Formal Jacobi identity over :
ProvedQseriesFormalization.Pending.JacobiCubeAnalyticToFormal.qPochInfPS_pow_three_eq_jacobiThetaPS_complexchan-invitation-qseriesformal-power-seriesq-pochhammertheta-functions
In ,
where is qPochInfPS ℂ, the inverse of the partition generating series, and the right-hand side is jacobiThetaPS ℂ. This is the formal power series version of Jacobi's identity. It is proved by showing both sides are Taylor series at of the same analytic function (the analytic Jacobi identity of Chapter 4), and uniqueness of power series expansions.
Preamble
import Definitions.Def_QseriesFormalization_DefCoreQSeries
import Mathlib.Combinatorics.Enumerative.Partition.GenFun
import Mathlib.RingTheory.PowerSeries.Basic
import Mathlib.RingTheory.PowerSeries.Expand
import Mathlib.RingTheory.PowerSeries.Inverse
import Mathlib.Analysis.SpecificLimits.Normed
import Mathlib.Analysis.Normed.Ring.InfiniteSum
import Mathlib
import Mathlib.Combinatorics.Enumerative.Partition.Basic
import Mathlib.Tactic.Linarith
import Mathlib.Analysis.SpecialFunctions.Log.Summable
import Mathlib.Analysis.Normed.Group.Tannery
/-!
# Pending: B2 via analytic → formal Taylor bridge
Independent attack route on the cube convolution identity. Avoids
Sylvester's combinatorial bijection by leveraging the already-proved
**analytic** Jacobi cube identity from Chapter 4:
`jacobiIdentity_pochhammer (q : ℂ) (hq : ‖q‖ < 1) :
(eulerPentagonalInfiniteProduct q)^3
= ∑' n : ℕ, ((-1)^n · (2n+1) · q^(n(n+1)/2))`
The plan: extract coefficient `n` from both sides via analytic Taylor
uniqueness, and identify both with the corresponding formal-PS coefficient.
## Roadmap
Let `f q := (eulerPentagonalInfiniteProduct q)^3` and
`g q := ∑' n, (-1)^n (2n+1) q^{n(n+1)/2}`.
Both are analytic on `Metric.ball 0 1`, and equal there by
`jacobiIdentity_pochhammer`.
1. **Polynomial truncation agreement (analytic side).** For `N > n`,
`f q ≡ ((∏_{k=1}^N (1 - q^k))^3 : polynomial) (mod q^{N+1})` near 0.
*Reason*: the tail factors `∏_{k>N}(1 - q^k)` have constant coefficient
1 and start at order > N, so they don't affect coefficients of degree
≤ N.
2. **Polynomial coefficient = formal-PS coefficient.** For
`P : ℂ[X]` and `n < deg(P)+1`, `P.coeff n` (polynomial) =
`(P : ℂ⟦X⟧).coeff n` (formal PS) = analytic Taylor coefficient of
`P.eval q` at `0`.
3. **Truncation = formal qPochInfPS truncation.** Already proved as
`coeff_qPochInfPS_pow_eq_coeff_finite_product_pow`.
4. **RHS Taylor extraction.** The Taylor coefficient at `0` of `g`
of order `n` equals `((jacobiTripleSign n : ℤ) : ℂ)` — by index
unpacking and dominated convergence on the geometric tail.
5. **Combine via analytic uniqueness.** `HasFPowerSeriesAt.eq_formal-
MultilinearSeries_of_eventually` (Mathlib) gives that the formal
multilinear series at 0 of `f` and `g` coincide, hence their
coefficient functions agree.
Substantial Mathlib infrastructure (`HasFPowerSeriesAt`,
`AnalyticAt.taylor`, `Polynomial.hasFPowerSeries`) is involved. Each
step below is a cleanly-stated sorry; closing them is multi-day Lean
work but follows standard analytic patterns.
Once all five are closed, the headline goal in
`Pending/Chapter19_B2_FromCubeConvolution.lean` follows in two lines
(rewrite + ext, matched against this file's bridge theorems).
-/
open QseriesFormalization
open QseriesFormalization.PartI.Ch04 (eulerPentagonalInfiniteProduct
)
open QseriesFormalization.PartIV.Ch19 (qPochInfPS jacobiThetaPS jacobiTripleSign
)
/-! ### Step 1: polynomial truncation agreement
For each `n`, there exists `N` such that
`fAnalytic q = ((∏_{k=0}^{N-1} (1 - q^(k+1)))^3) + (higher order tail)`
where the tail contributes 0 to the coefficient of `q^n` for `n < N`.
The right Lean machinery: `HasFPowerSeriesOnBall` + tail estimates.
-/
/- **Step 1.** The analytic Taylor coefficient at `0` of `fAnalytic` of
order `n` equals the polynomial coefficient at order `n` of the finite
truncation `(∏_{k=0}^{n}(1 - X^(k+1)))^3 ∈ ℂ[X]`.
Statement deferred until the Mathlib idiom is fixed (one of
`iteratedFDeriv`, `HasFPowerSeriesAt`, `Analytic.taylorCoeff`). A
placeholder `True` here would violate the playbook (trivially-true
RHS), so we leave the precise statement out of the file and document
it here as the next concrete dependency. -/
/-! ### Step 2: polynomial coefficient = formal PS coefficient.
This direction is essentially `Polynomial.coe_powerSeries` plus
`Polynomial.coeff_coe` or `PowerSeries.coeff_polynomial`. Mathlib has
the natural ring hom `ℂ[X] →+* ℂ⟦X⟧` (`Polynomial.toPowerSeries`) and
it preserves coefficients.
-/
/-! ### Step 3: truncation = formal qPochInfPS truncation.
Already proved in `Chapter19.coeff_qPochInfPS_pow_eq_coeff_finite_product_pow`.
-/
/-! ### Step 4: RHS Taylor extraction. **CLOSED.**
The reindex `∑' k, jacobiTripleSign k · q^k = ∑' n, (-1)^n (2n+1) q^{T_n}`
is now proved unconditionally in `Chapter19_JacobiTripleSignChar` as
`tsum_jacobiTripleSign_eq_jacobi_series`. -/
/-! ### Step 5: combine via analytic uniqueness.
`Mathlib.Analysis.Analytic.Uniqueness.HasFPowerSeriesAt.eq_formalMultilinearSeries_of_eventually`
gives: if `f` and `g` are analytic at `0` with formal multilinear series
`p` and `q` respectively, and `f = g` eventually near `0`, then `p = q`.
Applied here: `fAnalytic` has formal series `ofScalars ℂ (jts ·)` on
`ball 0 1` (the `hasFPowerSeriesOnBall_fAnalytic` below); the formal-PS
`(qPochInfPS ℂ) ^ 3` is also an analytic-PS at 0 (via its coefficients
matching a polynomial truncation up to high order). By the uniqueness
of formal multilinear series for an analytic function, the coefficient
sequences coincide. -/Formal statement
theorem QseriesFormalization.Pending.JacobiCubeAnalyticToFormal.qPochInfPS_pow_three_eq_jacobiThetaPS_complex :
(qPochInfPS ℂ) ^ 3 = jacobiThetaPS ℂ := by sorrySource
Lean formalization (Xiang Huang) of Hei-Chi Chan, An Invitation to q-Series: From Jacobi's Triple Product Identity to Ramanujan's "Most Beautiful Identity" (World Scientific, 2011). Lean source: https://github.com/xiangyazi24/invitation-to-qseries-lean/blob/5df34b280b6a9c4a68ab09105b4e5d3c7bbe88d1/QseriesFormalization/Pending/JacobiCubeAnalyticToFormal.lean#L282-L335