Documentation

Signals.CanonicalForms

Affine interval and planar canonical-form examples in fixed oriented coordinates.

Forms are position-dependent alternating covectors, not just reciprocal scalars. Residues are punctured-neighborhood limits of regularized coefficients. These examples do not assert a general projective canonical-form existence theorem or an optical evaluation operation.

noncomputable def Signals.CanonicalForms.intervalCoefficient (coordinate : ℝ) :
ℝ

Coefficient of dt on the oriented interval (0,1).

Instances For
    noncomputable def Signals.CanonicalForms.intervalForm (coordinate : ℝ) :
    (Fin 1 → ℝ) [⋀^Fin 1]→ₗ[ℝ] ℝ

    The interval one-form acts on a one-dimensional tangent vector by its coefficient.

    Instances For
      theorem Signals.CanonicalForms.intervalForm_apply (coordinate : ℝ) (vectors : Matrix (Fin 1) (Fin 1) ℝ) :
      (intervalForm coordinate) vectors = intervalCoefficient coordinate * vectors 0 0

      The alternating one-form has the stated dt coefficient.

      theorem Signals.CanonicalForms.intervalCoefficient_normalization (coordinate : ℝ) (lower : 0 < coordinate) (upper : coordinate < 1) :
      intervalCoefficient coordinate * (coordinate * (1 - coordinate)) = 1

      The interval coefficient is normalized away from its two boundary points.

      theorem Signals.CanonicalForms.residue_limit {coefficient regularized : ℝ → ℝ} {boundary : ℝ} (continuous : ContinuousAt regularized boundary) (regularization : ∀ (coordinate : ℝ), coordinate ≠ boundary → (coordinate - boundary) * coefficient coordinate = regularized coordinate) :
      Filter.Tendsto (fun (coordinate : ℝ) => (coordinate - boundary) * coefficient coordinate) (nhdsWithin boundary {boundary}ᶜ) (nhds (regularized boundary))

      A continuous regularization determines the simple-pole residue as a genuine limit.

      theorem Signals.CanonicalForms.interval_left_regularization (coordinate : ℝ) (distinct : coordinate ≠ 0) :
      coordinate * intervalCoefficient coordinate = 1 / (1 - coordinate)

      Multiplying by the left boundary coordinate removes its pole, including at the other boundary.

      theorem Signals.CanonicalForms.interval_residue_zero :
      Filter.Tendsto (fun (coordinate : ℝ) => coordinate * intervalCoefficient coordinate) (nhdsWithin 0 {0}ᶜ) (nhds 1)

      The residue at the left endpoint is +1 in the dt convention.

      theorem Signals.CanonicalForms.interval_right_regularization (coordinate : ℝ) (distinct : coordinate ≠ 1) :
      (coordinate - 1) * intervalCoefficient coordinate = -1 / coordinate

      Multiplying by t-1 removes the right pole with the oriented negative residue.

      theorem Signals.CanonicalForms.interval_residue_one :
      Filter.Tendsto (fun (coordinate : ℝ) => (coordinate - 1) * intervalCoefficient coordinate) (nhdsWithin 1 {1}ᶜ) (nhds (-1))

      The residue at the right endpoint is -1 in the dt convention.

      theorem Signals.CanonicalForms.intervalCoefficient_continuousAt (coordinate : ℝ) (left : coordinate ≠ 0) (right : coordinate ≠ 1) :
      ContinuousAt intervalCoefficient coordinate

      The interval coefficient has no other affine pole.

      @[reducible, inline]

      Two real affine coordinates and their tangent-vector space.

      Instances For
        noncomputable def Signals.CanonicalForms.triangleCoefficient (point : Plane) :
        ℝ

        Coefficient of dx wedge dy on x>0, y>0, x+y<1.

        Instances For
          noncomputable def Signals.CanonicalForms.triangleForm (point : Plane) :
          Plane [⋀^Fin 2]→ₗ[ℝ] ℝ

          A genuine alternating two-form in the standard oriented affine coordinates.

