coordinate-system 5.2.1__cp313-cp313-win_amd64.whl → 6.0.0__cp313-cp313-win_amd64.whl

coordinate_system/differential_geometry.py

@@ -1,33 +1,94 @@
 """
-Differential Geometry Module for Coordinate System Package
-===========================================================
+Differential Geometry Module
+============================
 
-This module provides tools for discrete differential geometry computations on surfaces,
-using the CORRECT Intrinsic Gradient Operator framework based on the proven algorithm.
+Comprehensive discrete differential geometry computations on surfaces,
+combining classical methods and intrinsic gradient operator framework.
 
-Key Formula:
-    G_μ = (c(u+h) - c(u-h)) / (2h) / c(u) then extract normal derivative using .VZ()
+Key Features:
+- Surface parametrization (Sphere, Torus, custom surfaces)
+- Metric tensor computation
+- Intrinsic gradient operator for curvature calculation
+- Classical curvature methods with high-order finite differences
+- Gaussian, mean, principal, and Riemann curvatures
 
-Author: PanGuoJun
-Date: 2025-10-31
+Author: Coordinate System Package
+Date: 2025-12-03
 """
 
 import math
 import numpy as np
-from typing import Tuple, Optional, Callable, Union
+from typing import Tuple, Optional, Callable, Union, Dict, List
 from .coordinate_system import coord3, vec3
 
 
-# ========== Surface Base Class ==========
+# ============================================================
+# High-Order Finite Difference Operators
+# ============================================================
+
+def derivative_5pt(f: Callable[[float], np.ndarray], x: float, h: float) -> np.ndarray:
+    """
+    5-point finite difference formula for first derivative.
+
+    Accuracy: O(h^4)
+
+    Args:
+        f: Function to differentiate
+        x: Point at which to compute derivative
+        h: Step size
+
+    Returns:
+        First derivative approximation
+    """
+    return (-f(x + 2*h) + 8*f(x + h) - 8*f(x - h) + f(x - 2*h)) / (12*h)
+
+
+def derivative_2nd_5pt(f: Callable[[float], np.ndarray], x: float, h: float) -> np.ndarray:
+    """
+    5-point finite difference formula for second derivative.
+
+    Accuracy: O(h^4)
+
+    Args:
+        f: Function to differentiate
+        x: Point at which to compute derivative
+        h: Step size
+
+    Returns:
+        Second derivative approximation
+    """
+    return (-f(x + 2*h) + 16*f(x + h) - 30*f(x) + 16*f(x - h) - f(x - 2*h)) / (12*h*h)
+
+
+def richardson_extrapolation(f_h: float, f_2h: float, order: int = 4) -> float:
+    """
+    Richardson extrapolation for accelerating convergence.
+
+    Args:
+        f_h: Value computed with step size h
+        f_2h: Value computed with step size 2h
+        order: Order of the method
+
+    Returns:
+        Extrapolated value with improved accuracy
+    """
+    return (2**order * f_h - f_2h) / (2**order - 1)
+
+
+# ============================================================
+# Surface Base Class and Common Surfaces
+# ============================================================
 
 class Surface:
     """
-    Base class for parametric surfaces r(u, v)
+    Base class for parametric surfaces r(u, v).
+
+    Subclasses must implement the position(u, v) method.
     """
 
     def __init__(self, h: float = 1e-6):
         """
-        Initialize surface
+        Initialize surface.
 
         Args:
             h: Step size for numerical differentiation
@@ -35,23 +96,27 @@ class Surface:
         self.h = h
 
     def position(self, u: float, v: float) -> vec3:
-        """Compute position on surface at parameters (u, v)"""
+        """
+        Compute position on surface at parameters (u, v).
+
+        Must be implemented by subclasses.
+        """
         raise NotImplementedError("Subclass must implement position(u, v)")
 
     def tangent_u(self, u: float, v: float) -> vec3:
-        """Compute tangent vector in u direction"""
+        """Compute tangent vector in u direction using central difference."""
         r_plus = self.position(u + self.h, v)
         r_minus = self.position(u - self.h, v)
         return (r_plus - r_minus) * (1.0 / (2.0 * self.h))
 
     def tangent_v(self, u: float, v: float) -> vec3:
-        """Compute tangent vector in v direction"""
+        """Compute tangent vector in v direction using central difference."""
         r_plus = self.position(u, v + self.h)
         r_minus = self.position(u, v - self.h)
         return (r_plus - r_minus) * (1.0 / (2.0 * self.h))
 
     def normal(self, u: float, v: float) -> vec3:
-        """Compute unit normal vector"""
+        """Compute unit normal vector."""
         r_u = self.tangent_u(u, v)
         r_v = self.tangent_v(u, v)
         n = r_u.cross(r_v)
@@ -62,13 +127,16 @@ class Surface:
             return vec3(0.0, 0.0, 1.0)
 
 
-# ========== Common Surface Types ==========
-
 class Sphere(Surface):
     """
-    Sphere surface
-    Parametrization: r(θ, φ) = R(sin θ cos φ, sin θ sin φ, cos θ)
-    where θ ∈ [0, π] is polar angle, φ ∈ [0, 2π] is azimuthal angle
+    Sphere surface.
+
+    Parametrization: r(theta, phi) = R(sin(theta)cos(phi), sin(theta)sin(phi), cos(theta))
+    where theta in [0, pi] is polar angle, phi in [0, 2*pi] is azimuthal angle.
+
+    Theoretical curvatures:
+    - Gaussian curvature: K = 1/R^2
+    - Mean curvature: H = 1/R
     """
 
     def __init__(self, radius: float = 1.0, h: float = 1e-6):
@@ -76,43 +144,75 @@ class Sphere(Surface):
         self.R = radius
 
     def position(self, theta: float, phi: float) -> vec3:
-        """Position on sphere"""
+        """Position on sphere."""
         x = self.R * math.sin(theta) * math.cos(phi)
         y = self.R * math.sin(theta) * math.sin(phi)
         z = self.R * math.cos(theta)
         return vec3(x, y, z)
 
