Documentation

TauCeti.AlgebraicGeometry.AdicSpace.Spa.Localization.Basic

The adic spectrum of a rational localisation lies over the rational subset #

Roadmap Layer 3.1 attaches to a rational subset U = R(T/s) of X = Spa(A, A⁺) the complete topological coordinate ring A_U = A⟨T/s⟩ together with its ring of integral elements A_U⁺, and asks for a natural homeomorphism

Spa (A_U, A_U⁺) ≃ U.

This file builds the map underlying that homeomorphism and proves that it lands in U. The structure map A → A⟨T/s⟩ is continuous and carries A⁺ into A_U⁺, so TauCeti.ValuationSpectrum.spaComap already gives a continuous map

Spa (A_U, A_U⁺) → Spa (A, A⁺),

and the content here is that its image is contained in R(T/s), so that it corestricts to a continuous map into the rational subset.

The two valuation-theoretic conditions cutting out R(T/s) come from the two defining features of the localisation. The denominator s becomes a unit in A⟨T/s⟩, so no point of Spa (A_U, A_U⁺) has it in its support; and each fraction t/s lies in A_U⁺, so every point is sub-unit on it, which after clearing the denominator says v(t) ≤ v(s). Neither uses a Huber hypothesis, so both are extracted first as a statement about an arbitrary continuous homomorphism inverting s.

The reverse map — extending a point of R(T/s) to a continuous valuation on the completed localisation — is not constructed here; it is the remaining half of the roadmap's homeomorphism. Its pre-completion form is TauCeti.AlgebraicGeometry.AdicSpace.Spa.Localization.Surjective, which extends a point of R(T/s) to the uncompleted localisation Aₛ and so identifies the image of Spa (Aₛ, Aₛ⁺) with R(T/s); passing from Aₛ to A⟨T/s⟩ is still open.

Main definitions #

Main results #

Provenance #

The mathematics is Wedhorn's §8.1 description of the coordinate ring of a rational subset; the proofs here are direct and follow no existing formalisation. AINTLIB — the roadmap's designated prior formalisation of this material — was not consulted for this file: no checkout of it was available in the authoring environment. Nothing is ported.

References #

The rational localisation #

Throughout this section S is an algebraic localisation of A away from s, carrying the localisation topology of TauCeti.Huber.PairOfDefinition.locTopology, and A⟨T/s⟩ is its separated completion. The three letIs that name the uniformity and its two companions are the ones every statement about A⟨T/s⟩ carries.

noncomputable def TauCeti.ValuationSpectrum.spaComapLoc {A : Type u_1} [CommRing A] [TopologicalSpace A] [IsTopologicalRing A] (P : Huber.PairOfDefinition A) (Aplus : Subring A) (T : Finset A) (s : A) (S : Type u_2) [CommRing S] [Algebra A S] [IsLocalization.Away s S] (hden : P.HasDenominatorPower T s S) :
↑(spa (P.completedPlusSubring Aplus T s S hden)) → ↑(spa Aplus)

The map Spa (A_U, A_U⁺) → Spa (A, A⁺) induced by the structure map A → A⟨T/s⟩: the structure map is continuous and carries A⁺ into A_U⁺, which is all TauCeti.ValuationSpectrum.spaComap needs.

Equations
Instances For
    @[simp]
    theorem TauCeti.ValuationSpectrum.spaComapLoc_val {A : Type u_1} [CommRing A] [TopologicalSpace A] [IsTopologicalRing A] (P : Huber.PairOfDefinition A) (Aplus : Subring A) (T : Finset A) (s : A) (S : Type u_2) [CommRing S] [Algebra A S] [IsLocalization.Away s S] (hden : P.HasDenominatorPower T s S) (v : ↑(spa (P.completedPlusSubring Aplus T s S hden))) :
    ↑(spaComapLoc P Aplus T s S hden v) = comap (P.toCompletionLoc T s S hden) ↑v

    The underlying point of spaComapLoc is the pullback along the structure map. The body is sealed across the module boundary, so this is how a consumer computes with it.

    theorem TauCeti.ValuationSpectrum.continuous_spaComapLoc {A : Type u_1} [CommRing A] [TopologicalSpace A] [IsTopologicalRing A] (P : Huber.PairOfDefinition A) (Aplus : Subring A) (T : Finset A) (s : A) (S : Type u_2) [CommRing S] [Algebra A S] [IsLocalization.Away s S] (hden : P.HasDenominatorPower T s S) :
    Continuous (spaComapLoc P Aplus T s S hden)

    spaComapLoc is continuous.

    theorem TauCeti.ValuationSpectrum.spaComapLoc_eq_comp {A : Type u_1} [CommRing A] [TopologicalSpace A] [IsTopologicalRing A] (P : Huber.PairOfDefinition A) (Aplus : Subring A) (T : Finset A) (s : A) (S : Type u_2) [CommRing S] [Algebra A S] [IsLocalization.Away s S] (hden : P.HasDenominatorPower T s S) (hloc : Continuous ⇑(algebraMap A S)) (hlocp : ∀ a ∈ Aplus, (algebraMap A S) a ∈ (integralClosure (↥(Algebra.adjoin (↥Aplus) (Set.range fun (t : ↥T) => Localization.divBy (↑t) s))) S).toSubring) (hcpl : Continuous ⇑UniformSpace.Completion.coeRingHom) (hcplp : ∀ x ∈ (integralClosure (↥(Algebra.adjoin (↥Aplus) (Set.range fun (t : ↥T) => Localization.divBy (↑t) s))) S).toSubring, UniformSpace.Completion.coeRingHom x ∈ P.completedPlusSubring Aplus T s S hden) :
    spaComapLoc P Aplus T s S hden = spaComap (algebraMap A S) hloc Aplus (integralClosure (↥(Algebra.adjoin (↥Aplus) (Set.range fun (t : ↥T) => Localization.divBy (↑t) s))) S).toSubring hlocp ∘ spaComap UniformSpace.Completion.coeRingHom hcpl (integralClosure (↥(Algebra.adjoin (↥Aplus) (Set.range fun (t : ↥T) => Localization.divBy (↑t) s))) S).toSubring (P.completedPlusSubring Aplus T s S hden) hcplp

    Pullback along the structure map factors through the uncompleted localization. The structure map ρ : A → A⟨T/s⟩ is the localization map A → Aₛ followed by the completion map Aₛ → A⟨T/s⟩, so spaComapLoc is the composite of the two induced maps of adic spectra.

