Documentation

Atlas.GeometryOfManifolds.code.WeinsteinNeighborhood

The linear map $V \to E^*$ sending $v \mapsto (e \mapsto \Omega(v, e))$, where $\Omega$ is a bilinear form on $V$ and $E \subseteq V$ a subspace.

Instances For

    The kernel of symplecticToDual Ω E is the symplectic orthogonal complement $E^{\perp_\Omega} = \{v \in V : \Omega(v, e) = 0 \text{ for all } e \in E\}$.

    A Lagrangian subspace $E$ is contained in the kernel of symplecticToDual Ω E, since $E = E^{\perp_\Omega}$.

    For $E \subseteq V$ Lagrangian, the descent of symplecticToDual Ω E to a linear map $V/E \to E^*$ given by $[v] \mapsto \Omega(v, \cdot)|_E$.

    Instances For

      The induced map $V/E \to E^*$ is injective for $E$ Lagrangian, because the kernel of the original map is exactly $E^{\perp_\Omega} = E$.

      For $E$ a Lagrangian subspace, $\dim(V/E) = \dim E^* = \dim E$, since $\dim E = \tfrac12 \dim V$.

      Linear Lagrangian normal bundle identification. For $E$ a Lagrangian subspace of a symplectic vector space $(V, \Omega)$, there is a natural linear isomorphism $V/E \xrightarrow{\sim} E^*$, identifying the "normal space" to $E$ with $T^*E$.

      Instances For
        theorem cohomology_isomorphism_surjectivity {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_U : Type u_3} {VF_U : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_U : DifferentialFormSpace Ω_U VF_U] (i : DFSMorphism Ω_X VF_X Ω_U VF_U) (π : DFSMorphism Ω_U VF_U Ω_X VF_X) (h_retraction : ∀ {p : } (α : Ω_X p), i.pullback (π.pullback α) = α) {p : } (α : Ω_X p) :
        ∃ (β : Ω_U p), i.pullback β = α

        Surjectivity on cohomology from a retraction. If $\pi : U \to X$ is a retraction of $i : X \hookrightarrow U$ (so $i^* \pi^* = \mathrm{id}$), then $i^*$ is surjective on all forms: every $\alpha \in \Omega^p(X)$ lifts to $\beta = \pi^*\alpha \in \Omega^p(U)$ with $i^*\beta = \alpha$.

        structure HasRelativeHomotopyOperator (Ω : Type u_1) (VF : Type u_2) [inst : DifferentialFormSpace Ω VF] (Ω_X : Type u_3) (VF_X : Type u_4) [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) :
        Type (max u_1 u_3)

        A relative homotopy operator for an inclusion $i : X \hookrightarrow U$ packages the chain-level data showing that $i^*$ is a homotopy equivalence: a retraction $\pi : U \to X$, a chain homotopy $K : \Omega^{p+1}(U) \to \Omega^p(U)$ satisfying $dK + Kd = \mathrm{id} - \pi^* i^*$, and the additional condition $i^* K = 0$ (so that primitives vanish on $X$).

        Instances For

          Cartan's magic formula applied to a chosen "scaling" vector field $V$: $d\iota_V \beta + \iota_V d\beta = \mathcal{L}_V \beta$.

          theorem homotopy_formula_from_cartan_and_ftc {Ω : Type u_1} {VF : Type u_2} [inst : DifferentialFormSpace Ω VF] {Ω_X : Type u_3} {VF_X : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) (π : DFSMorphism Ω VF Ω_X VF_X) (V : VF) (Φ : {p : } → Ω pΩ p) (K : {p : } → Ω (p + 1)Ω p) (hK_def : ∀ {p : } (β : Ω (p + 1)), K β = Φ (DifferentialFormSpace.ι V β)) (hΦ_add : ∀ {p : } (γ₁ γ₂ : Ω p), Φ (γ₁ + γ₂) = Φ γ₁ + Φ γ₂) (hΦ_d : ∀ {p : } (γ : Ω p), DifferentialFormSpace.d VF (Φ γ) = Φ (DifferentialFormSpace.d VF γ)) (hFTC : ∀ {p : } (β : Ω (p + 1)), Φ (DifferentialFormSpace.L V β) = β - π.pullback (i.pullback β)) {p : } (β : Ω (p + 1)) :

          Derivation of the chain homotopy formula $dK + Kd = \mathrm{id} - \pi^* i^*$ from Cartan's formula together with a "fundamental theorem of calculus" identity $\Phi(\mathcal{L}_V \beta) = \beta - \pi^*(i^*\beta)$, where $K := \Phi \circ \iota_V$.

          theorem K_smul_from_axioms {Ω : Type u_1} {VF : Type u_2} [inst : DifferentialFormSpace Ω VF] (V : VF) (Φ : {p : } → Ω pΩ p) (hΦ_smul : ∀ {p : } (r : ) (γ : Ω p), Φ (r γ) = r Φ γ) {p : } (r : ) (α : Ω (p + 1)) :

          $\mathbb{R}$-linearity of $K = \Phi \circ \iota_V$ in its argument, derived from the linearity of $\iota_V$ and the scalar-multiplicativity of $\Phi$.

          theorem K_vanishes_on_X_from_axioms {Ω : Type u_1} {VF : Type u_2} [inst : DifferentialFormSpace Ω VF] {Ω_X : Type u_3} {VF_X : Type u_4} [_inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) (V : VF) (Φ : {p : } → Ω pΩ p) (hιV_vanish : ∀ {p : } (β : Ω (p + 1)), i.pullback (DifferentialFormSpace.ι V β) = 0) (hΦ_preserve : ∀ {p : } (γ : Ω p), i.pullback γ = 0i.pullback (Φ γ) = 0) {p : } (β : Ω (p + 1)) :

          $K\beta = \Phi(\iota_V \beta)$ pulls back to zero on $X$, since $\iota_V$ already does and $\Phi$ preserves vanishing on $X$.

          structure HasTubularHomotopyPrimitives (Ω : Type u_1) (VF : Type u_2) [inst : DifferentialFormSpace Ω VF] (Ω_X : Type u_3) (VF_X : Type u_4) [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) :
          Type (max (max u_1 u_2) u_3)

          Tubular homotopy primitives for an inclusion $i : X \hookrightarrow U$: the data of a retraction $\pi$, an Euler-like vector field $V$, and a "radial integration" operator $\Phi$ satisfying additivity, scalar-multiplicativity, commutation with $d$, an "FTC" identity $\Phi(\mathcal{L}_V \beta) = \beta - \pi^* i^*\beta$, and the conditions $i^*(\iota_V \beta) = 0$ and $i^* \Phi = i^*$.

          Such primitives canonically produce a relative homotopy operator via $K := \Phi \circ \iota_V$.

          Instances For
            theorem HasTubularHomotopyPrimitives.Φ_preserves_vanishing {Ω : Type u_1} {VF : Type u_2} [inst : DifferentialFormSpace Ω VF] {Ω_X : Type u_3} {VF_X : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] {i : DFSMorphism Ω_X VF_X Ω VF} (hp : HasTubularHomotopyPrimitives Ω VF Ω_X VF_X i) {p : } (γ : Ω p) ( : i.pullback γ = 0) :
            i.pullback (hp.Φ γ) = 0

            If $i^*\gamma = 0$ and $\Phi$ fixes pullbacks to $X$, then $i^*(\Phi \gamma) = 0$.

            noncomputable def hasRelativeHomotopyOperator_of_primitives {Ω : Type u_1} {VF : Type u_2} [inst : DifferentialFormSpace Ω VF] {Ω_X : Type u_3} {VF_X : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) (hp : HasTubularHomotopyPrimitives Ω VF Ω_X VF_X i) :
            HasRelativeHomotopyOperator Ω VF Ω_X VF_X i

            Construct a HasRelativeHomotopyOperator from the more granular tubular homotopy primitives, via $K := \Phi \circ \iota_V$.