+    @property
+    def theoretical_gaussian_curvature(self) -> float:
+        """Theoretical Gaussian curvature K = 1/R^2."""
+        return 1.0 / (self.R * self.R)
+
+    @property
+    def theoretical_mean_curvature(self) -> float:
+        """Theoretical mean curvature H = 1/R."""
+        return 1.0 / self.R
+
 
 class Torus(Surface):
     """
-    Torus surface
-    Parametrization: r(u, v) = ((R + r cos u) cos v, (R + r cos u) sin v, r sin u)
+    Torus surface.
+
+    Parametrization: r(u, v) = ((R + r*cos(u))*cos(v), (R + r*cos(u))*sin(v), r*sin(u))
+    where R is major radius, r is minor radius.
+
+    Theoretical curvatures:
+    - Gaussian curvature: K = cos(u) / (r * (R + r*cos(u)))
+    - Mean curvature: H = (R + 2*r*cos(u)) / (2*r*(R + r*cos(u)))
     """
 
     def __init__(self, major_radius: float = 3.0, minor_radius: float = 1.0, h: float = 1e-6):
         super().__init__(h)
-        self.R = major_radius  # Major radius
-        self.r = minor_radius  # Minor radius
+        self.R = major_radius
+        self.r = minor_radius
 
     def position(self, u: float, v: float) -> vec3:
-        """Position on torus"""
+        """Position on torus."""
         x = (self.R + self.r * math.cos(u)) * math.cos(v)
         y = (self.R + self.r * math.cos(u)) * math.sin(v)
         z = self.r * math.sin(u)
         return vec3(x, y, z)
 
+    def theoretical_gaussian_curvature(self, u: float) -> float:
+        """Theoretical Gaussian curvature at parameter u."""
+        return math.cos(u) / (self.r * (self.R + self.r * math.cos(u)))
+
+    def theoretical_mean_curvature(self, u: float) -> float:
+        """Theoretical mean curvature at parameter u."""
+        return (self.R + 2 * self.r * math.cos(u)) / (2 * self.r * (self.R + self.r * math.cos(u)))
 
-# ========== Metric Tensor ==========
+
+# ============================================================
+# Metric Tensor
+# ============================================================
 
 class MetricTensor:
     """
-    First fundamental form (metric tensor) of a surface
-    g_ij = <∂r/∂u^i, ∂r/∂u^j>
+    First fundamental form (metric tensor) of a surface.
+
+    g_ij = <dr/du^i, dr/du^j>
+
+    Components:
+    - E = g_11 = <r_u, r_u>
+    - F = g_12 = <r_u, r_v>
+    - G = g_22 = <r_v, r_v>
     """
 
     def __init__(self, E: float, F: float, G: float):
         """
-        Initialize metric tensor
+        Initialize metric tensor.
 
         Args:
             E: g_11 = <r_u, r_u>
@@ -126,7 +226,7 @@ class MetricTensor:
 
     @classmethod
     def from_surface(cls, surface: Surface, u: float, v: float) -> 'MetricTensor':
-        """Create metric tensor from surface at point (u, v)"""
+        """Create metric tensor from surface at point (u, v)."""
         r_u = surface.tangent_u(u, v)
         r_v = surface.tangent_v(u, v)
         E = r_u.dot(r_u)
@@ -135,171 +235,193 @@ class MetricTensor:
         return cls(E, F, G)
 
     def determinant(self) -> float:
-        """Get determinant of metric tensor"""
+        """Get determinant of metric tensor."""
         return self.det
 
+    def inverse(self) -> np.ndarray:
+        """Get inverse metric tensor as numpy array."""
+        if abs(self.det) < 1e-14:
+            return np.eye(2)
+        return np.array([[self.G, -self.F], [-self.F, self.E]]) / self.det
+
+    def as_matrix(self) -> np.ndarray:
+        """Get metric tensor as numpy array."""
+        return np.array([[self.E, self.F], [self.F, self.G]])
+
     def __repr__(self) -> str:
         return f"MetricTensor(E={self.E:.6f}, F={self.F:.6f}, G={self.G:.6f}, det={self.det:.6f})"
 
 
-# ========== Intrinsic Gradient Operator ==========
+# ============================================================
+# Gradient Result
+# ============================================================
+
+class GradientResult:
+    """
+    Gradient result containing normal vector derivative.
+    """
+
+    def __init__(self, dn: vec3, direction: str):
+        """
+        Initialize gradient result.
+
+        Args:
+            dn: Normal vector derivative
+            direction: Parameter direction ('u' or 'v')
+        """
+        self.dn = dn
+        self.direction = direction
+
+    def __repr__(self) -> str:
+        return f"GradientResult({self.direction}: [{self.dn.x:.6f}, {self.dn.y:.6f}, {self.dn.z:.6f}])"
+
+
+# ============================================================
+# Intrinsic Gradient Operator
+# ============================================================
 
