# Spaces

## Creation of spaces

`quadratic_space`

— Method`quadratic_space(K::NumField, n::Int; cached::Bool = true) -> QuadSpace`

Create the quadratic space over `K`

with dimension `n`

and Gram matrix equals to the identity matrix.

`hermitian_space`

— Method`hermitian_space(E::NumField, n::Int; cached::Bool = true) -> HermSpace`

Create the hermitian space over `E`

with dimension `n`

and Gram matrix equals to the identity matrix. The number field `E`

must be a quadratic extension, that is, $degree(E) == 2$ must hold.

`quadratic_space`

— Method`quadratic_space(K::NumField, G::MatElem; cached::Bool = true) -> QuadSpace`

Create the quadratic space over `K`

with Gram matrix `G`

. The matrix `G`

must be square and symmetric.

`hermitian_space`

— Method`hermitian_space(E::NumField, gram::MatElem; cached::Bool = true) -> HermSpace`

Create the hermitian space over `E`

with Gram matrix equals to `gram`

. The matrix `gram`

must be square and hermitian with respect to the non-trivial automorphism of `E`

. The number field `E`

must be a quadratic extension, that is, $degree(E) == 2$ must hold.

### Examples

Here are easy examples to see how these constructors work. We will keep the two following spaces for the rest of this section:

`julia> K, a = CyclotomicRealSubfield(7);`

`julia> Kt, t = K["t"];`

`julia> E, b = number_field(t^2-a*t+1, "b");`

`julia> Q = quadratic_space(K, K[0 1; 1 0])`

`Quadratic space of dimension 2 over maximal real subfield of cyclotomic field of order 7 with gram matrix [0 1] [1 0]`

`julia> H = hermitian_space(E, 3)`

`Hermitian space of dimension 3 over relative number field with defining polynomial t^2 - (z_7 + 1/z_7)*t + 1 over number field with defining polynomial $^3 + $^2 - 2*$ - 1 over rational field with gram matrix [1 0 0] [0 1 0] [0 0 1]`

## Attributes

Let $(V, \Phi)$ be a space over $E/K$. We define its *dimension* to be its dimension as a vector space over its base ring $E$ and its *rank* to be the rank of its Gram matrix. If these two invariants agree, the space is said to be *regular*.

While dealing with lattices, one always works with regular ambient spaces.

The *determinant* $\text{det}(V, \Phi)$ of $(V, \Phi)$ is defined to be the class of the determinant of its Gram matrix in $K^{\times}/N(E^{\times})$ (which is similar to $K^{\times}/(K^{\times})^2$ in the quadratic case). The *discriminant* $\text{disc}(V, \Phi)$ of $(V, \Phi)$ is defined to be $(-1)^{(m(m-1)/2)}\text{det}(V, \Phi)$, where $m$ is the rank of $(V, \Phi)$.

`rank`

— Method`rank(V::AbstractSpace) -> Int`

Return the rank of the space `V`

.

`dim`

— Method`dim(V::AbstractSpace) -> Int`

Return the dimension of the space `V`

.

`gram_matrix`

— Method`gram_matrix(V::AbstractSpace) -> MatElem`

Return the Gram matrix of the space `V`

.

`involution`

— Method`involution(V::AbstractSpace) -> NumFieldMor`

Return the involution of the space `V`

.

`base_ring`

— Method`base_ring(V::AbstractSpace) -> NumField`

Return the algebra over which the space `V`

is defined.

`fixed_field`

— Method`fixed_field(V::AbstractSpace) -> NumField`

Return the fixed field of the space `V`

.

`det`

— Method`det(V::AbstractSpace) -> FieldElem`

Return the determinant of the space `V`

as an element of its fixed field.

`discriminant`

— Method`discriminant(V::AbstractSpace) -> FieldElem`

Return the discriminant of the space `V`

as an element of its fixed field.