          Instances For
            theorem Signals.CanonicalForms.triangleForm_apply (point : Plane) (vectors : Matrix (Fin 2) (Fin 2) ℝ) :
            (triangleForm point) vectors = triangleCoefficient point * (vectors 0 0 * vectors 1 1 - vectors 0 1 * vectors 1 0)

            Evaluation on two tangent vectors is the coefficient times their oriented area.

            theorem Signals.CanonicalForms.triangleCoefficient_normalization (point : Plane) (first : point 0 ≠ 0) (second : point 1 ≠ 0) (diagonal : 1 - point 0 - point 1 ≠ 0) :
            triangleCoefficient point * (point 0 * point 1 * (1 - point 0 - point 1)) = 1

            The triangle coefficient has unit numerator on its regular domain.

            theorem Signals.CanonicalForms.triangleCoefficient_continuousAt (point : Plane) (first : point 0 ≠ 0) (second : point 1 ≠ 0) (diagonal : 1 - point 0 - point 1 ≠ 0) :
            ContinuousAt triangleCoefficient point

            The triangle coefficient has no affine singularity away from its three boundary lines.

            theorem Signals.CanonicalForms.cancel_pole_factor (coordinate remaining : ℝ) (distinct : coordinate ≠ 0) :
            coordinate * (1 / (coordinate * remaining)) = 1 / remaining

            Cancel a simple real pole without requiring the remaining factor to be nonzero.

            theorem Signals.CanonicalForms.triangle_x_regularization (coordinate normal : ℝ) (distinct : normal ≠ 0) :
            normal * (triangleForm ![normal, coordinate]) !![1, 0; 0, 1] = 1 / (coordinate * (1 - normal - coordinate))

            Contracting dx wedge dy with the x-normal leaves the positive oriented dy residue.

            theorem Signals.CanonicalForms.triangle_y_regularization (coordinate normal : ℝ) (distinct : normal ≠ 0) :
            normal * (triangleForm ![coordinate, normal]) !![0, 1; 1, 0] = -1 / (coordinate * (1 - coordinate - normal))

            Contracting with the y-normal gives -dx, fixing the second boundary's orientation.

            theorem Signals.CanonicalForms.triangle_diagonal_regularization (coordinate normal : ℝ) (distinct : normal ≠ 0) :
            normal * (triangleForm ![coordinate, 1 - coordinate - normal]) !![0, -1; 1, -1] = 1 / (coordinate * (1 - coordinate - normal))

            The diagonal normal z=1-x-y and tangent x produce the positive dx orientation.

            theorem Signals.CanonicalForms.triangle_residue_x (coordinate : ℝ) (lower : 0 < coordinate) (upper : coordinate < 1) :
            Filter.Tendsto (fun (normal : ℝ) => normal * (triangleForm ![normal, coordinate]) !![1, 0; 0, 1]) (nhdsWithin 0 {0}ᶜ) (nhds (intervalCoefficient coordinate))

            The x-boundary residue is the interval form's coefficient on the surviving coordinate.

            theorem Signals.CanonicalForms.triangle_residue_y (coordinate : ℝ) (lower : 0 < coordinate) (upper : coordinate < 1) :
            Filter.Tendsto (fun (normal : ℝ) => normal * (triangleForm ![coordinate, normal]) !![0, 1; 1, 0]) (nhdsWithin 0 {0}ᶜ) (nhds (-intervalCoefficient coordinate))

            The y-boundary residue is the negative interval coefficient in the dx convention.

            theorem Signals.CanonicalForms.triangle_residue_diagonal (coordinate : ℝ) (lower : 0 < coordinate) (upper : coordinate < 1) :
            Filter.Tendsto (fun (normal : ℝ) => normal * (triangleForm ![coordinate, 1 - coordinate - normal]) !![0, -1; 1, -1]) (nhdsWithin 0 {0}ᶜ) (nhds (intervalCoefficient coordinate))

            The diagonal-boundary residue is the positive interval coefficient in the chosen z,x coordinates.

            noncomputable def Signals.CanonicalForms.squareCoefficient (point : Plane) :
            ℝ

            The unit square coefficient, whose only affine poles are its four boundary lines.