 class IntrinsicGradientOperator:
     """
-    CORRECT implementation based on proven test code
+    Intrinsic gradient operator implementation.
+
+    Key formula:
+        G_mu = (c(u+h) - c(u-h)) / (2h) / c(u)
+    Then extract normal derivative using .VZ()
     """
-    
+
     def __init__(self, surface: Surface, step_size: float = 1e-3):
         self.surface = surface
         self.h = step_size
 
     def calc_intrinsic_frame(self, u: float, v: float) -> coord3:
         """
-        Calculate intrinsic frame - EXACTLY like the working test code
+        Calculate intrinsic frame at point (u, v).
+
+        For Sphere and Torus, uses analytical expressions.
+        For other surfaces, uses numerical derivatives.
         """
-        # 对于球面，使用解析表达式（与测试代码一致）
         if isinstance(self.surface, Sphere):
             R = self.surface.R
             theta, phi = u, v
-            
-            # 位置向量
+
             pos = self.surface.position(theta, phi)
-            
-            # 切向量（解析表达式，与测试代码一致）
+
+            # Analytical tangent vectors
             r_theta = vec3(
                 R * math.cos(theta) * math.cos(phi),
                 R * math.cos(theta) * math.sin(phi),
                 -R * math.sin(theta)
             )
-            
             r_phi = vec3(
                 -R * math.sin(theta) * math.sin(phi),
                 R * math.sin(theta) * math.cos(phi),
                 0
             )
-            
-            # 单位法向量
+
             n = r_theta.cross(r_phi).normalized()
-            
-            # 单位切向量
             e1 = r_theta.normalized()
             e2 = r_phi.normalized()
-            
+
         elif isinstance(self.surface, Torus):
-            # 对于环面，使用解析表达式
-            R = self.surface.R  # 主半径
-            r = self.surface.r  # 副半径
+            R = self.surface.R
+            r = self.surface.r
             u_param, v_param = u, v
-            
-            # 位置向量
+
             pos = self.surface.position(u_param, v_param)
-            
-            # 切向量（解析表达式）
+
+            # Analytical tangent vectors
             r_u = vec3(
                 -r * math.sin(u_param) * math.cos(v_param),
                 -r * math.sin(u_param) * math.sin(v_param),
                 r * math.cos(u_param)
             )
-            
             r_v = vec3(
                 -(R + r * math.cos(u_param)) * math.sin(v_param),
                 (R + r * math.cos(u_param)) * math.cos(v_param),
                 0
             )
-            
-            # 单位法向量
+
             n = r_u.cross(r_v).normalized()
-            
-            # 单位切向量
             e1 = r_u.normalized()
             e2 = r_v.normalized()
-            
+
         else:
-            # 对于其他曲面，使用数值方法
+            # Numerical method for general surfaces
             pos = self.surface.position(u, v)
             r_u = self.surface.tangent_u(u, v)
             r_v = self.surface.tangent_v(u, v)
-            
-            # 单位法向量
+
             n = r_u.cross(r_v).normalized()
-            
-            # 单位切向量
             e1 = r_u.normalized()
             e2 = r_v.normalized()
-        
-        # 创建内禀标架（与测试代码一致）
+
+        # Create intrinsic frame
         frame = coord3()
         frame.o = pos
         frame.ux = e1
         frame.uy = e2
         frame.uz = n
-        
+
         return frame
 
-    def compute_both(self, u: float, v: float) -> Tuple['GradientResult', 'GradientResult', coord3]:
+    def compute_both(self, u: float, v: float) -> Tuple[GradientResult, GradientResult, coord3]:
         """
-        Compute gradients using CENTRAL DIFFERENCES - like test code
+        Compute gradients in both u and v directions using central differences.
+
+        Returns:
+            Tuple of (G_u, G_v, center_frame)
         """
-        # 计算中心点和偏移点的标架
         c_center = self.calc_intrinsic_frame(u, v)
-        
-        # u方向：中心差分
+
+        # u direction: central difference
         c_u_plus = self.calc_intrinsic_frame(u + self.h, v)
         c_u_minus = self.calc_intrinsic_frame(u - self.h, v)
         dn_du = ((c_u_plus - c_u_minus) / (2 * self.h)).VZ()
-        
-        # v方向：中心差分  
+
+        # v direction: central difference
         c_v_plus = self.calc_intrinsic_frame(u, v + self.h)
         c_v_minus = self.calc_intrinsic_frame(u, v - self.h)
         dn_dv = ((c_v_plus - c_v_minus) / (2 * self.h)).VZ()
-        
-        # 创建梯度结果
+
         G_u = GradientResult(dn_du, "u")
         G_v = GradientResult(dn_dv, "v")
-        
+
         return G_u, G_v, c_center
 
+    def compute_u(self, u: float, v: float) -> GradientResult:
+        """Compute gradient in u direction only."""
+        c_plus = self.calc_intrinsic_frame(u + self.h, v)
+        c_minus = self.calc_intrinsic_frame(u - self.h, v)
+        dn = ((c_plus - c_minus) / (2 * self.h)).VZ()
+        return GradientResult(dn, "u")
 
-class GradientResult:
-    """
-    Gradient result containing normal vector derivative
-    """
+    def compute_v(self, u: float, v: float) -> GradientResult:
+        """Compute gradient in v direction only."""
+        c_plus = self.calc_intrinsic_frame(u, v + self.h)
+        c_minus = self.calc_intrinsic_frame(u, v - self.h)
+        dn = ((c_plus - c_minus) / (2 * self.h)).VZ()
+        return GradientResult(dn, "v")
 