### Examples

So for instance, one could get the following information about the hermitian space $H$:

`julia> K, a = CyclotomicRealSubfield(7);`

`julia> Kt, t = K["t"];`

`julia> E, b = number_field(t^2-a*t+1, "b");`

`julia> H = hermitian_space(E, 3);`

`julia> rank(H), dim(H)`

`(3, 3)`

`julia> gram_matrix(H)`

`[1 0 0] [0 1 0] [0 0 1]`

`julia> involution(H)`

`Map with following data Domain: ======= Relative number field of degree 2 over maximal real subfield of cyclotomic field of order 7 Codomain: ========= Relative number field of degree 2 over maximal real subfield of cyclotomic field of order 7`

`julia> base_ring(H)`

`Relative number field with defining polynomial t^2 - (z_7 + 1/z_7)*t + 1 over number field with defining polynomial $^3 + $^2 - 2*$ - 1 over rational field`

`julia> fixed_field(H)`

`Number field with defining polynomial $^3 + $^2 - 2*$ - 1 over rational field`

`julia> det(H), discriminant(H)`

`(1, -1)`

## Predicates

Let $(V, \Phi)$ be a hermitian space over $E/K$ (resp. quadratic space $K$). We say that $(V, \Phi)$ is *definite* if $E/K$ is CM (resp. $K$ is totally real) and if there exists an orthogonal basis of $V$ for which the diagonal elements of the associated Gram matrix of $(V, \Phi)$ are either all totally positive or all totally negative. In the former case, $V$ is said to be *positive definite*, while in the latter case it is *negative definite*. In all the other cases, we say that $V$ is *indefinite*.

`is_regular`

— Method`is_regular(V::AbstractSpace) -> Bool`

Return whether the space `V`

is regular, that is, if the Gram matrix has full rank.

`is_quadratic`

— Method`is_quadratic(V::AbstractSpace) -> Bool`

Return whether the space `V`

is quadratic.

`ishermitian`

— Method`is_hermitian(V::AbstractSpace) -> Bool`

Return whether the space `V`

is hermitian.

`is_positive_definite`

— Method`is_positive_definite(V::AbstractSpace) -> Bool`

Return whether the space `V`

is positive definite.

`is_negative_definite`

— Method`is_negative_definite(V::AbstractSpace) -> Bool`

Return whether the space `V`

is negative definite.

`is_definite`

— Method`is_definite(V::AbstractSpace) -> Bool`

Return whether the space `V`

is definite.

Note that the `is_hermitian`

function tests whether the space is non-quadratic.

### Examples

`julia> K, a = CyclotomicRealSubfield(7);`

`julia> Kt, t = K["t"];`

`julia> E, b = number_field(t^2-a*t+1, "b");`

`julia> Q = quadratic_space(K, K[0 1; 1 0]);`

`julia> H = hermitian_space(E, 3);`

`julia> is_regular(Q), is_regular(H)`

`(true, true)`

`julia> is_quadratic(Q), ishermitian(H)`

`(true, true)`

`julia> is_definite(Q), is_positive_definite(H)`

`(false, true)`

## Inner products and diagonalization

`gram_matrix`

— Method`gram_matrix(V::AbstractSpace, M::MatElem) -> MatElem`

Return the Gram matrix of the rows of `M`

with respect to the Gram matrix of the space `V`

.

`gram_matrix`

— Method`gram_matrix(V::AbstractSpace, S::Vector{Vector}) -> MatElem`

Return the Gram matrix of the sequence `S`

with respect to the Gram matrix of the space `V`

.

`inner_product`

— Method`inner_product(V::AbstractSpace, v::Vector, w::Vector) -> FieldElem`

Return the inner product of `v`

and `w`

with respect to the bilinear form of the space `V`

.

`orthogonal_basis`

— Method`orthogonal_basis(V::AbstractSpace) -> MatElem`

Return a matrix `M`

, such that the rows of `M`

form an orthogonal basis of the space `V`

.

`diagonal`

— Method`diagonal(V::AbstractSpace) -> Vector{FieldElem}`

Return a vector of elements $a_1,\dotsc,a_n$ such that the space `V`

is isometric to the diagonal space $\langle a_1,\dotsc,a_n \rangle$.

The elements are contained in the fixed field of `V`

.