            Instances For
              theorem relative_poincare_lemma {Ω_U : Type u_1} {VF_U : Type u_2} [inst_U : DifferentialFormSpace Ω_U VF_U] {Ω_X : Type u_3} {VF_X : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω_U VF_U) (hK : HasRelativeHomotopyOperator Ω_U VF_U Ω_X VF_X i) { : } (β : Ω_U ( + 1)) (hclosed : DifferentialFormSpace.d VF_U β = 0) (hvanish : i.pullback β = 0) :
              ∃ (μ : Ω_U ), DifferentialFormSpace.d VF_U μ = β i.pullback μ = 0

              Relative Poincaré lemma. Given a relative homotopy operator for $i : X \hookrightarrow U$, every closed $(\ell+1)$-form $\beta \in \Omega^{\ell+1}(U)$ whose pullback $i^*\beta = 0$ vanishes admits a primitive $\mu \in \Omega^\ell(U)$ with $d\mu = \beta$ and $i^*\mu = 0$.

              theorem relative_poincare_lemma_from_primitives {Ω_U : Type u_1} {VF_U : Type u_2} [inst_U : DifferentialFormSpace Ω_U VF_U] {Ω_X : Type u_3} {VF_X : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω_U VF_U) (hp : HasTubularHomotopyPrimitives Ω_U VF_U Ω_X VF_X i) { : } (β : Ω_U ( + 1)) (hclosed : DifferentialFormSpace.d VF_U β = 0) (hvanish : i.pullback β = 0) :
              ∃ (μ : Ω_U ), DifferentialFormSpace.d VF_U μ = β i.pullback μ = 0

              Variant of the relative Poincaré lemma stated directly in terms of HasTubularHomotopyPrimitives: the explicit primitive is $\mu = \Phi(\iota_V \beta)$.

              structure HasDeformationRetract {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_U : Type u_3} {VF_U : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_U : DifferentialFormSpace Ω_U VF_U] (i : DFSMorphism Ω_X VF_X Ω_U VF_U) :
              Type (max u_1 u_3)

              An algebraic deformation retract of $U$ onto $X$: a relative homotopy operator together with an injectivity condition at degree 0 (forms of degree 0 with vanishing differential and vanishing pullback are zero).

              Instances For
                theorem relative_poincare_lemma_of_retract {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_U : Type u_3} {VF_U : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_U : DifferentialFormSpace Ω_U VF_U] (i : DFSMorphism Ω_X VF_X Ω_U VF_U) (hdr : HasDeformationRetract i) { : } (β : Ω_U ( + 1)) (hclosed : DifferentialFormSpace.d VF_U β = 0) (hvanish : i.pullback β = 0) :
                ∃ (μ : Ω_U ), DifferentialFormSpace.d VF_U μ = β i.pullback μ = 0

                Relative Poincaré lemma packaged for the HasDeformationRetract interface.

                theorem cohomology_isomorphism_injectivity_of_retract {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_U : Type u_3} {VF_U : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_U : DifferentialFormSpace Ω_U VF_U] (i : DFSMorphism Ω_X VF_X Ω_U VF_U) (hdr : HasDeformationRetract i) (p : ) (β : Ω_U (p + 1)) :
                DifferentialFormSpace.d VF_U β = 0(∃ (γ : Ω_X p), i.pullback β = DifferentialFormSpace.d VF_X γ)∃ (η : Ω_U p), β = DifferentialFormSpace.d VF_U η

                Injectivity on cohomology from a deformation retract. If $\beta \in \Omega^{p+1}(U)$ is closed and its restriction $i^*\beta$ is exact on $X$ (i.e. $i^*\beta = d\gamma$), then $\beta$ itself is exact on $U$.

                theorem cohomology_isomorphism {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_U₁ : Type u_3} {VF_U₁ : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_U₁ : DifferentialFormSpace Ω_U₁ VF_U₁] (i : DFSMorphism Ω_X VF_X Ω_U₁ VF_U₁) (hdr : HasDeformationRetract i) :
                (∀ {p : } (α : Ω_X p), ∃ (β : Ω_U₁ p), i.pullback β = α) (∀ (p : ) (β : Ω_U₁ (p + 1)), DifferentialFormSpace.d VF_U₁ β = 0(∃ (γ : Ω_X p), i.pullback β = DifferentialFormSpace.d VF_X γ)∃ (η : Ω_U₁ p), β = DifferentialFormSpace.d VF_U₁ η) ∀ (β : Ω_U₁ 0), DifferentialFormSpace.d VF_U₁ β = 0i.pullback β = 0β = 0

                Cohomology isomorphism. A deformation retract of $U_1$ onto $X$ gives the three algebraic statements together asserting that $i^* : H^*(U_1, \mathbb{R}) \to H^*(X, \mathbb{R})$ is an isomorphism: surjectivity on forms, injectivity on cohomology in positive degree, and injectivity in degree $0$.

                structure NormalBundleDFS (Ω_X : Type u_1) (VF_X : Type u_2) (Ω_NX : Type u_3) (VF_NX : Type u_4) [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_NX : DifferentialFormSpace Ω_NX VF_NX] :
                Type (max u_1 u_3)

                Abstract data for the normal bundle $NX \to X$ at the level of differential forms: a zero-section morphism $X \to NX$ and a projection $NX \to X$ whose composition pulled back to forms is the identity on $\Omega^*(X)$.

                Instances For
                  structure IsSubmanifoldInclusion {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_M : Type u_3} {VF_M : Type u_4} [DifferentialFormSpace Ω_X VF_X] [DifferentialFormSpace Ω_M VF_M] (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (dimM dimX : ) :

                  Predicate that the DFS morphism $i : X \hookrightarrow M$ comes from a smooth submanifold inclusion of dimensions $\dim X \le \dim M$.

                  • dim_le : dimX dimM
                  Instances For
                    structure IsOpenNeighborhoodOf {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_U : Type u_3} {VF_U : Type u_4} {Ω_N : Type u_5} {VF_N : Type u_6} [DifferentialFormSpace Ω_X VF_X] [inst_U : DifferentialFormSpace Ω_U VF_U] [DifferentialFormSpace Ω_N VF_N] (incl_X_N : DFSMorphism Ω_X VF_X Ω_N VF_N) (incl_U_N : DFSMorphism Ω_U VF_U Ω_N VF_N) (incl_X_U : DFSMorphism Ω_X VF_X Ω_U VF_U) :

                    Compatibility predicate stating that $U \subseteq N$ is an open neighborhood of $X \subseteq N$: the inclusion $X \hookrightarrow N$ factors as $X \hookrightarrow U \hookrightarrow N$ on differential forms.

                    Instances For
                      structure IsDFSDiffeomorphism {Ω₁ : Type u_1} {VF₁ : Type u_2} {Ω₂ : Type u_3} {VF₂ : Type u_4} [inst₁ : DifferentialFormSpace Ω₁ VF₁] [inst₂ : DifferentialFormSpace Ω₂ VF₂] (φ : DFSMorphism Ω₁ VF₁ Ω₂ VF₂) (φ_inv : DFSMorphism Ω₂ VF₂ Ω₁ VF₁) :

                      A pair of DFS morphisms $\varphi, \varphi^{-1}$ form a DFS diffeomorphism if their pullbacks are mutually inverse on all differential forms.

                      Instances For
                        class HasTubularExpMapData {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_NX : Type u_3} {VF_NX : Type u_4} {Ω_M : Type u_5} {VF_M : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_NX : DifferentialFormSpace Ω_NX VF_NX] [inst_M : DifferentialFormSpace Ω_M VF_M] (nb : NormalBundleDFS Ω_X VF_X Ω_NX VF_NX) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (dimM dimX : ) (hSubmfd : IsSubmanifoldInclusion i dimM dimX) :
                        Type (max (max (max (max u_1 u_3) u_5) (u_7 + 1)) (u_8 + 1))

                        The data of an exponential map used in proving the tubular neighborhood theorem: an open neighborhood $U_0$ of $X$ in $NX$ together with a smooth map $\exp : U_0 \to M$ whose restriction to the zero section is $i : X \hookrightarrow M$.

