Reference: Set-Like Objects

Constraint

class namedisl.Constraint(_obj: IslObjectT_co, space: Space, _isl_names_ok: bool = False)

static equality_from_aff(aff: Aff) → Constraint

static inequality_from_aff(aff: Aff) → Constraint

is_equality

as_aff() → Aff

as_basic_set()

as_basic_map()

namedisl.make_constraint(obj: Constraint) → Constraint

Quasiconvex set

class namedisl.BasicSet(_obj: IslObjectT_co, space: Space, _isl_names_ok: bool = False)

add_constraint(cns: Constraint, /) → BasicSet

var_affs

Returns a lazily-evaluated mapping from dimension names (or zero) to PwAffs.

Note

Lazy evaluation means you do not pay for the creation of unused dimensions.

as_set() → Set

get_constraints()

is_bounded() → bool

is_empty() → bool

plain_is_empty() → bool

plain_is_universe() → bool

universe_like_me() → Self

empty_like_me() → Self

fix_dim(name: str, value: Val | int) → Self

eliminate(names: Collection[str], *, cache: Cache | None = None) → Self

Keeps the dimensions, but eliminates constraints.

project_out(names: str | Collection[str], *, cache: Cache | None = None) → Self

Eliminates the dimensions and constraints.

project_out_except(names_to_keep: Collection[str], *, cache: Cache | None = None) → Self

Eliminates the dimensions and constraints.

gist(context: Self) → Self

remove_divs() → Self

__and__(other: Self) → Self

__or__(other: Self) → Self

Return self|value.

__sub__(other: Self) → Self

__eq__(other: object) → bool

This is intended to be cheap and strict, suitable mainly for hashing. Some subclasses (e.g. namedisl.Set) may provide a different, ‘mathematically exact’ notion of equality as a different method that is more expensive to check for.

equals(other: Self) → bool

__lt__(other: Self) → bool

Return self<value.

__le__(other: Self) → bool

Return self<=value.

namedisl.make_basic_set(src: str, ctx: Context | None = None) → BasicSet

namedisl.make_basic_set(src: BasicSet) → BasicSet

General set

class namedisl.Set(_obj: IslObjectT_co, space: Space, _isl_names_ok: bool = False)

complement() → Set

convex_hull() → BasicSet

get_basic_sets() → list[BasicSet]

dim_max(name: str, *, cache: Cache | None = None) → PwAff

dim_min(name: str, *, cache: Cache | None = None) → PwAff

var_pw_affs

Returns a lazily-evaluated mapping from dimension names (or zero) to PwAffs.

Note

Lazy evaluation means you do not pay for the creation of unused dimensions.

as_map(in_names: Collection[str]) → Map

as_basic() → BasicSet

Will only succeed if self consists of a single piece. Try coalesce() if not.

equate_dims(names: Sequence[tuple[str, str]]) → Self

as_pw_multi_aff() → PwMultiAff

remove_redundancies()

coalesce() → Self

is_bounded() → bool

is_empty() → bool

plain_is_empty() → bool

plain_is_universe() → bool

universe_like_me() → Self

empty_like_me() → Self

fix_dim(name: str, value: Val | int) → Self

eliminate(names: Collection[str], *, cache: Cache | None = None) → Self

Keeps the dimensions, but eliminates constraints.

project_out(names: str | Collection[str], *, cache: Cache | None = None) → Self

Eliminates the dimensions and constraints.

project_out_except(names_to_keep: Collection[str], *, cache: Cache | None = None) → Self

Eliminates the dimensions and constraints.

gist(context: Self) → Self

remove_divs() → Self

__and__(other: Self) → Self

__or__(other: Self) → Self

Return self|value.

__sub__(other: Self) → Self

__eq__(other: object) → bool

This is intended to be cheap and strict, suitable mainly for hashing. Some subclasses (e.g. namedisl.Set) may provide a different, ‘mathematically exact’ notion of equality as a different method that is more expensive to check for.

equals(other: Self) → bool

__lt__(other: Self) → bool

Return self<value.