-    def __init__(self, dn: vec3, direction: str):
-        """
-        Initialize gradient result
 
-        Args:
-            dn: Normal vector derivative (computed using proven algorithm)
-            direction: Parameter direction ('u' or 'v')
-        """
-        self.dn = dn
-        self.direction = direction
-
-    def __repr__(self) -> str:
-        return f"GradientResult({self.direction}: [{self.dn.x:.6f}, {self.dn.y:.6f}, {self.dn.z:.6f}])"
-
-
-# ========== Curvature Calculator ==========
+# ============================================================
+# Intrinsic Gradient Curvature Calculator
+# ============================================================
 
 class IntrinsicGradientCurvatureCalculator:
     """
-    Curvature calculator using the CORRECTED intrinsic gradient method
+    Curvature calculator using intrinsic gradient method.
+
+    Computes curvatures from the shape operator derived
+    from the intrinsic gradient of the normal field.
     """
-    
+
     def __init__(self, surface: Surface, step_size: float = 1e-3):
         self.surface = surface
         self.h = step_size
         self.grad_op = IntrinsicGradientOperator(surface, step_size)
 
-    def compute_gaussian_curvature(self, u: float, v: float) -> float:
-        """
-        Compute Gaussian curvature - CORRECTED implementation
-        """
-        # 计算梯度算子
-        G_u, G_v, _ = self.grad_op.compute_both(u, v)
-
-        # 计算切向量（使用解析表达式）
+    def _get_tangent_vectors(self, u: float, v: float) -> Tuple[vec3, vec3]:
+        """Get tangent vectors (analytical for known surfaces, numerical otherwise)."""
         if isinstance(self.surface, Sphere):
             R = self.surface.R
             theta, phi = u, v
@@ -314,42 +436,48 @@ class IntrinsicGradientCurvatureCalculator:
                 0
             )
         elif isinstance(self.surface, Torus):
-            R = self.surface.R  # 主半径
-            r = self.surface.r  # 副半径
-            u_param, v_param = u, v
+            R = self.surface.R
+            r = self.surface.r
             r_u = vec3(
-                -r * math.sin(u_param) * math.cos(v_param),
-                -r * math.sin(u_param) * math.sin(v_param),
-                r * math.cos(u_param)
+                -r * math.sin(u) * math.cos(v),
+                -r * math.sin(u) * math.sin(v),
+                r * math.cos(u)
             )
             r_v = vec3(
-                -(R + r * math.cos(u_param)) * math.sin(v_param),
-                (R + r * math.cos(u_param)) * math.cos(v_param),
+                -(R + r * math.cos(u)) * math.sin(v),
+                (R + r * math.cos(u)) * math.cos(v),
                 0
             )
         else:
-            # 对于其他曲面，使用数值导数
             r_u = self.surface.tangent_u(u, v)
             r_v = self.surface.tangent_v(u, v)
 
-        # 提取法向量导数
+        return r_u, r_v
+
+    def compute_gaussian_curvature(self, u: float, v: float) -> float:
+        """
+        Compute Gaussian curvature K = det(II) / det(I).
+        """
+        G_u, G_v, _ = self.grad_op.compute_both(u, v)
+        r_u, r_v = self._get_tangent_vectors(u, v)
+
         dn_du = G_u.dn
         dn_dv = G_v.dn
 
-        # 计算度量张量
+        # First fundamental form
         E = r_u.dot(r_u)
         F = r_u.dot(r_v)
         G = r_v.dot(r_v)
         metric_det = E * G - F * F
 
-        # 计算第二基本形式
+        # Second fundamental form
         L = -dn_du.dot(r_u)
         M1 = -dn_du.dot(r_v)
         M2 = -dn_dv.dot(r_u)
         N = -dn_dv.dot(r_v)
         M = (M1 + M2) / 2.0
 
-        # 高斯曲率
+        # Gaussian curvature
         if abs(metric_det) > 1e-14:
             K = (L * N - M * M) / metric_det
         else:
@@ -357,142 +485,12 @@ class IntrinsicGradientCurvatureCalculator:
 