                        Instances
                          class HasIFTData {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_NX : Type u_3} {VF_NX : Type u_4} {Ω_M : Type u_5} {VF_M : Type u_6} {Ω_U₀ : Type u_7} {VF_U₀ : Type u_8} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_NX : DifferentialFormSpace Ω_NX VF_NX] [inst_M : DifferentialFormSpace Ω_M VF_M] [inst_U₀ : DifferentialFormSpace Ω_U₀ VF_U₀] (nb : NormalBundleDFS Ω_X VF_X Ω_NX VF_NX) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (exp_map : DFSMorphism Ω_U₀ VF_U₀ Ω_M VF_M) (j₀ : DFSMorphism Ω_U₀ VF_U₀ Ω_NX VF_NX) (s₀ : DFSMorphism Ω_X VF_X Ω_U₀ VF_U₀) (hU₀_nbhd : IsOpenNeighborhoodOf nb.zeroSection j₀ s₀) (hd_exp_id : ∀ {p : } (α : Ω_M p), s₀.pullback (exp_map.pullback α) = i.pullback α) :
                          Type (max (max (max (u_10 + 1) u_5) u_7) (u_9 + 1))

                          The output of the inverse function theorem applied to the exponential map: a DFS-diffeomorphism $\varphi : U_0 \xrightarrow{\sim} U_1$ where $U_1 \subseteq M$, together with the factorization $\exp = \mathrm{incl}_{U_1 \hookrightarrow M} \circ \varphi$.

                          Instances
                            structure TubularNeighborhoodData {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_NX : Type u_3} {VF_NX : Type u_4} {Ω_M : Type u_5} {VF_M : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_NX : DifferentialFormSpace Ω_NX VF_NX] [inst_M : DifferentialFormSpace Ω_M VF_M] (nb : NormalBundleDFS Ω_X VF_X Ω_NX VF_NX) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) :
                            Type (max (max (max (max (max (max u_1 (u_10 + 1)) u_3) u_5) (u_7 + 1)) (u_8 + 1)) (u_9 + 1))

                            Packaging of the conclusion of the tubular neighborhood theorem: open neighborhoods $U_0$ of $X$ in $NX$ and $U_1$ of $X$ in $M$, a DFS-diffeomorphism $\varphi : U_0 \xrightarrow{\sim} U_1$ that restricts to the identity on $X$, together with the relevant inclusions and the open-neighborhood data on both sides.

                            Instances For
                              def tubular_neighborhood_theorem {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_NX : Type u_3} {VF_NX : Type u_4} {Ω_M : Type u_5} {VF_M : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_NX : DifferentialFormSpace Ω_NX VF_NX] [inst_M : DifferentialFormSpace Ω_M VF_M] (nb : NormalBundleDFS Ω_X VF_X Ω_NX VF_NX) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (dimM dimX : ) (hSubmfd : IsSubmanifoldInclusion i dimM dimX) [expData : HasTubularExpMapData nb i dimM dimX hSubmfd] [iftData : HasIFTData nb i HasTubularExpMapData.exp_map HasTubularExpMapData.j₀ HasTubularExpMapData.s₀ ] :

                              Tubular neighborhood theorem (assembly). Given an exponential map data and the output of the inverse function theorem, this assembles the full tubular neighborhood data: $X \subseteq M$ admits a tubular neighborhood $U_1 \cong U_0 \subseteq NX$.

                              Instances For
                                noncomputable def tubularExpMapData_exists {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_NX : Type u_3} {VF_NX : Type u_4} {Ω_M : Type u_5} {VF_M : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_NX : DifferentialFormSpace Ω_NX VF_NX] [inst_M : DifferentialFormSpace Ω_M VF_M] (nb : NormalBundleDFS Ω_X VF_X Ω_NX VF_NX) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (dimM dimX : ) (hSubmfd : IsSubmanifoldInclusion i dimM dimX) :
                                HasTubularExpMapData nb i dimM dimX hSubmfd

                                Existence of exponential-map data for a submanifold inclusion (axiomatic; ultimately to be supplied by Riemannian or local-coordinate methods).

                                Instances For
                                  noncomputable def iftData_exists {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_NX : Type u_3} {VF_NX : Type u_4} {Ω_M : Type u_5} {VF_M : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_NX : DifferentialFormSpace Ω_NX VF_NX] [inst_M : DifferentialFormSpace Ω_M VF_M] (nb : NormalBundleDFS Ω_X VF_X Ω_NX VF_NX) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (dimM dimX : ) (hSubmfd : IsSubmanifoldInclusion i dimM dimX) (expData : HasTubularExpMapData nb i dimM dimX hSubmfd) :

                                  Existence of the inverse-function-theorem data from exponential-map data (axiomatic; the proof goes through the IFT applied to the differential of $\exp$ which is the identity on the zero section).

                                  Instances For
                                    noncomputable def tubularNeighborhoodTheorem_DFS {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_NX : Type u_3} {VF_NX : Type u_4} {Ω_M : Type u_5} {VF_M : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_NX : DifferentialFormSpace Ω_NX VF_NX] [inst_M : DifferentialFormSpace Ω_M VF_M] (nb : NormalBundleDFS Ω_X VF_X Ω_NX VF_NX) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (dimM dimX : ) (hSubmfd : IsSubmanifoldInclusion i dimM dimX) :

                                    Tubular neighborhood theorem (DFS version). For any submanifold inclusion $i : X \hookrightarrow M$ with normal-bundle data $nb$, there exist open neighborhoods $U_0$ of $X$ in $NX$ and $U_1$ of $X$ in $M$, and a DFS-diffeomorphism $\varphi : U_0 \xrightarrow{\sim} U_1$ that is the identity on $X$.

                                    Instances For
                                      theorem moser_cartan_computation {Ω : Type u_1} {VF : Type u_2} [inst : DifferentialFormSpace Ω VF] (v : VF) (ω : Ω 2) (μ : Ω 1) (hω_closed : DifferentialFormSpace.d VF ω = 0) ( : DifferentialFormSpace.ι v ω = -μ) :

                                      Cartan-formula computation underlying Moser's trick: if $\omega$ is closed and $\iota_v \omega = -\mu$, then $\mathcal{L}_v \omega = -d\mu$.

                                      theorem moser_cancellation {Ω : Type u_1} {VF : Type u_2} [inst : DifferentialFormSpace Ω VF] (v : VF) (ω : Ω 2) (μ : Ω 1) (σ : Ω 2) (hω_closed : DifferentialFormSpace.d VF ω = 0) ( : DifferentialFormSpace.ι v ω = -μ) (hdμ : DifferentialFormSpace.d VF μ = σ) :

                                      Moser's cancellation identity: with $\omega$ closed, $\iota_v\omega = -\mu$, and $d\mu = \sigma$, one has $\mathcal{L}_v\omega + \sigma = 0$, the time-derivative of the family $\omega_t$ along the Moser flow.

                                      theorem moser_difference_closed_and_vanishes {Ω : Type u_1} {VF : Type u_2} [inst : DifferentialFormSpace Ω VF] {Ω_X : Type u_3} {VF_X : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) (ω₀ ω₁ : Ω 2) (hclosed₀ : DifferentialFormSpace.d VF ω₀ = 0) (hclosed₁ : DifferentialFormSpace.d VF ω₁ = 0) (hagree : i.pullback ω₀ = i.pullback ω₁) :
                                      DifferentialFormSpace.d VF (ω₁ - ω₀) = 0 i.pullback (ω₁ - ω₀) = 0

                                      If $\omega_0, \omega_1$ are closed 2-forms with the same pullback to $X$, then their difference $\omega_1 - \omega_0$ is closed and vanishes when pulled back to $X$.

                                      theorem moser_interpolation_closed {Ω : Type u_1} {VF : Type u_2} [inst : DifferentialFormSpace Ω VF] (ω₀ ω₁ : Ω 2) (hclosed₀ : DifferentialFormSpace.d VF ω₀ = 0) (hclosed₁ : DifferentialFormSpace.d VF ω₁ = 0) (t : ) :
                                      DifferentialFormSpace.d VF ((1 - t) ω₀ + t ω₁) = 0

                                      The interpolated 2-form $\omega_t = (1-t)\omega_0 + t\omega_1$ is closed whenever both $\omega_0, \omega_1$ are.

                                      def ContractionSurjective {Ω : Type u_1} {VF : Type u_2} [inst : DifferentialFormSpace Ω VF] (ω : Ω 2) :

                                      The 2-form $\omega$ has the property that contraction $v \mapsto \iota_v \omega$ hits every 1-form — equivalently, $\omega^\flat : TM \to T^*M$ is surjective.

                                      Instances For
                                        theorem moser_equation_solvable {Ω : Type u_1} {VF : Type u_2} [inst : DifferentialFormSpace Ω VF] (ω : Ω 2) (μ : Ω 1) (hsurj : ContractionSurjective ω) :
                                        ∃ (v : VF), DifferentialFormSpace.ι v ω = -μ

                                        If contraction with $\omega$ is surjective, the Moser equation $\iota_v \omega = -\mu$ has a solution $v$.

