variable("a", bounds=(0,2)).evaluate({"a": (2,3)})
Bounds(lower=2, upper=3)

>>> variable("a", bounds=(0,2)).evaluate({"a": Bounds(1,1)})
Bounds(lower=1, upper=1)

>>> variable("a", bounds=(-1,0)).evaluate({"a": 1})
Bounds(lower=1, upper=1)

>>> variable("a", bounds=(1,1)).evaluate({"a": 0})
Bounds(lower=0, upper=0)

>>> variable("a", bounds=(1,1)).evaluate({"x": 1})
Bounds(lower=1, upper=1)

evaluate_propositions(interpretation: ~typing.Dict[str, ~typing.Union[int, ~typing.Tuple[int, int], ~puan.Bounds]], out: ~typing.Callable[[~puan.Bounds], ~typing.Union[~puan.Bounds, ~typing.Tuple[int, int], int]] = <function variable.<lambda>>) → Dict[str, Bounds]#

Evaluates interpretation on this proposition.

Parameters

interpretationDict[Union[str, variable], int]: the values of the variables in the model to evaluate it for
outCallback[[Bounds], Union[Bounds, Tuple[int,int], int]]: an optional callback function for changing output data type.

Returns

outOptional[int]

Examples

>>> variable("a", bounds=(0,1)).evaluate_propositions({"x": 1})
{'a': Bounds(lower=0, upper=1)}

>>> variable("a", bounds=(0,1)).evaluate_propositions({"a": (1,1)})
{'a': Bounds(lower=1, upper=1)}

>>> variable("a", bounds=(0,1)).evaluate_propositions({"a": Bounds(1,1)})
{'a': Bounds(lower=1, upper=1)}

>>> variable("a", bounds=(0,2)).evaluate_propositions({"a": 2})
{'a': Bounds(lower=2, upper=2)}

>>> variable("a", bounds=(-1,0)).evaluate_propositions({"a": 1})
{'a': Bounds(lower=1, upper=1)}

>>> variable("a", bounds=(1,1)).evaluate_propositions({"a": 0})
{'a': Bounds(lower=0, upper=0)}

>>> variable("a", bounds=(1,1)).evaluate_propositions({"x": 1})
{'a': Bounds(lower=1, upper=1)}

flatten() → List[Proposition]#

Returns all its propositions and their sub propositions as a unique list of propositions.

Returns

outList[Proposition]

Examples

>>> variable("a", bounds=(0,2)).flatten()
[variable(id='a', bounds=Bounds(lower=0, upper=2))]

static from_json(data: Dict[str, Any], class_map: List[Type] = []) → variable#

Creates variable from json format.

Notes

Bounds are defaulted to Bounds(lower=0, upper=1) if not provided. class_map is ignored.

Examples

>>> variable.from_json({'id': 'x'})
variable(id='x', bounds=Bounds(lower=0, upper=1))

>>> variable.from_json({'id': 'x', 'bounds': {'lower': -2, 'upper': 3}})
variable(id='x', bounds=Bounds(lower=-2, upper=3))

static from_mixed(*variables: List[Union[str, int, tuple, list, variable]], default_bounds: Tuple[int, int] = (0, 1)) → List[variable]#

Returns a list of variable from a list of mixed data type.

Parameters

variablesA list of either str, int, tuple, list or variable
default_boundsTuple of ints (lower bound, upper bound)

Returns

outList[variable]

Notes

Every item x in *variables that is not an instance of variable will be converted to a variable with id set to x and default bounds.
List of variables are returned sorted.

Examples

>>> variable.from_mixed("a",4,("b","c"),variable("x",(1,2)))
[variable(id="('b', 'c')", bounds=Bounds(lower=0, upper=1)), variable(id='4', bounds=Bounds(lower=0, upper=1)), variable(id='a', bounds=Bounds(lower=0, upper=1)), variable(id='x', bounds=Bounds(lower=1, upper=2))]

>>> variable.from_mixed("a",4,variable("x",(2,4)), default_bounds=(-1, 1))
[variable(id='4', bounds=Bounds(lower=-1, upper=1)), variable(id='a', bounds=Bounds(lower=-1, upper=1)), variable(id='x', bounds=Bounds(lower=2, upper=4))]

static from_strings(*variables: List[str], default_bounds: Tuple[int, int] = (0, 1)) → List[variable]#

Returns a list of variable from a list of strings (ids)

Parameters

variablesA list of strings (ids)
default_boundsTuple of ints (lower bound, upper bound)

Returns

outList[variable]

Notes

List of variables are returned sorted

Examples

>>> variable.from_strings("a","b")
[variable(id='a', bounds=Bounds(lower=0, upper=1)), variable(id='b', bounds=Bounds(lower=0, upper=1))]

>>> variable.from_strings("a","b", default_bounds=(-1,10))
[variable(id='a', bounds=Bounds(lower=-1, upper=10)), variable(id='b', bounds=Bounds(lower=-1, upper=10))]

id: str#

static support_vector_variable() → variable#

Default support vector variable settings.

Returns

outvariable

Examples

>>> variable.support_vector_variable()
variable(id=0, bounds=Bounds(lower=1, upper=1))

to_json() → Dict[str, Any]#

Converts into JSON data type.

Examples

>>> variable(id="x", bounds=Bounds(lower=0, upper=10)).to_json()
{'id': 'x', 'bounds': {'lower': 0, 'upper': 10}}

to_short() → Tuple[str, int, List, int, Tuple[int, int]]#

Converts into short data type.

Examples

>>> variable(id="x", bounds=Bounds(lower=0, upper=10)).to_short()
('x', 1, [], 0, (0, 10))

Bounds#

class Bounds(lower: int, upper: int)#

Bases: object

The Bounds class defines lower and upper bounds of a variable

Raises

ValueError: If lower is greater than upper

Methods

as_tuple()

Bounds as tuple

as_tuple() → Tuple[int, int]#

Bounds as tuple

Returns

outTuple[int, int]

property constant: Optional[int]#

If lower and upper bounds are the same, that value is returned. Else None is returned.

Returns

outOptional[int]

Examples

>>> Bounds(0, 1).constant

>>> Bounds(-2, -2).constant
-2

Sign#

class Sign(value)#

Bases: IntEnum

Sign, either positive or negative

NEGATIVE = -1#

POSITIVE = 1#

API reference

Numpy Array objects