`restrict_scalars`

— Method```
restrict_scalars(V::AbstractSpace, K::QQField,
alpha::FieldElem = one(base_ring(V)))
-> QuadSpace, AbstractSpaceRes
```

Given a space $(V, \Phi)$ and a subfield `K`

of the base algebra `E`

of `V`

, return the quadratic space `W`

obtained by restricting the scalars of $(V, \alpha\Phi)$ to `K`

, together with the map `f`

for extending the scalars back. The form on the restriction is given by $Tr \circ \Phi$ where $Tr: E \to K$ is the trace form. The rescaling factor $\alpha$ is set to 1 by default.

Note that for now one can only restrict scalars to $\mathbb Q$.

### Examples

`julia> K, a = CyclotomicRealSubfield(7);`

`julia> Kt, t = K["t"];`

`julia> E, b = number_field(t^2-a*t+1, "b");`

`julia> Q = quadratic_space(K, K[0 1; 1 0]);`

`julia> H = hermitian_space(E, 3);`

`julia> gram_matrix(Q, K[1 1; 2 0])`

`[2 2] [2 0]`

`julia> gram_matrix(H, E[1 0 0; 0 1 0; 0 0 1])`

`[1 0 0] [0 1 0] [0 0 1]`

`julia> inner_product(Q, K[1 1], K[0 2])`

`[2]`

`julia> orthogonal_basis(H)`

`[1 0 0] [0 1 0] [0 0 1]`

`julia> diagonal(Q), diagonal(H)`

`(nf_elem[1, -1], nf_elem[1, 1, 1])`

## Equivalence

Let $(V, \Phi)$ and $(V', \Phi')$ be spaces over the same extension $E/K$. A *homomorphism of spaces* from $V$ to $V'$ is a $E$-linear mapping $f \colon V \to V'$ such that for all $x,y \in V$, one has

\[ \Phi'(f(x), f(y)) = \Phi(x,y).\]

An automorphism of spaces is called an *isometry* and a monomorphism is called an *embedding*.

`hasse_invariant`

— Method`hasse_invariant(V::QuadSpace, p::Union{InfPlc, NfOrdIdl}) -> Int`

Returns the Hasse invariant of the quadratic space `V`

at `p`

. This is equal to the product of local Hilbert symbols $(a_i, a_j)_p$, $i < j$, where $V$ is isometric to $\langle a_1, \dotsc, a_n\rangle$. If `V`

is degenerate return the hasse invariant of `V/radical(V)`

.

`witt_invariant`

— Method`witt_invariant(V::QuadSpace, p::Union{InfPlc, NfOrdIdl}) -> Int`

Returns the Witt invariant of the quadratic space `V`

at `p`

.

See [Definition 3.2.1, Kir16].

`is_isometric`

— Method`is_isometric(L::AbstractSpace, M::AbstractSpace) -> Bool`

Return whether the spaces `L`

and `M`

are isometric.

`is_isometric`

— Method`is_isometric(L::AbstractSpace, M::AbstractSpace, p::Union{InfPlc, NfOrdIdl}) -> Bool`

Return whether the spaces `L`

and `M`

are isometric over the completion at `p`

.

`invariants`

— Method```
invariants(M::QuadSpace)
-> FieldElem, Dict{NfOrdIdl, Int}, Vector{Tuple{InfPlc, Int}}
```

Returns a tuple `(n, k, d, H, I)`

of invariants of `M`

, which determine the isometry class completely. Here `n`

is the dimension. The dimension of the kernel is `k`

. The element `d`

is the determinant of a Gram matrix of the non-degenerate part, `H`

contains the non-trivial Hasse invariants and `I`

contains for each real place the negative index of inertia.

Note that `d`

is determined only modulo squares.

### Examples

For instance, for the case of $Q$ and the totally ramified prime $\mathfrak p$ of $O_K$ above $7$, one can get:

`julia> K, a = CyclotomicRealSubfield(7);`

`julia> Q = quadratic_space(K, K[0 1; 1 0]);`

`julia> OK = maximal_order(K);`

`julia> p = prime_decomposition(OK, 7)[1][1];`

`julia> hasse_invariant(Q, p), witt_invariant(Q, p)`

`(1, 1)`

`julia> Q2 = quadratic_space(K, K[-1 0; 0 1]);`

`julia> is_isometric(Q, Q2, p)`

`true`

`julia> is_isometric(Q, Q2)`

`true`

`julia> invariants(Q2)`

`(2, 0, -1, Dict{NfOrdIdl, Int64}(), Tuple{InfPlc{AnticNumberField, Hecke.NumFieldEmbNfAbs}, Int64}[(Infinite place corresponding to (Complex embedding corresponding to -1.80 of maximal real subfield of cyclotomic field of order 7), 1), (Infinite place corresponding to (Complex embedding corresponding to -0.45 of maximal real subfield of cyclotomic field of order 7), 1), (Infinite place corresponding to (Complex embedding corresponding to 1.25 of maximal real subfield of cyclotomic field of order 7), 1)])`

