YukonModule.ProximityPrize.SubmissionLower.GenericSlicePoints6807.part0
DefinitionYukon_84e6160981e0608a7da1d940better-codes
Source module ProximityPrize.SubmissionLower.GenericSlicePoints6807.
Definition code
/-
UNCOMPILED. Concrete generic-fiber point construction from the pinned APIs.
The nonlinear elimination theorem RCN202.embedding_point_certificate is used
at k=0, H=U=1, G=0. Its equation becomes t-ell and its cut becomes A.
In particular IsolatedPoint is a conclusion here, not an assumed cut budget.
-/
import Definitions.Def_Yukon_196ff1de7425c8d8df8d3a11
import Definitions.Def_Yukon_07bb1fdf83fc478e7c5449e7
import Definitions.Def_Yukon_755f5ab5e1dfa660f8e11f09
set_option backward.isDefEq.respectTransparency.types false
namespace ProximityPrize.SubmissionLower.GenericSlicePoints6807
open scoped Classical BigOperators
open RCN002 RCN072 RCN264 RCN207 RCN208 RCN134 RCN084
noncomputable section
set_option autoImplicit false
set_option maxRecDepth 20000
set_option maxHeartbeats 400000
set_option Elab.async false
set_option synthInstance.maxHeartbeats 300000
variable {K E : Type} [Field K] [Field E] [IsAlgClosed E]
[Algebra K E] [Algebra (RatFunc K) E] [IsScalarTower K (RatFunc K) E]
local notation "Poly" => MvPolynomial (Fin 3) K
local notation "PE" => MvPolynomial (Fin 3) E
theorem filteredCut_zero {R : Type*} [CommRing R] (A H G : R) :
filteredCut 0 (fun _ : Fin 1 => A) H G = A := by
simp [filteredCut, Fin.sum_univ_succ]
def sliceEquation (ell : Poly) : PE :=
MvPolynomial.C (algebraMap (RatFunc K) E (RCN202.rationalVariable K)) -
scalarPolynomialMap K E ell
/-- Drop only the auxiliary regularity gate; the underlying prime is unchanged.
This is what permits the old k=0 theorem to use H=1 while preserving the original
H*c exclusion separately in the resulting point certificate. -/
abbrev gateOneComponent (F A R : Poly) (C : RegularComponent K F A R) :
RegularComponent K F (filteredCut 0 (fun _ : Fin 1 => A) 1 0) 1 :=
⟨C.1, by
classical
apply (mem_regularComponents K).mpr
constructor
· simpa only [filteredCut_zero] using regularComponent_mem K F A R C
· intro h
have hz : coordinateEvaluation K C.1 (1 : Poly) = 0 := by
apply RingHom.mem_ker.mp
change (1 : Poly) ∈ RingHom.ker (coordinateEvaluation K C.1).toRingHom
rwa [coordinateEvaluation_ker K C.1]
exact one_ne_zero (by simpa only [map_one] using hz)⟩
omit [IsAlgClosed E] [Algebra (RatFunc K) E] [IsScalarTower K (RatFunc K) E] in
theorem scalar_eval_embedding (P : Ideal Poly) [P.IsPrime]
(f : CoordinateField K P →ₐ[K] E) (A : Poly) :
MvPolynomial.eval (embeddingPoint P f) (scalarPolynomialMap K E A) =
f (coordinateEvaluation K P A) := by
have h := AlgHom.congr_fun (embeddingPoint_aeval P f) A
change MvPolynomial.eval (embeddingPoint P f)
(MvPolynomial.map (algebraMap K E) A) = _
rw [MvPolynomial.eval_map]
exact h
omit [IsAlgClosed E] [Algebra (RatFunc K) E] [IsScalarTower K (RatFunc K) E] in
theorem scalar_eval_ne_zero (P : Ideal Poly) [P.IsPrime]
(f : CoordinateField K P →ₐ[K] E) (A : Poly) (hA : A ∉ P) :