    Both factors are ordinary spaComaps with a visible plus ring, which is what makes the generic descent results for a localization and for a map with dense range applicable; neither is available for ρ itself.

    The two continuity proofs and the two plus-ring conditions are quantified rather than fixed, so that a consumer can supply exactly the proofs its own spaComap was built from.

    theorem TauCeti.ValuationSpectrum.spaComapLoc_mem_rationalSubset {A : Type u_1} [CommRing A] [TopologicalSpace A] [IsTopologicalRing A] (P : Huber.PairOfDefinition A) (Aplus : Subring A) (T : Finset A) (s : A) (S : Type u_2) [CommRing S] [Algebra A S] [IsLocalization.Away s S] (hden : P.HasDenominatorPower T s S) (v : ↑(spa (P.completedPlusSubring Aplus T s S hden))) :
    ↑(spaComapLoc P Aplus T s S hden v) ∈ rationalSubset Aplus T s

    Every point of Spa (A_U, A_U⁺) lies over the rational subset R(T/s) — the half of roadmap Layer 3.1's homeomorphism Spa (A_U, A_U⁺) ≃ R(T/s) that the localisation supplies directly.

    The denominator is inverted in A⟨T/s⟩, so it is off the support of every point, and each t/s lies in A_U⁺, so every point is sub-unit on it.

    theorem TauCeti.ValuationSpectrum.range_spaComapLoc_subset {A : Type u_1} [CommRing A] [TopologicalSpace A] [IsTopologicalRing A] (P : Huber.PairOfDefinition A) (Aplus : Subring A) (T : Finset A) (s : A) (S : Type u_2) [CommRing S] [Algebra A S] [IsLocalization.Away s S] (hden : P.HasDenominatorPower T s S) :
    Set.range (spaComapLoc P Aplus T s S hden) ⊆ Subtype.val ⁻¹' rationalSubset Aplus T s

    The range of spaComapLoc is contained in the rational subset R(T/s), as a subset of the subtype ↥(Spa (A, A⁺)).

    noncomputable def TauCeti.ValuationSpectrum.spaLocToRationalSubset {A : Type u_1} [CommRing A] [TopologicalSpace A] [IsTopologicalRing A] (P : Huber.PairOfDefinition A) (Aplus : Subring A) (T : Finset A) (s : A) (S : Type u_2) [CommRing S] [Algebra A S] [IsLocalization.Away s S] (hden : P.HasDenominatorPower T s S) :
    ↑(spa (P.completedPlusSubring Aplus T s S hden)) → ↑(Subtype.val ⁻¹' rationalSubset Aplus T s)

    The canonical map Spa (A_U, A_U⁺) → R(T/s): spaComapLoc corestricted to the rational subset it lands in. This is the map that roadmap Layer 3.1 asks to be a homeomorphism.

    Equations
    • One or more equations did not get rendered due to their size.
    Instances For
      @[simp]
      theorem TauCeti.ValuationSpectrum.spaLocToRationalSubset_val {A : Type u_1} [CommRing A] [TopologicalSpace A] [IsTopologicalRing A] (P : Huber.PairOfDefinition A) (Aplus : Subring A) (T : Finset A) (s : A) (S : Type u_2) [CommRing S] [Algebra A S] [IsLocalization.Away s S] (hden : P.HasDenominatorPower T s S) (v : ↑(spa (P.completedPlusSubring Aplus T s S hden))) :
      ↑(spaLocToRationalSubset P Aplus T s S hden v) = spaComapLoc P Aplus T s S hden v

      The corestriction forgets to spaComapLoc.

      The canonical map into the rational subset is continuous.

      theorem TauCeti.ValuationSpectrum.rationalSubset_image_toCompletionLoc_eq_spa {A : Type u_1} [CommRing A] [TopologicalSpace A] [IsTopologicalRing A] (P : Huber.PairOfDefinition A) (Aplus : Subring A) (T : Finset A) (s : A) (S : Type u_2) [CommRing S] [Algebra A S] [IsLocalization.Away s S] (hden : P.HasDenominatorPower T s S) :
      rationalSubset (P.completedPlusSubring Aplus T s S hden) (Finset.image (⇑(P.toCompletionLoc T s S hden)) T) ((P.toCompletionLoc T s S hden) s) = spa (P.completedPlusSubring Aplus T s S hden)

      In the coordinate ring of U = R(T/s), the rational subset cut out by the images of the defining data is the whole adic spectrum.

      This is the point of passing to A_U: the conditions v(t) ≤ v(s) ≠ 0 that carve U out of Spa (A, A⁺) become vacuous over A⟨T/s⟩, because there s is a unit and each t/s is a sub-unit. It is the degenerate case of Wedhorn's comparison of rational subsets of U with rational subsets of X (§8.2), and it is what makes Spa (A_U, A_U⁺) a candidate for U rather than for a proper subset of it.

      The coordinate ring of an empty rational subset has empty adic spectrum. Every point of Spa (A_U, A_U⁺) lies over a point of R(T/s), so there is none to have when R(T/s) is empty.