## Embeddings

Let $(V, \Phi)$ and $(V', \Phi')$ be two spaces over the same extension $E/K$, and let $\sigma \colon V \to V'$ be an $E$-linear morphism. $\sigma$ is called a *representation* of $V$ into $V'$ if for all $x \in V$

\[ \Phi'(\sigma(x), \sigma(x)) = \Phi(x,x).\]

In such a case, $V$ is said to be *represented* by $V'$ and $\sigma$ can be seen as an embedding of $V$ into $V'$. This representation property can be also tested locally with respect to the completions at some finite places. Note that in both quadratic and hermitian cases, completions are taken at finite places of the fixed field $K$.

`is_locally_represented_by`

— Method`is_locally_represented_by(U::T, V::T, p::NfOrdIdl) where T <: AbstractSpace -> Bool`

Given two spaces `U`

and `V`

over the same algebra `E`

, and a prime ideal `p`

in the maximal order $\mathcal O_K$ of their fixed field `K`

, return whether `U`

is represented by `V`

locally at `p`

, i.e. whether $U_p$ embeds in $V_p$.

`is_represented_by`

— Method`is_represented_by(U::T, V::T) where T <: AbstractSpace -> Bool`

Given two spaces `U`

and `V`

over the same algebra `E`

, return whether `U`

is represented by `V`

, i.e. whether `U`

embeds in `V`

.

### Examples

Still using the spaces $Q$ and $H$, we can decide whether some other spaces embed respectively locally or globally into $Q$ or $H$:

`julia> K, a = CyclotomicRealSubfield(7);`

`julia> Kt, t = K["t"];`

`julia> E, b = number_field(t^2-a*t+1, "b");`

`julia> Q = quadratic_space(K, K[0 1; 1 0]);`

`julia> H = hermitian_space(E, 3);`

`julia> OK = maximal_order(K);`

`julia> p = prime_decomposition(OK, 7)[1][1];`

`julia> Q2 = quadratic_space(K, K[-1 0; 0 1]);`

`julia> H2 = hermitian_space(E, E[-1 0 0; 0 1 0; 0 0 -1]);`

`julia> is_locally_represented_by(Q2, Q, p)`

`true`

`julia> is_represented_by(Q2, Q)`

`true`

`julia> is_locally_represented_by(H2, H, p)`

`true`

`julia> is_represented_by(H2, H)`

`false`

## Categorical constructions

One can construct direct sums of spaces of the same kind. Since those are also direct products, they are called biproducts in this context. Depending on the user usage, one of the following three methods can be called to obtain the direct sum of a finite collection of spaces. Note that the corresponding copies of the original spaces in the direct sum are pairwise orthogonal.

`direct_sum`

— Method`direct_sum(M::ModuleFP{T}...; task::Symbol = :sum) where T`

Given modules $M_1\dots M_n$, say, return the direct sum $\bigoplus_{i=1}^n M_i$.

Additionally, return

- a vector containing the canonical injections $M_i\to\bigoplus_{i=1}^n M_i$ if
`task = :sum`

(default), - a vector containing the canonical projections $\bigoplus_{i=1}^n M_i\to M_i$ if
`task = :prod`

, - two vectors containing the canonical injections and projections, respectively, if
`task = :both`

, - none of the above maps if
`task = :none`

.

```
direct_sum(x::Vararg{T}) where T <: AbstractSpace -> T, Vector{AbstractSpaceMor}
direct_sum(x::Vector{T}) where T <: AbstractSpace -> T, Vector{AbstractSpaceMor}
```

Given a collection of quadratic or hermitian spaces $V_1, \ldots, V_n$, return their direct sum $V := V_1 \oplus \ldots \oplus V_n$, together with the injections $V_i \to V$.