         return K
 
-    def compute_riemann_curvature(self, u: float, v: float, proportional_correction: bool = True) -> float:
-        """
-        Compute Riemann curvature tensor component R^1_212
-        Based on the intrinsic gradient operator method with correct implementation
-        """
-        delta = self.h
-
-        # 计算中心标架
-        c_center = self.grad_op.calc_intrinsic_frame(u, v)
-
-        # 计算u方向的内禀梯度算子（使用中心差分）
-        c_u_plus = self.grad_op.calc_intrinsic_frame(u + delta, v)
-        c_u_minus = self.grad_op.calc_intrinsic_frame(u - delta, v)
-        G_u = (c_u_plus - c_u_minus) / (2 * delta)
-
-        # 计算v方向的内禀梯度算子（使用中心差分）
-        c_v_plus = self.grad_op.calc_intrinsic_frame(u, v + delta)
-        c_v_minus = self.grad_op.calc_intrinsic_frame(u, v - delta)
-        G_v = (c_v_plus - c_v_minus) / (2 * delta)
-
-        # 计算李括号 [G_u, G_v] = G_u ∘ G_v - G_v ∘ G_u
-        # 这需要计算二阶混合偏导数
-
-        # 计算 ∂²c/∂u∂v 和 ∂²c/∂v∂u
-        c_uv_pp = self.grad_op.calc_intrinsic_frame(u + delta, v + delta)
-        c_uv_pm = self.grad_op.calc_intrinsic_frame(u + delta, v - delta)
-        c_uv_mp = self.grad_op.calc_intrinsic_frame(u - delta, v + delta)
-        c_uv_mm = self.grad_op.calc_intrinsic_frame(u - delta, v - delta)
-
-        # 二阶混合偏导数 ∂²c/∂u∂v
-        d2c_dudv = (c_uv_pp - c_uv_pm - c_uv_mp + c_uv_mm) / (4 * delta * delta)
-
-        # 对于对称的联络，∂²c/∂v∂u = ∂²c/∂u∂v，所以李括号简化为零
-        # 但是我们需要考虑标架的非对易性
-
-        # 更精确的方法：直接计算联络系数的变化
-        # 提取法向量的导数（这是关键）
-        dn_du = G_u.VZ()  # 法向量在u方向的导数
-        dn_dv = G_v.VZ()  # 法向量在v方向的导数
-
-        # 计算切向量（根据曲面类型）
-        if isinstance(self.surface, Sphere):
-            R = self.surface.R
-            theta, phi = u, v
-            r_u = vec3(
-                R * math.cos(theta) * math.cos(phi),
-                R * math.cos(theta) * math.sin(phi),
-                -R * math.sin(theta)
-            )
-            r_v = vec3(
-                -R * math.sin(theta) * math.sin(phi),
-                R * math.sin(theta) * math.cos(phi),
-                0
-            )
-        elif isinstance(self.surface, Torus):
-            R = self.surface.R
-            r = self.surface.r
-            u_param, v_param = u, v
-            r_u = vec3(
-                -r * math.sin(u_param) * math.cos(v_param),
-                -r * math.sin(u_param) * math.sin(v_param),
-                r * math.cos(u_param)
-            )
-            r_v = vec3(
-                -(R + r * math.cos(u_param)) * math.sin(v_param),
-                (R + r * math.cos(u_param)) * math.cos(v_param),
-                0
-            )
-        else:
-            r_u = self.surface.tangent_u(u, v)
-            r_v = self.surface.tangent_v(u, v)
-
-        # 计算第二基本形式系数
-        L = -dn_du.dot(r_u)
-        M = -dn_du.dot(r_v)
-        N = -dn_dv.dot(r_v)
-
-        # 计算度量张量
-        E = r_u.dot(r_u)
-        F = r_u.dot(r_v)
-        G_metric = r_v.dot(r_v)
-        det_g = E * G_metric - F * F
-
-        # 对于2D曲面，黎曼曲率张量只有一个独立分量
-        # R^1_212 与高斯曲率的关系为：K = R^1_212 / det(g)
-        # 所以 R^1_212 = K * det(g) = (LN - M²)
-
-        R_1212 = L * N - M * M
-
-        # 比例修正（对于某些参数化可能需要）
-        if proportional_correction and isinstance(self.surface, Sphere):
-            # 对于球面的某些参数化，可能需要额外的修正
-            # 但是基于第二基本形式的计算通常不需要
-            pass
-
-        return R_1212
-
     def compute_mean_curvature(self, u: float, v: float) -> float:
         """
-        Compute mean curvature
+        Compute mean curvature H = (EN - 2FM + GL) / (2*det(I)).
         """
-        # 使用与高斯曲率相同的计算方法
         G_u, G_v, _ = self.grad_op.compute_both(u, v)
-
-        # 计算切向量（使用解析表达式）
-        if isinstance(self.surface, Sphere):
-            R = self.surface.R
-            theta, phi = u, v
-            r_u = vec3(
-                R * math.cos(theta) * math.cos(phi),
-                R * math.cos(theta) * math.sin(phi),
-                -R * math.sin(theta)
-            )
-            r_v = vec3(
-                -R * math.sin(theta) * math.sin(phi),
-                R * math.sin(theta) * math.cos(phi),
-                0
-            )
-        elif isinstance(self.surface, Torus):
-            R = self.surface.R  # 主半径
-            r = self.surface.r  # 副半径
-            u_param, v_param = u, v
-            r_u = vec3(
-                -r * math.sin(u_param) * math.cos(v_param),
-                -r * math.sin(u_param) * math.sin(v_param),
-                r * math.cos(u_param)
-            )
-            r_v = vec3(
-                -(R + r * math.cos(u_param)) * math.sin(v_param),
-                (R + r * math.cos(u_param)) * math.cos(v_param),
-                0
-            )
-        else:
-            # 对于其他曲面，使用数值导数
-            r_u = self.surface.tangent_u(u, v)
-            r_v = self.surface.tangent_v(u, v)
+        r_u, r_v = self._get_tangent_vectors(u, v)
 
         dn_du = G_u.dn
         dn_dv = G_v.dn
@@ -508,7 +506,6 @@ class IntrinsicGradientCurvatureCalculator:
         N = -dn_dv.dot(r_v)
         M = (M1 + M2) / 2.0
 