                                        theorem local_moser_diffeomorphism_core {Ω : Type u_1} {VF : Type u_2} [inst : DifferentialFormSpace Ω VF] {Ω_X : Type u_3} {VF_X : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) (ω₀ ω₁ : Ω 2) (hclosed₀ : DifferentialFormSpace.d VF ω₀ = 0) (hclosed₁ : DifferentialFormSpace.d VF ω₁ = 0) (_hnd₀ : Function.Injective fun (X : VF) => DifferentialFormSpace.ι X ω₀) (_hnd₁ : Function.Injective fun (X : VF) => DifferentialFormSpace.ι X ω₁) (_hagree : ∀ (V : VF), i.pullback (DifferentialFormSpace.ι V ω₀) = i.pullback (DifferentialFormSpace.ι V ω₁)) (h_homotopy_op : HasRelativeHomotopyOperator Ω VF Ω_X VF_X i) (h_pb_agree : (∀ (V : VF), i.pullback (DifferentialFormSpace.ι V ω₀) = i.pullback (DifferentialFormSpace.ι V ω₁))i.pullback ω₀ = i.pullback ω₁) (h_contr_surj : (Function.Injective fun (X : VF) => DifferentialFormSpace.ι X ω₀)ContractionSurjective ω₀) (h_ode_flow : ∀ (μ : Ω 1) (v : VF), DifferentialFormSpace.d VF μ = ω₁ - ω₀i.pullback μ = 0DifferentialFormSpace.ι v ω₀ = -μDifferentialFormSpace.L v ω₀ + (ω₁ - ω₀) = 0∃ (φ : DFSMorphism Ω VF Ω VF) (φ_inv : DFSMorphism Ω VF Ω VF), φ.pullback ω₁ = ω₀ (∀ {p : } (α : Ω p), φ_inv.pullback (φ.pullback α) = α) (∀ {p : } (α : Ω p), φ.pullback (φ_inv.pullback α) = α) ∀ {p : } (α : Ω p), i.pullback (φ.pullback α) = i.pullback α) :
                                        ∃ (φ : DFSMorphism Ω VF Ω VF) (φ_inv : DFSMorphism Ω VF Ω VF), φ.pullback ω₁ = ω₀ (∀ {p : } (α : Ω p), φ_inv.pullback (φ.pullback α) = α) (∀ {p : } (α : Ω p), φ.pullback (φ_inv.pullback α) = α) ∀ {p : } (α : Ω p), i.pullback (φ.pullback α) = i.pullback α

                                        Local Moser theorem (algebraic core). Given closed nondegenerate 2-forms $\omega_0, \omega_1$ that agree on $X$ (in the strong sense that $i^*(\iota_V \omega_0) = i^*(\iota_V \omega_1)$ for all vector fields $V$), and given a relative homotopy operator, the surjectivity of contraction with $\omega_0$, and an ODE flow existence assumption, there exists a diffeomorphism $\varphi$ with $\varphi^*\omega_1 = \omega_0$ that is the identity on $X$.

                                        def IsProperSubDFS {Ω₁ : Type u_1} {VF₁ : Type u_2} [DifferentialFormSpace Ω₁ VF₁] {Ω₂ : Type u_3} {VF₂ : Type u_4} [DifferentialFormSpace Ω₂ VF₂] (i : DFSMorphism Ω₁ VF₁ Ω₂ VF₂) :

                                        The DFS morphism $i : U \hookrightarrow M$ comes from a proper open inclusion $U \subsetneq M$: pullback on smooth functions is not surjective.

                                        Instances For
                                          theorem moserFlowNeighborhoodRestriction {Ω : Type u_Ω} {VF : Type u_VF} [inst : DifferentialFormSpace Ω VF] {Ω_X : Type u_1} {VF_X : Type u_2} [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) (ω₀ ω₁ : Ω 2) (φ_M φ_M_inv : DFSMorphism Ω VF Ω VF) (hpull : φ_M.pullback ω₁ = ω₀) (hinv_l : ∀ {p : } (α : Ω p), φ_M_inv.pullback (φ_M.pullback α) = α) (hinv_r : ∀ {p : } (α : Ω p), φ_M.pullback (φ_M_inv.pullback α) = α) (hfix : ∀ {p : } (α : Ω p), i.pullback (φ_M.pullback α) = i.pullback α) :
                                          ∃ (Ω_U₀ : Type u_Ω) (VF_U₀ : Type u_VF) (inst_U₀ : DifferentialFormSpace Ω_U₀ VF_U₀) (Ω_U₁ : Type u_Ω) (VF_U₁ : Type u_VF) (inst_U₁ : DifferentialFormSpace Ω_U₁ VF_U₁) (incl₀ : DFSMorphism Ω_U₀ VF_U₀ Ω VF) (incl₁ : DFSMorphism Ω_U₁ VF_U₁ Ω VF) (sub₀ : DFSMorphism Ω_X VF_X Ω_U₀ VF_U₀) (sub₁ : DFSMorphism Ω_X VF_X Ω_U₁ VF_U₁) (φ : DFSMorphism Ω_U₀ VF_U₀ Ω_U₁ VF_U₁) (φ_inv : DFSMorphism Ω_U₁ VF_U₁ Ω_U₀ VF_U₀), φ.pullback (incl₁.pullback ω₁) = incl₀.pullback ω₀ (∀ {p : } (α : Ω_U₀ p), φ.pullback (φ_inv.pullback α) = α) (∀ {p : } (α : Ω_U₁ p), φ_inv.pullback (φ.pullback α) = α) (∀ {p : } (α : Ω_U₁ p), sub₀.pullback (φ.pullback α) = sub₁.pullback α) (∀ {p : } (α : Ω p), sub₀.pullback (incl₀.pullback α) = i.pullback α) (∀ {p : } (α : Ω p), sub₁.pullback (incl₁.pullback α) = i.pullback α) IsProperSubDFS incl₀ IsProperSubDFS incl₁

                                          A global Moser flow $\varphi_M : M \to M$ matching $\omega_1$ to $\omega_0$ and fixing $X$ restricts to a local diffeomorphism between open neighborhoods $U_0, U_1$ of $X$.

                                          theorem ode_flow_exists {Ω : Type u_Ω} {VF : Type u_VF} [inst : DifferentialFormSpace Ω VF] {Ω_X : Type u_1} {VF_X : Type u_2} [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) (ω₀ ω₁ : Ω 2) (μ : Ω 1) (v : VF) (hdμ : DifferentialFormSpace.d VF μ = ω₁ - ω₀) (hμ_vanish : i.pullback μ = 0) ( : DifferentialFormSpace.ι v ω₀ = -μ) (hcancel : DifferentialFormSpace.L v ω₀ + (ω₁ - ω₀) = 0) :
                                          ∃ (φ : DFSMorphism Ω VF Ω VF) (φ_inv : DFSMorphism Ω VF Ω VF), φ.pullback ω₁ = ω₀ (∀ {p : } (α : Ω p), φ_inv.pullback (φ.pullback α) = α) (∀ {p : } (α : Ω p), φ.pullback (φ_inv.pullback α) = α) ∀ {p : } (α : Ω p), i.pullback (φ.pullback α) = i.pullback α

                                          Existence of the Moser ODE flow (axiomatic). Given the data of the Moser equation $\iota_v\omega_0 = -\mu$ and Cartan's cancellation $\mathcal{L}_v\omega_0 + (\omega_1-\omega_0) = 0$ with $\mu$ vanishing on $X$, the time-1 flow of $v$ yields a diffeomorphism $\varphi$ with $\varphi^*\omega_1 = \omega_0$ that is the identity on $X$.

                                          theorem pullbackAgreement_of_contractionAgreement {Ω : Type u_Ω} {VF : Type u_VF} [inst : DifferentialFormSpace Ω VF] {Ω_X : Type u_1} {VF_X : Type u_2} [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) (ω₀ ω₁ : Ω 2) (hagree : ∀ (V : VF), i.pullback (DifferentialFormSpace.ι V ω₀) = i.pullback (DifferentialFormSpace.ι V ω₁)) :
                                          i.pullback ω₀ = i.pullback ω₁

                                          If $\iota_V \omega_0$ and $\iota_V \omega_1$ have the same pullback to $X$ for every $V$, then $\omega_0$ and $\omega_1$ themselves have the same pullback.

