import OAI.NumberTheory.DukePrimeDegree.Geometry.ControlledDistanceCuts import OAI.NumberTheory.DukePrimeDegree.Entropy.ContractingCutItineraries namespace OAI /-! # Finite observations from controlled distance cuts The observation records the membership bits of the selected distance balls. Equality of observations gives small relative lifts, while the dyadic estimates imply almost-sure avoidance of all finite cut boundaries. -/ noncomputable section open Set MeasureTheory Filter open scoped Topology ENNReal namespace DukePrimeDegree open PrimeDegreePackets def distanceCutObservation {n N : ℕ} (z : Fin N → X n) (c : Fin N → ℝ) (x : X n) (i : Fin N) : Bool := decide (quotientLogDistance (z i) x < c i) lemma distanceCutObservation_measurable {n N : ℕ} (z : Fin N → X n) (c : Fin N → ℝ) : Measurable (distanceCutObservation z c) := by apply measurable_pi_iff.mpr intro i apply measurable_to_bool have he : (fun x => distanceCutObservation z c x i) ⁻¹' {true} = {x | quotientLogDistance (z i) x < c i} := by ext x simp [distanceCutObservation] rw [he] exact measurableSet_lt (continuous_quotientLogDistance n _).measurable measurable_const lemma distanceCutObservation_small_lifts {n N : ℕ} (z : Fin N → X n) (c : Fin N → ℝ) {r : ℝ} (hsmall : ∀ i, ∀ x, quotientLogDistance (z i) x < c i → ∀ y, quotientLogDistance (z i) y < c i → ∃ g ∈ matrixBall n r, rightAction x g=y) {x y : X n} (hx : ∃ i, quotientLogDistance (z i) x < c i) (hxy : distanceCutObservation z c y=distanceCutObservation z c x) : ∃ g ∈ matrixBall n r, rightAction x g=y := by obtain ⟨i,hi⟩ := hx apply hsmall i x hi y have H := congrFun hxy i simpa only [distanceCutObservation,hi,decide_true,decide_eq_true_eq] using H theorem ae_eventually_avoid_finite_dyadic_boundaries {Y : Type*} [MeasurableSpace Y] {N : ℕ} (μ : Measure Y) {T : Y → Y} (hT : MeasurePreserving T μ μ) (f : Fin N → Y → ℝ) (hf : ∀ i, Measurable (f i)) (c C : Fin N → ℝ) (hbound : ∀ i k, μ (realBoundaryStrip (f i) (c i) ((1/4:ℝ)^k)) ≤ ENNReal.ofReal (C i*(1/2:ℝ)^k)) : ∀ᵐ x ∂μ, ∀ᶠ k : ℕ in atTop, ∀ i, (1/4:ℝ)^k < |f i (T^[k] x)-c i| := by have hall (i : Fin N) : ∀ᵐ x ∂μ, ∀ᶠ k : ℕ in atTop, (1/4:ℝ)^k < |f i (T^[k] x)-c i| := by apply ae_eventually_avoid_realBoundaryStrip μ hT (hf i) have hs := ENNReal.tsum_le_tsum (hbound i) exact (lt_of_le_of_lt hs (summable_geometric_two.mul_left (C i)).tsum_ofReal_lt_top).ne filter_upwards [ae_all_iff.mpr hall] with x hx exact Filter.eventually_all.mpr hx end DukePrimeDegree end end OAI