Sparse sparse line-anchor GCD certificates with degree-sensitive cardinality bound
ProvedWeierstrassEllipticZeta.line_anchor_gcd_compressionLet Q have homogeneous bidegree (m,n) in the existing elliptic-extension coordinates. At every fixed coordinate b and slope alpha, the selected slice indices J lie in {0,...,m+n}, their GCD equals the existing line-anchor GCD G exactly, and
|J| + natDegree(G) <= n+1.
Thus |J|<=n+1 always and |J|<=n when G has positive degree, the condition for a viable line. The selected family also has minimum cardinality among all subsets of the canonical sample indices giving this same GCD. The zero-obstruction convention is included, with J empty and G zero. No positivity assumptions on m,n or analytic assumptions on S are needed.
For every predicate on the underlying finite locus candidate, existence of an old GCD-tested anchor satisfying it is equivalent to existence of a compressed-GCD anchor satisfying it. The equivalence preserves the exact candidate, coordinate b, period-pair index and point/fibre branches.
Previously the defining GCD used m+n+1 slices. It can now be represented exactly with at most n+1 slices, or at most n for a positive-degree GCD; more precisely at most n+1-deg(G). This is a bound on the size of a selected subfamily, not a claim that the first n samples suffice. The largest possible sample index remains m+n. Finding a minimum subfamily may still require the full family. The degree-n bound, integer-search bound and global geometric cost estimate are unchanged.
import Mathlib.Algebra.GCDMonoid.Finset import Mathlib.RingTheory.Polynomial.Content import Mathlib.Algebra.Polynomial.Div import Definitions.Def_WeierstrassEllipticZeta_SparseLineAnchorGCD import Mathlib.Tactic noncomputable section open scoped Classical open WeierstrassEllipticZeta
theorem WeierstrassEllipticZeta.line_anchor_gcd_compression
(S : Fin 5 → ℂ → ℂ) (Q : MvPolynomial (Fin 7) ℂ) (m n : ℕ)
(hQ : ∀ d ∈ Q.support, d 0 + d 1 = m ∧
d 2 + d 3 + d 4 + d 5 + d 6 = n) :
(∀ b α : ℂ,
lineAnchorGCDCore S Q m n b α ⊆ Finset.range (m + n + 1) ∧
sparseLineAnchorGCD S Q m n b α = lineAnchorGCD S Q m n b α ∧
(lineAnchorGCDCore S Q m n b α).card +
(lineAnchorGCD S Q m n b α).natDegree ≤ n + 1 ∧
(0 < (lineAnchorGCD S Q m n b α).natDegree →
(lineAnchorGCDCore S Q m n b α).card ≤ n) ∧
(∀ J : Finset ℕ, J ⊆ Finset.range (m + n + 1) →
J.gcd (fun j : ℕ => anchorSlice S Q b α j) = lineAnchorGCD S Q m n b α →
(lineAnchorGCDCore S Q m n b α).card ≤ J.card)) ∧
∀ (Λ : Submodule ℤ ℂ) (η : Λ →ₗ[ℤ] ℂ) (X : Finset ℂ)
(K : Set ℂ) (Z : Finset ℂ) (P : FiniteLocusCandidate Λ X → Prop),
(∃ a : GCDAnchorCandidate Λ η X S Q m n K Z,
P (gcdAnchorLocus Λ η X S Q m n K Z a)) ↔
(∃ a : SparseGCDAnchorCandidate Λ η X S Q m n K Z,
P (sparseGCDAnchorLocus Λ η X S Q m n K Z a)) := by sorry