                                          theorem contractionSurjective_of_injective {Ω : Type u_Ω} {VF : Type u_VF} [inst : DifferentialFormSpace Ω VF] (ω₀ : Ω 2) (hnd : Function.Injective fun (X : VF) => DifferentialFormSpace.ι X ω₀) :

                                          A 2-form $\omega_0$ for which contraction is injective is also nondegenerate in the sense that contraction is surjective (in finite dimension; here taken axiomatically).

                                          noncomputable def tubularHomotopyPrimitives_exists_uv {Ω : Type u_Ω} {VF : Type u_VF} [inst : DifferentialFormSpace Ω VF] {Ω_X : Type u_1} {VF_X : Type u_2} [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) :

                                          Axiomatic existence of tubular homotopy primitives for any submanifold inclusion $i : X \hookrightarrow U$ (universe-polymorphic version).

                                          Instances For
                                            theorem local_moser_theorem4_book {Ω : Type u_Ω} {VF : Type u_VF} [inst : DifferentialFormSpace Ω VF] {Ω_X : Type u_1} {VF_X : Type u_2} [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) (ω₀ ω₁ : Ω 2) (hclosed₀ : DifferentialFormSpace.d VF ω₀ = 0) (hclosed₁ : DifferentialFormSpace.d VF ω₁ = 0) (hnd₀ : Function.Injective fun (X : VF) => DifferentialFormSpace.ι X ω₀) (hnd₁ : Function.Injective fun (X : VF) => DifferentialFormSpace.ι X ω₁) (hagree : ∀ (V : VF), i.pullback (DifferentialFormSpace.ι V ω₀) = i.pullback (DifferentialFormSpace.ι V ω₁)) :
                                            ∃ (Ω_U₀ : Type u_Ω) (VF_U₀ : Type u_VF) (inst_U₀ : DifferentialFormSpace Ω_U₀ VF_U₀) (Ω_U₁ : Type u_Ω) (VF_U₁ : Type u_VF) (inst_U₁ : DifferentialFormSpace Ω_U₁ VF_U₁) (incl₀ : DFSMorphism Ω_U₀ VF_U₀ Ω VF) (incl₁ : DFSMorphism Ω_U₁ VF_U₁ Ω VF) (sub₀ : DFSMorphism Ω_X VF_X Ω_U₀ VF_U₀) (sub₁ : DFSMorphism Ω_X VF_X Ω_U₁ VF_U₁) (φ : DFSMorphism Ω_U₀ VF_U₀ Ω_U₁ VF_U₁) (φ_inv : DFSMorphism Ω_U₁ VF_U₁ Ω_U₀ VF_U₀), φ.pullback (incl₁.pullback ω₁) = incl₀.pullback ω₀ (∀ {p : } (α : Ω_U₀ p), φ.pullback (φ_inv.pullback α) = α) (∀ {p : } (α : Ω_U₁ p), φ_inv.pullback (φ.pullback α) = α) (∀ {p : } (α : Ω_U₁ p), sub₀.pullback (φ.pullback α) = sub₁.pullback α) (∀ {p : } (α : Ω p), sub₀.pullback (incl₀.pullback α) = i.pullback α) (∀ {p : } (α : Ω p), sub₁.pullback (incl₁.pullback α) = i.pullback α) IsProperSubDFS incl₀ IsProperSubDFS incl₁

                                            Local Moser theorem (Theorem 4 in the book). For a submanifold $X \hookrightarrow M$ and two closed nondegenerate 2-forms $\omega_0, \omega_1$ on $M$ that agree to first order along $X$, there exist open neighborhoods $U_0, U_1$ of $X$ in $M$ and a diffeomorphism $\varphi : U_0 \xrightarrow{\sim} U_1$ with $\varphi^*\omega_1 = \omega_0$ and $\varphi|_X = \mathrm{id}$.

                                            structure IsClosedSubmanifold {Ω_M : Type u_1} {VF_M : Type u_2} {Ω_X : Type u_3} {VF_X : Type u_4} [DifferentialFormSpace Ω_M VF_M] [DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω_M VF_M) :

                                            $i : X \hookrightarrow M$ is a closed submanifold inclusion: every differential form on $X$ extends to a form on $M$ (i.e. $i^* : \Omega^*(M) \to \Omega^*(X)$ is surjective).

                                            • pullback_surj (p : ) (α : Ω_X p) : ∃ (β : Ω_M p), i.pullback β = α
                                            Instances For
                                              structure TubularNeighborhoodGeometricData {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_M : Type u_3} {VF_M : Type u_4} {Ω_TX : Type u_5} {VF_TX : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_M : DifferentialFormSpace Ω_M VF_M] [inst_TX : DifferentialFormSpace Ω_TX VF_TX] [hCot : CotangentBundleDFS Ω_X VF_X Ω_TX VF_TX] (S_M : SymplecticManifold Ω_M VF_M) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (hLag : IsLagrangianSubmanifold S_M i (2 * CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX) (CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX)) :
                                              Type (max (max (max (max (max (max u_1 (u_10 + 1)) u_3) u_5) (u_7 + 1)) (u_8 + 1)) (u_9 + 1))

                                              All the geometric data produced by combining the tubular neighborhood theorem with the Lagrangian normal-bundle identification $NX \cong T^*X$: a common neighborhood $U$ of $X$ carrying two symplectic forms $\omega_0$ (from $T^*X$) and $\omega_1$ (from $M$) that agree on $X$, together with the auxiliary inclusions and a $V$-side neighborhood diffeomorphic to $U$ identifying $\omega_1$ with the symplectic form on $M$.

                                              Instances For
                                                noncomputable def tubularNeighborhood_geometric_exists {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_M : Type u_3} {VF_M : Type u_4} {Ω_TX : Type u_5} {VF_TX : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_M : DifferentialFormSpace Ω_M VF_M] [inst_TX : DifferentialFormSpace Ω_TX VF_TX] [hCot : CotangentBundleDFS Ω_X VF_X Ω_TX VF_TX] (S_M : SymplecticManifold Ω_M VF_M) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (_hClosed : IsClosedSubmanifold i) (hLag : IsLagrangianSubmanifold S_M i (2 * CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX) (CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX)) :

                                                Axiomatic existence of TubularNeighborhoodGeometricData for any closed Lagrangian submanifold $X$ of a symplectic manifold $(M, \omega)$.

                                                Instances For
                                                  theorem whitneyExtension_firstOrderAgreement {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_M : Type u_3} {VF_M : Type u_4} {Ω_TX : Type u_5} {VF_TX : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_M : DifferentialFormSpace Ω_M VF_M] [inst_TX : DifferentialFormSpace Ω_TX VF_TX] [hCot : CotangentBundleDFS Ω_X VF_X Ω_TX VF_TX] (S_M : SymplecticManifold Ω_M VF_M) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (_hClosed : IsClosedSubmanifold i) (hLag : IsLagrangianSubmanifold S_M i (2 * CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX) (CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX)) (geo : TubularNeighborhoodGeometricData S_M i hLag) :
                                                  have inst_U := geo.inst_U; ∀ (V : geo.VF_U), geo.j.pullback (DifferentialFormSpace.ι V geo.ω₀) = geo.j.pullback (DifferentialFormSpace.ι V geo.ω₁)

                                                  First-order agreement on $X$. A Whitney-extension argument shows that the two candidate symplectic forms $\omega_0$ and $\omega_1$ supplied by the geometric data agree on $X$ at first order, i.e. $j^*(\iota_V\omega_0) = j^*(\iota_V\omega_1)$ for every vector field $V$.

                                                  noncomputable def tubularHomotopyPrimitives_exists {Ω_U : Type u_1} {VF_U : Type u_2} [inst_U : DifferentialFormSpace Ω_U VF_U] {Ω_X : Type u_3} {VF_X : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] (j : DFSMorphism Ω_X VF_X Ω_U VF_U) :
                                                  HasTubularHomotopyPrimitives Ω_U VF_U Ω_X VF_X j

                                                  Axiomatic existence of tubular homotopy primitives for $j : X \hookrightarrow U$ (unrestricted universe version).