MvPolynomial.eval (embeddingPoint P f) (scalarPolynomialMap K E A) ≠ 0 := by
rw [scalar_eval_embedding]
intro h
apply hA
rw [← coordinateEvaluation_ker K P]
exact (map_eq_zero_iff f f.injective).mp h
theorem embedding_point_certificate
(F A R ell : Poly) (C : RegularComponent K F A R)
[Algebra (RatFunc K) (CoordinateField K C.1)]
[IsScalarTower K (RatFunc K) (CoordinateField K C.1)]
[FiniteDimensional (RatFunc K) (CoordinateField K C.1)]
(hj : algebraMap (RatFunc K) (CoordinateField K C.1)
(RCN202.rationalVariable K) = coordinateEvaluation K C.1 ell)
(f : CoordinateField K C.1 →ₐ[RatFunc K] E) :
let mu := scalarPolynomialMap K E
let v := embeddingPoint C.1 (f.restrictScalars K)
MvPolynomial.eval v (mu F) = 0 ∧
MvPolynomial.eval v (sliceEquation (E := E) ell) = 0 ∧
MvPolynomial.aeval v (mu A) = 0 ∧
MvPolynomial.eval v (mu R) ≠ 0 ∧
IsolatedPoint (mu F) (sliceEquation (E := E) ell) (mu A) v := by
let C0 := gateOneComponent F A R C
have hj0 : algebraMap (RatFunc K) (CoordinateField K C0.1)
(RCN202.rationalVariable K) = movingValue C0.1 1 0 ell 1 := by
simpa only [movingValue, map_one, map_zero, mul_zero, zero_div, add_zero]
using hj
have h1 : (1 : Poly) ∉ C0.1 := regularComponent_H_not_mem K _ _ _ C0
have hcert := RCN202.embedding_point_certificate F 1 0 ell 1 0
(fun _ : Fin 1 => A) C0 hj0 h1 f
have hc :
MvPolynomial.eval (embeddingPoint C.1 (f.restrictScalars K))
(scalarPolynomialMap K E F) = 0 ∧
MvPolynomial.eval (embeddingPoint C.1 (f.restrictScalars K))
(sliceEquation (E := E) ell) = 0 ∧
MvPolynomial.aeval (embeddingPoint C.1 (f.restrictScalars K))
(scalarPolynomialMap K E A) = 0 ∧
MvPolynomial.eval (embeddingPoint C.1 (f.restrictScalars K)) (1 : PE) ≠ 0 ∧
IsolatedPoint (scalarPolynomialMap K E F) (sliceEquation (E := E) ell)
(scalarPolynomialMap K E A) (embeddingPoint C.1 (f.restrictScalars K)) := by
simpa only [map_one, map_zero,
movingEquation, eliminatedCut, filteredCut_zero, one_mul, mul_zero,
sub_zero, sliceEquation] using hcert
exact ⟨hc.1, hc.2.1, hc.2.2.1,
scalar_eval_ne_zero C.1 (f.restrictScalars K) R
(regularComponent_H_not_mem K F A R C), hc.2.2.2.2⟩
section WeightedEmbeddings
variable {I : Type} [Fintype I]
(P : I → Ideal (MvPolynomial (Fin 3) K)) [∀ i, (P i).IsPrime]
[∀ i, Algebra (RatFunc K) (CoordinateField K (P i))]
[∀ i, IsScalarTower K (RatFunc K) (CoordinateField K (P i))]
[∀ i, FiniteDimensional (RatFunc K) (CoordinateField K (P i))]
[∀ i, Algebra.IsSeparable (RatFunc K) (CoordinateField K (P i))]
/-- The left side is weighted embeddings, not merely a number of points.
Separability is indispensable for AlgHom.card = finrank. -/
theorem weighted_embedding_sum (weight : I → ℕ) :
(∑ z : (Σ i, CoordinateField K (P i) →ₐ[RatFunc K] E), weight z.1) =
∑ i, weight i * Module.finrank (RatFunc K) (CoordinateField K (P i)) := by
classical
rw [Fintype.sum_sigma]
apply Finset.sum_congr rfl
intro i _
simp only [Finset.sum_const, Finset.card_univ, smul_eq_mul, AlgHom.card, mul_comm]
end WeightedEmbeddings
end
end ProximityPrize.SubmissionLower.GenericSlicePoints6807
Source