__le__(other: Self) → bool

Return self<=value.

namedisl.make_set(src: str, ctx: Context | None = None) → Set

namedisl.make_set(src: Set) → Set

Quasiconvex map

class namedisl.BasicMap(_obj: IslObjectT_co, space: Space, _isl_names_ok: bool = False)

domain() → BasicSet

range() → BasicSet

intersect_domain(domain: BasicSet) → Self

intersect_range(range: BasicSet) → Self

get_constraints()

active_dim_types: ClassVar[frozenset[DimType]] = frozenset({DimType.param, DimType.in_, DimType.out})

reverse() → Self

is_empty() → bool

plain_is_empty() → bool

plain_is_universe() → bool

universe_like_me() → Self

empty_like_me() → Self

fix_dim(name: str, value: Val | int) → Self

eliminate(names: Collection[str], *, cache: Cache | None = None) → Self

Keeps the dimensions, but eliminates constraints.

project_out(names: str | Collection[str], *, cache: Cache | None = None) → Self

Eliminates the dimensions and constraints.

project_out_except(names_to_keep: Collection[str], *, cache: Cache | None = None) → Self

Eliminates the dimensions and constraints.

gist(context: Self) → Self

remove_divs() → Self

__and__(other: Self) → Self

__or__(other: Self) → Self

Return self|value.

__sub__(other: Self) → Self

__eq__(other: object) → bool

This is intended to be cheap and strict, suitable mainly for hashing. Some subclasses (e.g. namedisl.Set) may provide a different, ‘mathematically exact’ notion of equality as a different method that is more expensive to check for.

equals(other: Self) → bool

__lt__(other: Self) → bool

Return self<value.

__le__(other: Self) → bool

Return self<=value.

namedisl.make_basic_map(src: str, ctx: Context | None = None) → BasicMap

namedisl.make_basic_map(src: BasicMap) → BasicMap

General map

class namedisl.Map(_obj: IslObjectT_co, space: Space, _isl_names_ok: bool = False)

complement() → Map

convex_hull() → BasicMap

get_basic_maps() → list[BasicMap]

domain() → Set

range() → Set

intersect_domain(domain: Set) → Self

intersect_range(range: Set) → Self

apply_range(other: Self) → Self

apply_domain(other: Self) → Self

as_basic() → BasicMap

Will only succeed if self consists of a single piece. Try coalesce() if not.

as_set() → Set

domain_var_pw_affs

Returns a lazily-evaluated mapping from dimension names (or zero) to PwAffs for the domain variables of self

Note

Lazy evaluation means you do not pay for the creation of unused dimensions.

range_var_pw_affs

Returns a lazily-evaluated mapping from dimension names (or zero) to PwAffs for the domain variables of self

Note

Lazy evaluation means you do not pay for the creation of unused dimensions.

is_empty() → bool

plain_is_empty() → bool

plain_is_universe() → bool

universe_like_me() → Self

empty_like_me() → Self

fix_dim(name: str, value: Val | int) → Self

eliminate(names: Collection[str], *, cache: Cache | None = None) → Self

Keeps the dimensions, but eliminates constraints.

project_out(names: str | Collection[str], *, cache: Cache | None = None) → Self

Eliminates the dimensions and constraints.

project_out_except(names_to_keep: Collection[str], *, cache: Cache | None = None) → Self

Eliminates the dimensions and constraints.

gist(context: Self) → Self

remove_divs() → Self

__and__(other: Self) → Self

__or__(other: Self) → Self

Return self|value.

__sub__(other: Self) → Self

__eq__(other: object) → bool

This is intended to be cheap and strict, suitable mainly for hashing. Some subclasses (e.g. namedisl.Set) may provide a different, ‘mathematically exact’ notion of equality as a different method that is more expensive to check for.

equals(other: Self) → bool

__lt__(other: Self) → bool

Return self<value.

__le__(other: Self) → bool

Return self<=value.

equate_dims(names: Sequence[tuple[str, str]]) → Self

as_pw_multi_aff() → PwMultiAff

remove_redundancies()

coalesce() → Self

active_dim_types: ClassVar[frozenset[DimType]] = frozenset({DimType.param, DimType.in_, DimType.out})

reverse() → Self

namedisl.make_map(src: str, ctx: Context | None = None) → Map

namedisl.make_map(src: Map) → Map