                                                  Instances For
                                                    theorem moserTransport {Ω_U : Type u_1} {VF_U : Type u_2} [inst_U : DifferentialFormSpace Ω_U VF_U] {Ω_X : Type u_3} {VF_X : Type u_4} [inst_X : DifferentialFormSpace Ω_X VF_X] (j : DFSMorphism Ω_X VF_X Ω_U VF_U) (ω₀ ω₁ : Ω_U 2) (hclosed₀ : DifferentialFormSpace.d VF_U ω₀ = 0) (hclosed₁ : DifferentialFormSpace.d VF_U ω₁ = 0) (h_nondeg : Function.Injective fun (X : VF_U) => DifferentialFormSpace.ι X ω₀) (hagree : ∀ (V : VF_U), j.pullback (DifferentialFormSpace.ι V ω₀) = j.pullback (DifferentialFormSpace.ι V ω₁)) (h_homotopy : HasRelativeHomotopyOperator Ω_U VF_U Ω_X VF_X j) :
                                                    ∃ (φ : DFSMorphism Ω_U VF_U Ω_U VF_U) (φ_inv : DFSMorphism Ω_U VF_U Ω_U VF_U), φ.pullback ω₁ = ω₀ (∀ {p : } (α : Ω_U p), φ_inv.pullback (φ.pullback α) = α) (∀ {p : } (α : Ω_U p), φ.pullback (φ_inv.pullback α) = α) ∀ {p : } (α : Ω_U p), j.pullback (φ.pullback α) = j.pullback α

                                                    Moser transport. On a single ambient $U$ containing $X$, given closed 2-forms $\omega_0, \omega_1$ with $\omega_0$ nondegenerate that agree to first order on $X$, together with a relative homotopy operator, there exists a DFS-diffeomorphism $\varphi : U \to U$ with $\varphi^*\omega_1 = \omega_0$ fixing $X$.

                                                    class HasMoserInfrastructure {Ω : Type u_Ω} {VF : Type u_VF} [inst : DifferentialFormSpace Ω VF] {Ω_X : Type u_1} {VF_X : Type u_2} [inst_X : DifferentialFormSpace Ω_X VF_X] (i : DFSMorphism Ω_X VF_X Ω VF) :
                                                    Type (max (max u_1 u_VF) u_Ω)

                                                    Typeclass-style bundle of all the Moser-transport infrastructure for an inclusion $i : X \hookrightarrow U$: tubular homotopy primitives plus an existence statement for the Moser diffeomorphism given the standard hypotheses.

                                                    Instances
                                                      class HasLagrangianTubularData {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_M : Type u_3} {VF_M : Type u_4} {Ω_TX : Type u_5} {VF_TX : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_M : DifferentialFormSpace Ω_M VF_M] [inst_TX : DifferentialFormSpace Ω_TX VF_TX] [hCot : CotangentBundleDFS Ω_X VF_X Ω_TX VF_TX] (S_M : SymplecticManifold Ω_M VF_M) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (hLag : IsLagrangianSubmanifold S_M i (2 * CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX) (CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX)) :
                                                      Type (max (max (max (max (max (max u_1 (u_10 + 1)) u_3) u_5) (u_7 + 1)) (u_8 + 1)) (u_9 + 1))

                                                      All the geometric and Moser-flow data needed to prove Weinstein's Lagrangian neighborhood theorem: a common neighborhood $U$ of $X$ with two symplectic forms $\omega_0$ (induced from $T^*X$) and $\omega_1$ (induced from $M$), a $V$-side neighborhood diffeomorphic to $U$ realizing $\omega_1$ as a pullback of the symplectic form on $M$, and a Moser flow $\varphi$ on $U$ matching $\omega_1$ to $\omega_0$ that fixes $X$.

                                                      Instances
                                                        theorem weinstein_lagrangian_neighborhood_core_from_data {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_M : Type u_3} {VF_M : Type u_4} {Ω_TX : Type u_5} {VF_TX : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_M : DifferentialFormSpace Ω_M VF_M] [inst_TX : DifferentialFormSpace Ω_TX VF_TX] [hCot : CotangentBundleDFS Ω_X VF_X Ω_TX VF_TX] (S_M : SymplecticManifold Ω_M VF_M) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (_hClosed : IsClosedSubmanifold i) (hLag : IsLagrangianSubmanifold S_M i (2 * CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX) (CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX)) (hTub : HasLagrangianTubularData S_M i hLag) :

                                                        Weinstein's Lagrangian neighborhood theorem (core, from data). From a HasLagrangianTubularData package, one constructs the symplectomorphism $\Phi := \psi \circ \varphi$ from the local model $(U, \omega_0)$ to $(V, S_V.\omega)$ that is the identity on $X$ in the appropriate sense.

                                                        @[reducible]
                                                        noncomputable def HasLagrangianTubularData_exists {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_M : Type u_3} {VF_M : Type u_4} {Ω_TX : Type u_5} {VF_TX : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_M : DifferentialFormSpace Ω_M VF_M] [inst_TX : DifferentialFormSpace Ω_TX VF_TX] [hCot : CotangentBundleDFS Ω_X VF_X Ω_TX VF_TX] (S_M : SymplecticManifold Ω_M VF_M) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (_hClosed : IsClosedSubmanifold i) (hLag : IsLagrangianSubmanifold S_M i (2 * CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX) (CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX)) :

                                                        Existence of HasLagrangianTubularData: combine the geometric tubular neighborhood data with the first-order agreement of $\omega_0$ and $\omega_1$ on $X$, then apply the local Moser theorem to produce the required diffeomorphism $\varphi$.

                                                        Instances For
                                                          theorem weinstein_lagrangian_neighborhood_book3 {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_M : Type u_3} {VF_M : Type u_4} {Ω_TX : Type u_5} {VF_TX : Type u_6} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_M : DifferentialFormSpace Ω_M VF_M] [inst_TX : DifferentialFormSpace Ω_TX VF_TX] [hCot : CotangentBundleDFS Ω_X VF_X Ω_TX VF_TX] (S_M : SymplecticManifold Ω_M VF_M) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (_hClosed : IsClosedSubmanifold i) (hLag : IsLagrangianSubmanifold S_M i (2 * CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX) (CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX)) :
                                                          ∃ (Ω_U₀ : Type u_7) (VF_U₀ : Type u_8) (inst_U₀ : DifferentialFormSpace Ω_U₀ VF_U₀) (ω₀ : Ω_U₀ 2) (hω₀_closed : DifferentialFormSpace.d VF_U₀ ω₀ = 0) (hω₀_nd : Function.Injective fun (X : VF_U₀) => DifferentialFormSpace.ι X ω₀) (ι_U₀ : DFSMorphism Ω_U₀ VF_U₀ Ω_TX VF_TX) (j₀ : DFSMorphism Ω_X VF_X Ω_U₀ VF_U₀) (_ : ∀ {p : } (α : Ω_TX p), j₀.pullback (ι_U₀.pullback α) = CotangentBundleDFS.zeroSection.pullback α) (Ω_V : Type u_9) (VF_V : Type u_10) (inst_V : DifferentialFormSpace Ω_V VF_V) (S_V : SymplecticManifold Ω_V VF_V) (ι_V : DFSMorphism Ω_V VF_V Ω_M VF_M) (j_V : DFSMorphism Ω_X VF_X Ω_V VF_V) (_ : ∀ {p : } (α : Ω_M p), j_V.pullback (ι_V.pullback α) = i.pullback α), have S_U₀ := { ω := ω₀, closed := hω₀_closed, nondegenerate := hω₀_nd }; ∃ (Φ : Symplectomorphism S_U₀ S_V), ∀ {p : } (α : Ω_V p), j₀.pullback (Φ.toMorphism.pullback α) = j_V.pullback α

                                                          Weinstein's Lagrangian neighborhood theorem (book form). For a symplectic manifold $(M, \omega)$ and a closed Lagrangian submanifold $i : X \hookrightarrow M$, there exist a neighborhood $U_0 \subseteq T^*X$ of the zero section, a neighborhood $V \subseteq M$ of $X$, and a symplectomorphism $\Phi : (U_0, \omega_0) \to (V, \omega)$ that intertwines the inclusion of $X$ into both.

                                                          structure DFSPointEvalX (Ω_X : Type u_1) (VF_X : Type u_2) (X : Type u_3) [inst_X : DifferentialFormSpace Ω_X VF_X] :
                                                          Type (max u_1 u_3)

                                                          Abstract "pointwise evaluation" for 1-forms on $X$ at points of an underlying point set: a predicate "$\mu$ vanishes at $x$" used to phrase critical-point arguments.