For objects of type `AbstractSpace`

, finite direct sums and finite direct products agree and they are therefore called biproducts. If one wants to obtain `V`

as a direct product with the projections $V \to V_i$, one should call `direct_product(x)`

. If one wants to obtain `V`

as a biproduct with the injections $V_i \to V$ and the projections $V \to V_i$, one should call `biproduct(x)`

.

`direct_sum(g1::QuadSpaceCls, g2::QuadSpaceCls) -> QuadSpaceCls`

Return the isometry class of the direct sum of two representatives.

`direct_product`

— Method`direct_product(F::FreeMod{T}...; task::Symbol = :prod) where T`

Given free modules $F_1\dots F_n$, say, return the direct product $\prod_{i=1}^n F_i$.

Additionally, return

- a vector containing the canonical projections $\prod_{i=1}^n F_i\to F_i$ if
`task = :prod`

(default), - a vector containing the canonical injections $F_i\to\prod_{i=1}^n F_i$ if
`task = :sum`

, - two vectors containing the canonical projections and injections, respectively, if
`task = :both`

, - none of the above maps if
`task = :none`

.

`direct_product(M::ModuleFP{T}...; task::Symbol = :prod) where T`

Given modules $M_1\dots M_n$, say, return the direct product $\prod_{i=1}^n M_i$.

Additionally, return

- a vector containing the canonical projections $\prod_{i=1}^n M_i\to M_i$ if
`task = :prod`

(default), - a vector containing the canonical injections $M_i\to\prod_{i=1}^n M_i$ if
`task = :sum`

, - two vectors containing the canonical projections and injections, respectively, if
`task = :both`

, - none of the above maps if
`task = :none`

.

```
direct_product(algebras::AlgAss...; task::Symbol = :sum)
-> AlgAss, Vector{AbsAlgAssMor}, Vector{AbsAlgAssMor}
direct_product(algebras::Vector{AlgAss}; task::Symbol = :sum)
-> AlgAss, Vector{AbsAlgAssMor}, Vector{AbsAlgAssMor}
```

Returns the algebra $A = A_1 \times \cdots \times A_k$. `task`

can be ":sum", ":prod", ":both" or ":none" and determines which canonical maps are computed as well: ":sum" for the injections, ":prod" for the projections.

```
direct_product(x::Vararg{T}) where T <: AbstractSpace -> T, Vector{AbstractSpaceMor}
direct_product(x::Vector{T}) where T <: AbstractSpace -> T, Vector{AbstractSpaceMor}
```

Given a collection of quadratic or hermitian spaces $V_1, \ldots, V_n$, return their direct product $V := V_1 \times \ldots \times V_n$, together with the projections $V \to V_i$.

For objects of type `AbstractSpace`

, finite direct sums and finite direct products agree and they are therefore called biproducts. If one wants to obtain `V`

as a direct sum with the injections $V_i \to V$, one should call `direct_sum(x)`

. If one wants to obtain `V`

as a biproduct with the injections $V_i \to V$ and the projections $V \to V_i$, one should call `biproduct(x)`

.