-        # 平均曲率
         if abs(metric_det) > 1e-14:
             H = (G * L - 2 * F * M + E * N) / (2 * metric_det)
         else:
@@ -516,14 +513,49 @@ class IntrinsicGradientCurvatureCalculator:
 
         return H
 
-    def compute_all_curvatures(self, u: float, v: float) -> dict:
+    def compute_riemann_curvature(self, u: float, v: float) -> float:
+        """
+        Compute Riemann curvature tensor component R^1_212.
+
+        For 2D surfaces: R^1_212 = K * det(g) = LN - M^2
+        """
+        G_u, G_v, _ = self.grad_op.compute_both(u, v)
+        r_u, r_v = self._get_tangent_vectors(u, v)
+
+        dn_du = G_u.dn
+        dn_dv = G_v.dn
+
+        L = -dn_du.dot(r_u)
+        M = -dn_du.dot(r_v)
+        N = -dn_dv.dot(r_v)
+
+        R_1212 = L * N - M * M
+        return R_1212
+
+    def compute_principal_curvatures(self, u: float, v: float) -> Tuple[float, float]:
+        """
+        Compute principal curvatures k1 and k2.
+
+        Uses: k1,k2 = H +/- sqrt(H^2 - K)
+        """
+        K = self.compute_gaussian_curvature(u, v)
+        H = self.compute_mean_curvature(u, v)
+
+        discriminant = max(0, H * H - K)
+        sqrt_disc = discriminant ** 0.5
+
+        k1 = H + sqrt_disc
+        k2 = H - sqrt_disc
+
+        return k1, k2
+
+    def compute_all_curvatures(self, u: float, v: float) -> Dict[str, Union[float, Tuple[float, float]]]:
         """
-        Compute all curvature quantities
+        Compute all curvature quantities at once.
         """
         K = self.compute_gaussian_curvature(u, v)
         H = self.compute_mean_curvature(u, v)
-        
-        # 主曲率
+
         discriminant = max(0, H * H - K)
         sqrt_disc = discriminant ** 0.5
         k1 = H + sqrt_disc
@@ -536,7 +568,172 @@ class IntrinsicGradientCurvatureCalculator:
         }
 
 
-# ========== Convenience Functions ==========
+# ============================================================
+# Classical Curvature Calculator
+# ============================================================
+
+class CurvatureCalculator:
+    """
+    High-precision discrete curvature calculator using classical differential geometry.
+
+    Uses high-order finite differences for computing derivatives
+    and fundamental forms.
+    """
+
+    def __init__(self, surface: Surface, step_size: float = 1e-3):
+        self.surface = surface
+        self.h = step_size
+
+    def _position_array(self, u: float, v: float) -> np.ndarray:
+        """Convert vec3 position to numpy array."""
+        pos = self.surface.position(u, v)
+        return np.array([pos.x, pos.y, pos.z])
+
+    def _compute_derivatives(self, u: float, v: float) -> Dict[str, np.ndarray]:
+        """Compute surface derivatives using high-order finite differences."""
+        effective_h = max(self.h, 1e-6)
+
+        r_u = derivative_5pt(lambda uu: self._position_array(uu, v), u, effective_h)
+        r_v = derivative_5pt(lambda vv: self._position_array(u, vv), v, effective_h)
+
+        r_uu = derivative_2nd_5pt(lambda uu: self._position_array(uu, v), u, effective_h)
+        r_vv = derivative_2nd_5pt(lambda vv: self._position_array(u, vv), v, effective_h)
+        r_uv = derivative_5pt(
+            lambda vv: derivative_5pt(
+                lambda uu: self._position_array(uu, vv), u, effective_h
+            ), v, effective_h
+        )
+
+        return {
+            'r_u': r_u, 'r_v': r_v,
+            'r_uu': r_uu, 'r_vv': r_vv, 'r_uv': r_uv
+        }
+
+    def compute_fundamental_forms(self, u: float, v: float) -> Tuple[np.ndarray, np.ndarray, np.ndarray]:
+        """
+        Compute first and second fundamental forms.
+
+        Returns:
+            (g, h, n): First form, second form, normal vector
+        """
+        derivs = self._compute_derivatives(u, v)
+        r_u = derivs['r_u']
+        r_v = derivs['r_v']
+        r_uu = derivs['r_uu']
+        r_vv = derivs['r_vv']
+        r_uv = derivs['r_uv']
+
+        # First fundamental form
+        E = np.dot(r_u, r_u)
+        F = np.dot(r_u, r_v)
+        G = np.dot(r_v, r_v)
+        g = np.array([[E, F], [F, G]])
+
+        # Normal vector
+        n_vec = np.cross(r_u, r_v)
+        n_norm = np.linalg.norm(n_vec)
+        if n_norm > 1e-14:
+            n = n_vec / n_norm
+        else:
+            n = np.array([0., 0., 1.])
+
+        # Second fundamental form
+        L = np.dot(r_uu, n)
+        M = np.dot(r_uv, n)
+        N = np.dot(r_vv, n)
+        h = np.array([[L, M], [M, N]])
+
+        return g, h, n
+
+    def compute_gaussian_curvature(self, u: float, v: float) -> float:
+        """Compute Gaussian curvature K = det(II) / det(I)."""
+        g, h, _ = self.compute_fundamental_forms(u, v)
+
+        det_g = np.linalg.det(g)
+        det_h = np.linalg.det(h)
+
+        if abs(det_g) < 1e-14:
+            return 0.0
+
+        return det_h / det_g
+
+    def compute_mean_curvature(self, u: float, v: float) -> float:
+        """Compute mean curvature H."""
+        g, h, _ = self.compute_fundamental_forms(u, v)
+
+        det_g = np.linalg.det(g)
+        if abs(det_g) < 1e-14:
+            return 0.0
+
+        trace_term = g[1, 1] * h[0, 0] - 2 * g[0, 1] * h[0, 1] + g[0, 0] * h[1, 1]
+        H = trace_term / (2 * det_g)
+
+        return abs(H)
+
+    def compute_principal_curvatures(self, u: float, v: float) -> Tuple[float, float, np.ndarray, np.ndarray]:
+        """
+        Compute principal curvatures and principal directions.
+
+        Returns:
+            (k1, k2, dir1, dir2): Principal curvatures and directions
+        """
+        g, h, _ = self.compute_fundamental_forms(u, v)
+        derivs = self._compute_derivatives(u, v)
+        r_u = derivs['r_u']
+        r_v = derivs['r_v']
+
+        det_g = np.linalg.det(g)
+        if abs(det_g) < 1e-14:
+            return 0.0, 0.0, np.array([1., 0., 0.]), np.array([0., 1., 0.])
+
+        # Shape operator S = g^{-1} h
+        g_inv = np.linalg.inv(g)
+        S = g_inv @ h
+
+        # Eigenvalue decomposition
+        eigenvalues, eigenvectors = np.linalg.eig(S)
+        k1, k2 = eigenvalues.real
+
+        # Ensure k1 >= k2 by absolute value
+        if abs(k1) < abs(k2):
+            k1, k2 = k2, k1
+            eigenvectors = eigenvectors[:, [1, 0]]
+
+        # Convert to 3D directions
+        dir1_2d = eigenvectors[:, 0]
+        dir2_2d = eigenvectors[:, 1]
+
+        dir1_3d = dir1_2d[0] * r_u + dir1_2d[1] * r_v
+        dir2_3d = dir2_2d[0] * r_u + dir2_2d[1] * r_v
+
+        dir1_3d = dir1_3d / (np.linalg.norm(dir1_3d) + 1e-14)
+        dir2_3d = dir2_3d / (np.linalg.norm(dir2_3d) + 1e-14)
+
+        return k1, k2, dir1_3d, dir2_3d
+
+    def compute_all_curvatures(self, u: float, v: float) -> Dict[str, Union[float, np.ndarray]]:
+        """Compute all curvature quantities at once."""
+        g, h, n = self.compute_fundamental_forms(u, v)
+        K = self.compute_gaussian_curvature(u, v)
+        H = self.compute_mean_curvature(u, v)
+        k1, k2, dir1, dir2 = self.compute_principal_curvatures(u, v)
+
+        return {
+            'K': K,
+            'H': H,
+            'k1': k1,
+            'k2': k2,
+            'dir1': dir1,
+            'dir2': dir2,
+            'g': g,
+            'h': h,
+            'n': n
+        }
+
+
+# ============================================================
+# Convenience Functions - Intrinsic Gradient Method (Default)
+# ============================================================
 