                                                          • isZeroAt : Ω_X 1XProp
                                                          Instances For
                                                            structure HasH1Vanishing (Ω_X : Type u_1) (VF_X : Type u_2) [inst_X : DifferentialFormSpace Ω_X VF_X] :

                                                            Axiomatic vanishing of $H^1(X, \mathbb{R})$: every closed 1-form on $X$ is exact, i.e. $\mu = dh$ for some smooth $h : X \to \mathbb{R}$.

                                                            Instances For
                                                              theorem compact_critical_ge_two_lagrangian {Ω_X : Type u_1} {VF_X : Type u_2} {X : Type u_3} [inst_X : DifferentialFormSpace Ω_X VF_X] [TopologicalSpace X] (ptEval : DFSPointEvalX Ω_X VF_X X) (hCompact : CompactSpace X) (h : Ω_X 0) :
                                                              ∃ (x : X) (y : X), x y ptEval.isZeroAt (DifferentialFormSpace.d VF_X h) x ptEval.isZeroAt (DifferentialFormSpace.d VF_X h) y

                                                              On a compact (Hausdorff) manifold $X$, every smooth function $h : X \to \mathbb{R}$ has at least two critical points (achieving its max and min).

                                                              structure C1CloseLagrangian {Ω_X : Type u_1} {VF_X : Type u_2} {Ω_M : Type u_3} {VF_M : Type u_4} {Ω_TX : Type u_5} {VF_TX : Type u_6} {X : Type u_7} [inst_X : DifferentialFormSpace Ω_X VF_X] [inst_M : DifferentialFormSpace Ω_M VF_M] [inst_TX : DifferentialFormSpace Ω_TX VF_TX] [hCot : CotangentBundleDFS Ω_X VF_X Ω_TX VF_TX] (ptEval : DFSPointEvalX Ω_X VF_X X) (S_M : SymplecticManifold Ω_M VF_M) (i : DFSMorphism Ω_X VF_X Ω_M VF_M) (hLag : IsLagrangianSubmanifold S_M i (2 * CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX) (CotangentBundleDFS.dimX Ω_X VF_X Ω_TX VF_TX)) :
                                                              Type u_1

                                                              The data witnessing that a $C^1$-close Lagrangian deformation $Y$ of $X$ is encoded by a closed 1-form $\mu \in \Omega^1(X)$ such that $X \cap Y$ corresponds to the zero set of $\mu$ in the Weinstein chart.

                                                              Instances For
                                                                class IsLagrangianEmbedding_mfld {E : Type u_1} [NormedAddCommGroup E] [NormedSpace E] {HH : Type u_2} [TopologicalSpace HH] (I : ModelWithCorners E HH) (M : Type u_3) [TopologicalSpace M] [ChartedSpace HH M] [IsManifold I M] (X : Type u_4) [TopologicalSpace X] [ChartedSpace HH X] [IsManifold I X] (i : XM) :

                                                                Mathlib-flavored predicate that $i : X \to M$ is a Lagrangian embedding: $i$ is smooth and injective.

                                                                Instances

                                                                  Mathlib-flavored axiomatization of $H^1(X, \mathbb{R}) = 0$ on a connected manifold: any smooth function with zero differential is constant.

                                                                  Instances
                                                                    structure C1CloseSubmanifold_mfld {E : Type u_1} [NormedAddCommGroup E] [NormedSpace E] {HH : Type u_2} [TopologicalSpace HH] (I : ModelWithCorners E HH) (M : Type u_3) [TopologicalSpace M] [ChartedSpace HH M] [IsManifold I M] (X : Type u_4) [TopologicalSpace X] [ChartedSpace HH X] [IsManifold I X] (i : XM) :
                                                                    Type u_4

                                                                    The Mathlib analog of C1CloseLagrangian: data of a smooth function $h : X \to \mathbb{R}$ whose critical points correspond exactly to intersection points of $X$ with a $C^1$-close submanifold $Y$.

                                                                    Instances For
                                                                      theorem weinstein_and_H1_give_primitive {E : Type u_1} [NormedAddCommGroup E] [NormedSpace E] {HH : Type u_2} [TopologicalSpace HH] (I : ModelWithCorners E HH) (M : Type u_3) [TopologicalSpace M] [ChartedSpace HH M] [IsManifold I M] (X : Type u_4) [TopologicalSpace X] [ChartedSpace HH X] [IsManifold I X] (i : XM) (hLag : IsLagrangianEmbedding_mfld I M X i) (hH1 : HasVanishingH1_mfld I X) (Y : C1CloseSubmanifold_mfld I M X i) :
                                                                      ∃ (h : X), ContMDiff I (modelWithCornersSelf ) h ∀ (x : X), Y.isIntersectionPt x mfderiv% h x = 0

                                                                      Combining Weinstein's theorem with $H^1(X) = 0$ produces a smooth primitive $h$ on $X$ whose critical points coincide with the intersection points of $X$ and $Y$.

                                                                      structure C1CloseLagrangianMfld {E : Type u_1} [NormedAddCommGroup E] [NormedSpace E] {HH : Type u_2} [TopologicalSpace HH] (I : ModelWithCorners E HH) (X : Type u_3) [TopologicalSpace X] [ChartedSpace HH X] [IsManifold I X] :
                                                                      Type u_3

                                                                      A slimmer variant of C1CloseSubmanifold_mfld carrying only the smooth function $h$ on $X$ and the intersection-vs-critical-point equivalence.

                                                                      Instances For
                                                                        theorem evt_ge_two_critical_points {E : Type u_1} [NormedAddCommGroup E] [NormedSpace E] {HH : Type u_2} [TopologicalSpace HH] (I : ModelWithCorners E HH) (X : Type u_3) [TopologicalSpace X] [ChartedSpace HH X] [IsManifold I X] [CompactSpace X] [Nontrivial X] (h : X) (hsmooth : ContMDiff I (modelWithCornersSelf ) h) :
                                                                        ∃ (x : X) (y : X), x y mfderiv% h x = 0 mfderiv% h y = 0

                                                                        Extreme value theorem giving $\ge 2$ critical points. On a compact nontrivial manifold $X$, every smooth $h : X \to \mathbb{R}$ has at least two distinct critical points.

                                                                        theorem lagrangian_intersection_c1_close {E : Type u_1} [NormedAddCommGroup E] [NormedSpace E] {HH : Type u_2} [TopologicalSpace HH] (I : ModelWithCorners E HH) (M : Type u_3) [TopologicalSpace M] [ChartedSpace HH M] [IsManifold I M] (X : Type u_4) [TopologicalSpace X] [ChartedSpace HH X] [IsManifold I X] [CompactSpace X] [Nontrivial X] (i : XM) [hLag : IsLagrangianEmbedding_mfld I M X i] [hH1 : HasVanishingH1_mfld I X] (Y : C1CloseSubmanifold_mfld I M X i) :
                                                                        ∃ (x : X) (y : X), x y Y.isIntersectionPt x Y.isIntersectionPt y

                                                                        Lagrangian intersection theorem for $C^1$-close deformations. If $X$ is compact, $H^1(X) = 0$, and $Y$ is a $C^1$-close Lagrangian deformation of $X$ in $M$, then $X$ and $Y$ intersect in at least two distinct points.

                                                                        Mathlib version of $H^1$ vanishing on a normed space $E$: every closed alternating 1-form on $E$ is the exterior derivative of an alternating 0-form.

                                                                        Instances For
                                                                          structure WeinsteinTubular {E : Type u_1} [NormedAddCommGroup E] [NormedSpace E] {H : Type u_2} [TopologicalSpace H] (I : ModelWithCorners E H) (M : Type u_3) [TopologicalSpace M] [ChartedSpace H M] [IsManifold I M] (X : Type u_4) [TopologicalSpace X] [ChartedSpace H X] [IsManifold I X] (ω : ME [⋀^Fin 2]→L[] ) (i : XM) (hi : ContMDiff I I i) :
                                                                          Type (max u_1 u_3)

                                                                          Mathlib-style packaging of the Weinstein tubular neighborhood: open sets $U \subseteq M$ containing $X$ and $U_0 \subseteq E \times E$ containing the zero section, with a diffeomorphism $\Phi : U_0 \xrightarrow{\sim} U$ taking the zero section into $X$ and (placeholder) preserving the symplectic structure.