`biproduct`

— Method```
biproduct(x::Vararg{T}) where T <: AbstractSpace -> T, Vector{AbstractSpaceMor}, Vector{AbstractSpaceMor}
biproduct(x::Vector{T}) where T <: AbstractSpace -> T, Vector{AbstractSpaceMor}, Vector{AbstractSpaceMor}
```

Given a collection of quadratic or hermitian spaces $V_1, \ldots, V_n$, return their biproduct $V := V_1 \oplus \ldots \oplus V_n$, together with the injections $V_i \to V$ and the projections $V \to V_i$.

For objects of type `AbstractSpace`

, finite direct sums and finite direct products agree and they are therefore called biproducts. If one wants to obtain `V`

as a direct sum with the injections $V_i \to V$, one should call `direct_sum(x)`

. If one wants to obtain `V`

as a direct product with the projections $V \to V_i$, one should call `direct_product(x)`

.

### Example

`julia> E, b = cyclotomix_field_as_cm_extensions(7);`

`ERROR: UndefVarError: cyclotomix_field_as_cm_extensions not defined`

`julia> H = hermitian_space(E, 3);`

`julia> H2 = hermitian_space(E, E[-1 0 0; 0 1 0; 0 0 -1]);`

`julia> H3, inj, proj = biproduct(H, H2)`

`(Hermitian space of dimension 6, AbstractSpaceMor[Map with following data Domain: ======= Hermitian space of dimension 3 Codomain: ========= Hermitian space of dimension 6, Map with following data Domain: ======= Hermitian space of dimension 3 Codomain: ========= Hermitian space of dimension 6], AbstractSpaceMor[Map with following data Domain: ======= Hermitian space of dimension 6 Codomain: ========= Hermitian space of dimension 3, Map with following data Domain: ======= Hermitian space of dimension 6 Codomain: ========= Hermitian space of dimension 3])`

`julia> is_one(matrix(compose(inj[1], proj[1])))`

`true`

`julia> is_zero(matrix(compose(inj[1], proj[2])))`

`true`

## Orthogonality operations

`orthogonal_complement`

— Method`orthogonal_complement(V::AbstractSpace, M::T) where T <: MatElem -> T`

Given a space `V`

and a subspace `W`

with basis matrix `M`

, return a basis matrix of the orthogonal complement of `W`

inside `V`

.

`orthogonal_projection`

— Method`orthogonal_projection(V::AbstractSpace, M::T) where T <: MatElem -> AbstractSpaceMor`

Given a space `V`

and a non-degenerate subspace `W`

with basis matrix `M`

, return the endomorphism of `V`

corresponding to the projection onto the complement of `W`

in `V`

.

### Example

`julia> K, a = CyclotomicRealSubfield(7);`

`julia> Kt, t = K["t"];`

`julia> Q = quadratic_space(K, K[0 1; 1 0]);`

`julia> orthogonal_complement(Q, matrix(K, 1, 2, [1 0]))`

`[1 0]`

## Isotropic spaces

Let $(V, \Phi)$ be a space over $E/K$ and let $\mathfrak p$ be a place in $K$. $V$ is said to be *isotropic* locally at $\mathfrak p$ if there exists an element $x \in V_{\mathfrak p}$ such that $\Phi_{\mathfrak p}(x,x) = 0$, where $\Phi_{\mathfrak p}$ is the continuous extension of $\Phi$ to $V_{\mathfrak p} \times V_{\mathfrak p}$.

`is_isotropic`

— Method`is_isotropic(V::AbstractSpace, p::Union{NfOrdIdl, InfPlc}) -> Bool`

Given a space `V`

and a place `p`

in the fixed field `K`

of `V`

, return whether the completion of `V`

at `p`

is isotropic.

### Example

`julia> K, a = CyclotomicRealSubfield(7);`

`julia> Kt, t = K["t"];`

`julia> E, b = number_field(t^2-a*t+1, "b");`

`julia> H = hermitian_space(E, 3);`

`julia> OK = maximal_order(K);`

`julia> p = prime_decomposition(OK, 7)[1][1];`

`julia> is_isotropic(H, p)`

`true`

## Hyperbolic spaces

Let $(V, \Phi)$ be a space over $E/K$ and let $\mathfrak p$ be a prime ideal of $\mathcal O_K$. $V$ is said to be *hyperbolic* locally at $\mathfrak p$ if the completion $V_{\mathfrak p}$ of $V$ can be decomposed as an orthogonal sum of hyperbolic planes. The hyperbolic plane is the space $(H, \Psi)$ of rank 2 over $E/K$ such that there exists a basis $e_1, e_2$ of $H$ such that $\Psi(e_1, e_1) = \Psi(e_2, e_2) = 0$ and $\Psi(e_1, e_2) = 1$.

`is_locally_hyperbolic`

— Method`is_locally_hyperbolic(V::Hermspace, p::NfOrdIdl) -> Bool`

Return whether the completion of the hermitian space `V`

over $E/K$ at the prime ideal `p`

of $\mathcal O_K$ is hyperbolic.

### Example

`julia> K, a = CyclotomicRealSubfield(7);`

`julia> Kt, t = K["t"];`

`julia> E, b = number_field(t^2-a*t+1, "b");`

`julia> H = hermitian_space(E, 3);`

`julia> OK = maximal_order(K);`

`julia> p = prime_decomposition(OK, 7)[1][1];`

`julia> is_locally_hyperbolic(H, p)`

`false`