 def compute_gaussian_curvature(
     surface: Surface,
@@ -545,15 +742,7 @@ def compute_gaussian_curvature(
     step_size: float = 1e-3
 ) -> float:
     """
-    Compute Gaussian curvature
-
-    Args:
-        surface: Surface object
-        u, v: Parameter values
-        step_size: Step size for numerical differentiation
-
-    Returns:
-        Gaussian curvature value
+    Compute Gaussian curvature using intrinsic gradient method.
     """
     calc = IntrinsicGradientCurvatureCalculator(surface, step_size)
     return calc.compute_gaussian_curvature(u, v)
@@ -566,7 +755,7 @@ def compute_mean_curvature(
     step_size: float = 1e-3
 ) -> float:
     """
-    Compute mean curvature
+    Compute mean curvature using intrinsic gradient method.
     """
     calc = IntrinsicGradientCurvatureCalculator(surface, step_size)
     return calc.compute_mean_curvature(u, v)
@@ -576,23 +765,13 @@ def compute_riemann_curvature(
     surface: Surface,
     u: float,
     v: float,
-    step_size: float = 1e-3,
-    proportional_correction: bool = True
+    step_size: float = 1e-3
 ) -> float:
     """
-    Compute Riemann curvature tensor component R^1_212
-
-    Args:
-        surface: Surface object
-        u, v: Parameter values
-        step_size: Step size for numerical differentiation
-        proportional_correction: Apply proportional correction for spherical coordinates
-
-    Returns:
-        Riemann curvature tensor component R^1_212
+    Compute Riemann curvature tensor component R^1_212.
     """
     calc = IntrinsicGradientCurvatureCalculator(surface, step_size)
-    return calc.compute_riemann_curvature(u, v, proportional_correction)
+    return calc.compute_riemann_curvature(u, v)
 
 
 def compute_all_curvatures(
@@ -600,9 +779,9 @@ def compute_all_curvatures(
     u: float,
     v: float,
     step_size: float = 1e-3
-) -> dict:
+) -> Dict[str, Union[float, Tuple[float, float]]]:
     """
-    Compute all curvature quantities
+    Compute all curvature quantities using intrinsic gradient method.
     """
     calc = IntrinsicGradientCurvatureCalculator(surface, step_size)
     return calc.compute_all_curvatures(u, v)
@@ -616,7 +795,7 @@ def compute_intrinsic_gradient(
     step_size: float = 1e-3
 ) -> GradientResult:
     """
-    Compute intrinsic gradient in specified direction
+    Compute intrinsic gradient in specified direction.
 