            Instances For
              noncomputable def Signals.CanonicalForms.squareForm (point : Plane) :
              Plane [⋀^Fin 2]→ₗ[ℝ] ℝ

              The square two-form in the same orientation as both constituent triangles.

              Instances For

                Affine pullback coefficient for the lower triangle, with unit positive Jacobian.

                Instances For

                  Affine pullback coefficient for the upper triangle, with unit positive Jacobian.

                  Instances For
                    theorem Signals.CanonicalForms.lowerTriangle_pullback (point : Plane) (vectors : Matrix (Fin 2) (Fin 2) ℝ) :
                    (triangleForm ![point 0 - point 1, point 1]) (vectors * !![1, 0; -1, 1]) = lowerTriangleCoefficient point * vectors.det

                    Pulling the standard triangle back by (x,y) -> (x-y,y) preserves oriented area.

                    theorem Signals.CanonicalForms.upperTriangle_pullback (point : Plane) (vectors : Matrix (Fin 2) (Fin 2) ℝ) :
                    (triangleForm ![point 0, point 1 - point 0]) (vectors * !![1, -1; 0, 1]) = upperTriangleCoefficient point * vectors.det

                    Pulling the standard triangle back by (x,y) -> (x,y-x) also preserves oriented area.

                    theorem Signals.CanonicalForms.square_triangulation (point : Plane) (first : point 0 ≠ 0) (second : point 1 ≠ 0) (firstUpper : point 0 ≠ 1) (secondUpper : point 1 ≠ 1) (diagonal : point 0 ≠ point 1) :

                    The two rational triangle extensions cancel their shared diagonal pole.

                    theorem Signals.CanonicalForms.squareForm_triangulation (point : Plane) (first : point 0 ≠ 0) (second : point 1 ≠ 0) (firstUpper : point 0 ≠ 1) (secondUpper : point 1 ≠ 1) (diagonal : point 0 ≠ point 1) :
                    lowerTriangleCoefficient point • Matrix.detRowAlternating + upperTriangleCoefficient point • Matrix.detRowAlternating = squareForm point

                    Triangulation equality holds at the level of alternating forms, not only coefficients.

                    theorem Signals.CanonicalForms.squareCoefficient_continuousAt (point : Plane) (first : point 0 ≠ 0) (second : point 1 ≠ 0) (firstUpper : point 0 ≠ 1) (secondUpper : point 1 ≠ 1) :
                    ContinuousAt squareCoefficient point

                    The summed square continuation is continuous even on the former interior diagonal.

                    The square is the product of interval coefficients in the fixed affine chart.

                    theorem Signals.CanonicalForms.square_residue_x_zero (coordinate : ℝ) :
                    Filter.Tendsto (fun (normal : ℝ) => normal * squareCoefficient ![normal, coordinate]) (nhdsWithin 0 {0}ᶜ) (nhds (intervalCoefficient coordinate))

                    At x=0 the oriented coefficient residue is the interval coefficient in y.

                    theorem Signals.CanonicalForms.square_residue_x_one (coordinate : ℝ) :
                    Filter.Tendsto (fun (normal : ℝ) => (normal - 1) * squareCoefficient ![normal, coordinate]) (nhdsWithin 1 {1}ᶜ) (nhds (-intervalCoefficient coordinate))

                    At x=1 the residue has the negative interval coefficient in y.

                    theorem Signals.CanonicalForms.square_residue_y_zero (coordinate : ℝ) :
                    Filter.Tendsto (fun (normal : ℝ) => normal * -squareCoefficient ![coordinate, normal]) (nhdsWithin 0 {0}ᶜ) (nhds (-intervalCoefficient coordinate))

                    Contracting with the y-normal gives a negative residue at y=0.

                    theorem Signals.CanonicalForms.square_residue_y_one (coordinate : ℝ) :
                    Filter.Tendsto (fun (normal : ℝ) => (normal - 1) * -squareCoefficient ![coordinate, normal]) (nhdsWithin 1 {1}ᶜ) (nhds (intervalCoefficient coordinate))

                    The y-normal orientation gives a positive residue at y=1.