                                                                          Instances For
                                                                            structure tubular_neighborhood_theorem_book_Mathlib {E : Type u_1} [NormedAddCommGroup E] [NormedSpace E] {H : Type u_2} [TopologicalSpace H] (I : ModelWithCorners E H) (M : Type u_3) [TopologicalSpace M] [ChartedSpace H M] [IsManifold I M] (X : Type u_4) [TopologicalSpace X] [ChartedSpace H X] [IsManifold I X] (i : XM) (hi : ContMDiff I I i) :
                                                                            Type (max u_1 u_3)

                                                                            Mathlib-style packaging of the (non-symplectic) tubular neighborhood theorem: neighborhoods $U_1 \ni X$ in $M$ and $U_0 \ni$ zero section in $E \times E$, and a diffeomorphism $\varphi : U_0 \to U_1$ mapping the zero section into $X$.

                                                                            Instances For
                                                                              theorem expMap_exists_Mathlib {EM : Type u_1} [NormedAddCommGroup EM] [NormedSpace EM] {HM : Type u_2} [TopologicalSpace HM] (IM : ModelWithCorners EM HM) {EX : Type u_3} [NormedAddCommGroup EX] [NormedSpace EX] {HX : Type u_4} [TopologicalSpace HX] (IX : ModelWithCorners EX HX) {EN : Type u_5} [NormedAddCommGroup EN] [NormedSpace EN] (M : Type u_6) [TopologicalSpace M] [ChartedSpace HM M] [IsManifold IM M] (X : Type u_7) [TopologicalSpace X] [ChartedSpace HX X] [IsManifold IX X] (i : XM) (hi : ContMDiff IX IM i) (hi_inj : Function.Injective i) :
                                                                              ∃ (U₀ : TopologicalSpace.Opens (EX × EN)) (exp_map : U₀M), (∀ (v : EX), (v, 0) U₀) (∀ (v : EX) (hv : (v, 0) U₀), exp_map (v, 0), hv Set.range i) (ContMDiff (modelWithCornersSelf (EX × EN)) IM fun (p : U₀) => exp_map p) (∀ (v : EX) (hv : (v, 0) U₀), ∃ (U : Set U₀), (v, 0), hv U Set.InjOn exp_map U) ∀ (x : X), i x Set.range exp_map

                                                                              Existence of an exponential-like map in the Mathlib setting (axiomatic). For any smooth injective $i : X \to M$, there exists an open set $U_0 \subseteq E_X \times E_N$ containing the zero section and a smooth map $\exp : U_0 \to M$ such that the zero section lands in $i(X)$, $\exp$ is locally injective, and $i(X) \subseteq \mathrm{im}(\exp)$.

                                                                              theorem ift_localDiffeo_Mathlib {EM : Type u_1} [NormedAddCommGroup EM] [NormedSpace EM] {HM : Type u_2} [TopologicalSpace HM] (IM : ModelWithCorners EM HM) {EX : Type u_3} [NormedAddCommGroup EX] [NormedSpace EX] {HX : Type u_4} [TopologicalSpace HX] (IX : ModelWithCorners EX HX) {EN : Type u_5} [NormedAddCommGroup EN] [NormedSpace EN] (M : Type u_6) [TopologicalSpace M] [ChartedSpace HM M] [IsManifold IM M] (X : Type u_7) [TopologicalSpace X] [ChartedSpace HX X] [IsManifold IX X] (i : XM) (hi : ContMDiff IX IM i) (hi_inj : Function.Injective i) (U₀_pre : TopologicalSpace.Opens (EX × EN)) (exp_map : U₀_preM) (hU₀_zero : ∀ (v : EX), (v, 0) U₀_pre) (hexp_zero : ∀ (v : EX) (hv : (v, 0) U₀_pre), exp_map (v, 0), hv Set.range i) (hexp_smooth : ContMDiff (modelWithCornersSelf (EX × EN)) IM fun (p : U₀_pre) => exp_map p) (hexp_local_inj : ∀ (v : EX) (hv : (v, 0) U₀_pre), ∃ (U : Set U₀_pre), (v, 0), hv U Set.InjOn exp_map U) (hexp_covers : ∀ (x : X), i x Set.range exp_map) :
                                                                              ∃ (U₀ : TopologicalSpace.Opens (EX × EN)) (U₁ : TopologicalSpace.Opens M) (φ : Diffeomorph (modelWithCornersSelf (EX × EN)) IM U₀ U₁ ), (∀ (v : EX), (v, 0) U₀) (∀ (x : X), i x U₁) (∀ (v : EX) (hv : (v, 0) U₀), (φ (v, 0), hv) Set.range i) R > 0, pU₀, p.2 < R

                                                                              IFT-based local diffeomorphism (axiomatic, Mathlib version). Starting from an exponential-like map $\exp : U_0 \to M$ with the standard properties, the inverse function theorem produces a smaller diffeomorphism $\varphi : U_0 \xrightarrow{\sim} U_1$ preserving the zero section, with $U_1 \supseteq i(X)$ and $U_0$ bounded in the "normal" direction.

                                                                              theorem tubularNeighborhoodTheorem {EM : Type u_1} [NormedAddCommGroup EM] [NormedSpace EM] {HM : Type u_2} [TopologicalSpace HM] (IM : ModelWithCorners EM HM) {EX : Type u_3} [NormedAddCommGroup EX] [NormedSpace EX] {HX : Type u_4} [TopologicalSpace HX] (IX : ModelWithCorners EX HX) {EN : Type u_5} [NormedAddCommGroup EN] [NormedSpace EN] (M : Type u_6) [TopologicalSpace M] [ChartedSpace HM M] [IsManifold IM M] (X : Type u_7) [TopologicalSpace X] [ChartedSpace HX X] [IsManifold IX X] (i : XM) (hi : ContMDiff IX IM i) (hi_inj : Function.Injective i) :
                                                                              ∃ (U₀ : TopologicalSpace.Opens (EX × EN)) (U₁ : TopologicalSpace.Opens M) (φ : Diffeomorph (modelWithCornersSelf (EX × EN)) IM U₀ U₁ ), (∀ (v : EX), (v, 0) U₀) (∀ (x : X), i x U₁) (∀ (v : EX) (hv : (v, 0) U₀), (φ (v, 0), hv) Set.range i) R > 0, pU₀, p.2 < R

                                                                              Tubular neighborhood theorem (Mathlib formulation). For any smooth injective $i : X \hookrightarrow M$, there exist open neighborhoods $U_0$ of the zero section of $E_X \times E_N$ and $U_1$ of $i(X)$ in $M$, and a diffeomorphism $\varphi : U_0 \xrightarrow{\sim} U_1$ preserving the zero section, with the normal factor of $U_0$ bounded. Combines expMap_exists_Mathlib and ift_localDiffeo_Mathlib.

                                                                              structure C1CloseLagrangian_Mathlib (E : Type u_1) [NormedAddCommGroup E] [NormedSpace E] (X : Type u_2) :
                                                                              Type (max u_1 u_2)

                                                                              Lightweight Mathlib structure recording the input data for the intersection problem: a chart $X \to E$, a closed 1-form $\mu$ on $E$, and a predicate isZeroAt on $X$ that is equivalent to $\mu$ vanishing at the chart image.

                                                                              Instances For
                                                                                theorem lagrangian_intersection_c1_close_Mathlib (E : Type u_1) [NormedAddCommGroup E] [NormedSpace E] (X : Type u_2) [TopologicalSpace X] [CompactSpace X] (hH1 : HasH1Vanishing_Mathlib E) (Y : C1CloseLagrangian_Mathlib E X) (hEVT : ∀ (f : EE [⋀^Fin 0]→L[] ), ∃ (x : X) (y : X), x y extDeriv f (Y.chart x) = 0 extDeriv f (Y.chart y) = 0) :
                                                                                ∃ (x : X) (y : X), x y Y.isZeroAt x Y.isZeroAt y

                                                                                Lagrangian intersection theorem (Mathlib version). Given $H^1$ vanishing on $E$, a $C^1$-close Lagrangian data $Y$, and an EVT-type hypothesis providing two distinct critical points of any function on $X$, we conclude that $X$ has at least two distinct intersection points with $Y$.