     Args:
         surface: Surface object
@@ -628,24 +807,142 @@ def compute_intrinsic_gradient(
         GradientResult object
     """
     grad_op = IntrinsicGradientOperator(surface, step_size)
-    
+
     if direction == 'u':
-        # Compute central difference for u direction
-        c_plus = grad_op.calc_intrinsic_frame(u + step_size, v)
-        c_minus = grad_op.calc_intrinsic_frame(u - step_size, v)
-        dn = ((c_plus - c_minus) / (2 * step_size)).VZ()
+        return grad_op.compute_u(u, v)
     elif direction == 'v':
-        # Compute central difference for v direction
-        c_plus = grad_op.calc_intrinsic_frame(u, v + step_size)
-        c_minus = grad_op.calc_intrinsic_frame(u, v - step_size)
-        dn = ((c_plus - c_minus) / (2 * step_size)).VZ()
+        return grad_op.compute_v(u, v)
     else:
         raise ValueError(f"direction must be 'u' or 'v', got: {direction}")
 
-    return GradientResult(dn, direction)
+
+# ============================================================
+# Convenience Functions - Classical Method
+# ============================================================
+
+def gaussian_curvature_classical(
+    surface: Surface,
+    u: float,
+    v: float,
+    step_size: float = 1e-3
+) -> float:
+    """Compute Gaussian curvature using classical method."""
+    calc = CurvatureCalculator(surface, step_size)
+    return calc.compute_gaussian_curvature(u, v)
+
+
+def mean_curvature_classical(
+    surface: Surface,
+    u: float,
+    v: float,
+    step_size: float = 1e-3
+) -> float:
+    """Compute mean curvature using classical method."""
+    calc = CurvatureCalculator(surface, step_size)
+    return calc.compute_mean_curvature(u, v)
+
+
+def principal_curvatures_classical(
+    surface: Surface,
+    u: float,
+    v: float,
+    step_size: float = 1e-3
+) -> Tuple[float, float]:
+    """Compute principal curvatures using classical method."""
+    calc = CurvatureCalculator(surface, step_size)
+    k1, k2, _, _ = calc.compute_principal_curvatures(u, v)
+    return k1, k2
+
+
+def all_curvatures_classical(
+    surface: Surface,
+    u: float,
+    v: float,
+    step_size: float = 1e-3
+) -> Dict[str, Union[float, np.ndarray]]:
+    """Compute all curvature quantities using classical method."""
+    calc = CurvatureCalculator(surface, step_size)
+    return calc.compute_all_curvatures(u, v)
+
+
+# ============================================================
+# Backward Compatibility Aliases
+# ============================================================
+
+def gaussian_curvature(
+    surface: Surface,
+    u: float,
+    v: float,
+    step_size: float = 1e-3
+) -> float:
+    """Compute Gaussian curvature (default: intrinsic gradient method)."""
+    return compute_gaussian_curvature(surface, u, v, step_size)
+
+
+def mean_curvature(
+    surface: Surface,
+    u: float,
+    v: float,
+    step_size: float = 1e-3
+) -> float:
+    """Compute mean curvature (default: intrinsic gradient method)."""
+    return compute_mean_curvature(surface, u, v, step_size)
+
+
+def principal_curvatures(
+    surface: Surface,
+    u: float,
+    v: float,
+    step_size: float = 1e-3
+) -> Tuple[float, float]:
+    """Compute principal curvatures (default: intrinsic gradient method)."""
+    calc = IntrinsicGradientCurvatureCalculator(surface, step_size)
+    return calc.compute_principal_curvatures(u, v)
+
+
+def all_curvatures(
+    surface: Surface,
+    u: float,
+    v: float,
+    step_size: float = 1e-3
+) -> Dict[str, Union[float, Tuple[float, float]]]:
+    """Compute all curvature quantities (default: intrinsic gradient method)."""
+    return compute_all_curvatures(surface, u, v, step_size)
 
 
-# ========== Export ==========
+# ============================================================
+# Method Comparison
+# ============================================================
+
+def compare_methods(
+    surface: Surface,
+    u: float,
+    v: float,
+    step_size: float = 1e-3
+) -> Dict[str, float]:
+    """
+    Compare classical and intrinsic gradient curvature methods.
+
+    Returns:
+        Dictionary with curvature values from both methods and error metrics
+    """
+    K_classical = gaussian_curvature_classical(surface, u, v, step_size)
+    K_intrinsic = compute_gaussian_curvature(surface, u, v, step_size)
+
+    difference = abs(K_classical - K_intrinsic)
+    relative_error = difference / abs(K_classical) if abs(K_classical) > 1e-14 else 0.0
+
+    return {
+        'K_classical': K_classical,
+        'K_intrinsic': K_intrinsic,
+        'difference': difference,
+        'relative_error': relative_error
+    }
+
+
+# ============================================================
+# Export
+# ============================================================
 
 __all__ = [
     # Surface classes
@@ -655,12 +952,35 @@ __all__ = [
 
     # Core classes
     'MetricTensor',
+    'GradientResult',
     'IntrinsicGradientOperator',
     'IntrinsicGradientCurvatureCalculator',
+    'CurvatureCalculator',
 
-    # Functions
+    # Intrinsic gradient method functions (default)
     'compute_gaussian_curvature',
     'compute_mean_curvature',
+    'compute_riemann_curvature',
     'compute_all_curvatures',
     'compute_intrinsic_gradient',
-]
+
+    # Classical method functions
+    'gaussian_curvature_classical',
+    'mean_curvature_classical',
+    'principal_curvatures_classical',
+    'all_curvatures_classical',
+
+    # Backward compatibility aliases
+    'gaussian_curvature',
+    'mean_curvature',
+    'principal_curvatures',
+    'all_curvatures',
+
+    # Method comparison
+    'compare_methods',
+
+    # Utility functions
+    'derivative_5pt',
+    'derivative_2nd_5pt',
+    'richardson_extrapolation',
+]