Bilateral partial fractions for a quadratic q-series
OpenRamanujanLostNotebook.PartI.entry_12_2_1Let q be complex with 0<|q|<1, and let a be a nonzero complex number. Conventions: qPoch(x,b,n)=∏{k=0}^{n−1}(1−x b^k); qPochInf(x,b)=∏{k≥0}(1−x b^k). Assume ∀ n : ℤ, (1 + (a * (q ^ (2 * n)))) ≠ 0; ∀ n : ℕ, ((qPoch ((-a)) ((q ^ 2)) ((n + 1))) * (qPoch (((-(q ^ 2)) / a)) ((q ^ 2)) (n))) ≠ 0; (qPochInf ((q ^ 2)) ((q ^ 2))) ≠ 0. Put s0(n)=((q ^ ((2 * (n ^ 2)) + n)) / (1 + (a * (q ^ (2 * n))))) (n∈ℤ); s1(n)=(((((-1) ^ n) * (qPoch (q) ((q ^ 2)) (n))) * (q ^ (n ^ 2))) / ((qPoch ((-a)) ((q ^ 2)) ((n + 1))) * (qPoch (((-(q ^ 2)) / a)) ((q ^ 2)) (n)))) (n∈ℕ). Then every listed series is absolutely convergent and, writing Sj=∑s_j(n) over its stated index set, (((qPochInf (q) ((q ^ 2))) / (qPochInf ((q ^ 2)) ((q ^ 2)))) * S0) = S1. Differs from the printed source: we make the implicit open-disk domain and all denominator nonvanishing conditions explicit, and assert convergence in the conclusion rather than assume it. We use 0<|q|<1 because integer powers in bilateral denominators require q≠0; the possible removable q=0 limit is not asserted.
Discrepancy from the printed source. Differs from the printed source: we make the implicit open-disk domain and all denominator nonvanishing conditions explicit, and assert convergence in the conclusion rather than assume it. We use 0<|q|<1 because integer powers in bilateral denominators require q≠0; the possible removable q=0 limit is not asserted.
import Mathlib import Definitions.Def_RamanujanLostNotebook_PartI_shared_qPoch import Definitions.Def_RamanujanLostNotebook_PartI_shared_qPochInf
namespace RamanujanLostNotebook.PartI
theorem entry_12_2_1 (q : ℂ) (hq : ‖q‖ < 1) (hq0 : q ≠ 0) (a : ℂ) (ha : a ≠ 0)
(hd0 : ∀ n : ℤ, (1 + (a * (q ^ (2 * n)))) ≠ 0)
(hd1 : ∀ n : ℕ, ((qPoch ((-a)) ((q ^ 2)) ((n + 1))) * (qPoch (((-(q ^ 2)) / a)) ((q ^ 2)) (n))) ≠ 0)
(hd2 : (qPochInf ((q ^ 2)) ((q ^ 2))) ≠ 0) :
let s0 : ℤ → ℂ := fun n => ((q ^ ((2 * (n ^ 2)) + n)) / (1 + (a * (q ^ (2 * n)))))
let s1 : ℕ → ℂ := fun n => (((((-1 : ℂ) ^ n) * (qPoch (q) ((q ^ 2)) (n))) * (q ^ (n ^ 2))) / ((qPoch ((-a)) ((q ^ 2)) ((n + 1))) * (qPoch (((-(q ^ 2)) / a)) ((q ^ 2)) (n))))
Summable s0 ∧ Summable s1 ∧
(((qPochInf (q) ((q ^ 2))) / (qPochInf ((q ^ 2)) ((q ^ 2)))) * (∑' n : ℤ, s0 n)) = (∑' n : ℕ, s1 n) := by sorry
end RamanujanLostNotebook.PartI