avlgmath 1.0.0

Advanced Vector & Linear Algebra / Numerical Methods Library

AVLGMath - Advanced Vector & Linear Algebra / Numerical Methods Library

A high-performance Python library for advanced linear algebra, numerical interpolation, numerical integration, matrix operations, and linear system solving. Optimized for numerical computing with NumPy BLAS/LAPACK backends.

Installation

pip install avlgmath

Features

• Interpolation Methods: Newton Forward/Backward Difference, Lagrangian
• Numerical Integration: Trapezoidal Rule, Simpson's 1/3 Rule with error estimation
• Linear Algebra: Matrix operations, eigenvalues, eigenvectors, rank computation
• Linear System Solving: Jacobi iteration method
• Vector Analysis: Dependence checking
• Differential Equations: First-order linear ODE solver
• Taylor Series: Taylor theorem expansions for 1 and 2 variables

Quick Start

import numpy as np
from avlgmath import *

# Newton Forward Difference Interpolation
X = np.array([0, 1, 2, 3])
Y = np.array([1, 3, 7, 13])
table = nfd(X, Y)
y_interp = nfd_interpolate(X, Y, 1.5)

# Matrix Operations
A = np.array([[1, 2], [3, 4]])
eigenvalues = eigenval(A)

# Jacobi Iteration
A = np.array([[4, -1], [-1, 3]])
b = np.array([7, 5])
result = jacobi_iter(A, b, n=20)

---

Function Reference

Interpolation Functions

nfd(X, Y)

Newton Forward Difference Table - Creates difference table for equally spaced points

Parameter	Type	Description
X	numpy array	x-coordinates (equally spaced)
Y	numpy array	y-coordinates
Returns	numpy array	2D array of differences

X = np.array([0, 1, 2, 3, 4])
Y = np.array([1, 2, 5, 10, 17])
table = nfd(X, Y)

nfd_interpolate(X, Y, x)

Newton Forward Difference Interpolation - Interpolates value at point x

Parameter	Type	Description
X	numpy array	x-coordinates (equally spaced)
Y	numpy array	y-coordinates
x	float	Point at which to interpolate
Returns	float	Interpolated value

X = np.array([0, 1, 2, 3])
Y = np.array([1, 2, 5, 10])
y_at_1_5 = nfd_interpolate(X, Y, 1.5)  # ≈ 3.375

nfdfrom_csv(path)

Load Newton forward difference from CSV file

Parameter	Type	Description
path	str	Path to CSV file with columns 'x' and 'y'
Returns	numpy array	Difference table

nbd(X, Y)

Newton Backward Difference Table - Creates difference table from end points

Parameter	Type	Description
X	numpy array	x-coordinates (equally spaced)
Y	numpy array	y-coordinates
Returns	numpy array	2D array of differences

nbd_interpolate(X, Y, x)

Newton Backward Difference Interpolation - Interpolates using backward differences

Parameter	Type	Description
X	numpy array	x-coordinates (equally spaced)
Y	numpy array	y-coordinates
x	float	Point at which to interpolate
Returns	float	Interpolated value

nbdfrom_csv(path)

Load Newton backward difference from CSV file

Parameter	Type	Description
path	str	Path to CSV file with columns 'x' and 'y'
Returns	numpy array	Difference table

lagrangian(X, Y)

Lagrangian Interpolation - Creates interpolating polynomial through all points

Parameter	Type	Description
X	numpy array	x-coordinates (must be distinct)
Y	numpy array	y-coordinates
Returns	dict	Contains polynomial object and evaluation function

X = np.array([1, 2, 3])
Y = np.array([2, 3, 5])
result = lagrangian(X, Y)
y_at_1_5 = result['evaluate'](1.5)

---

Numerical Integration Functions

ploytrap(coefficients, a, b, n=100)

Polynomial Trapezoidal Rule - Integrates polynomial using trapezoidal method

Parameter	Type	Description
coefficients	list/array	Polynomial coefficients [c₀, c₁, c₂, ...]
a	float	Lower limit of integration
b	float	Upper limit of integration
n	int	Number of subintervals (default 100)
Returns	float	Estimated integral value

coeffs = [1, 2, 1]  # 1 + 2x + x²
result = ploytrap(coeffs, a=0, b=3, n=100)

polytrap_e(coefficients, a, b, n=100)

Trapezoidal Rule Error Estimation

Parameter	Type	Description
coefficients	list/array	Polynomial coefficients
a	float	Lower limit
b	float	Upper limit
n	int	Number of subintervals
Returns	dict	Error bound and max second derivative

polysim(coefficients, a, b, n=100)

Polynomial Simpson's 1/3 Rule - More accurate than trapezoidal

Parameter	Type	Description
coefficients	list/array	Polynomial coefficients
a	float	Lower limit of integration
b	float	Upper limit of integration
n	int	Number of subintervals (must be even)
Returns	float	Estimated integral value

polysim_e(coefficients, a, b, n=100)

Simpson's 1/3 Rule Error Estimation

Parameter	Type	Description
coefficients	list/array	Polynomial coefficients
a	float	Lower limit
b	float	Upper limit
n	int	Number of subintervals (must be even)
Returns	dict	Error bound and max fourth derivative

---

Matrix Operations

matrixmul(A, B)

Efficient Matrix Multiplication - BLAS-optimized

Parameter	Type	Description
A	numpy array	First matrix (m × n)
B	numpy array	Second matrix (n × p)
Returns	numpy array	Result matrix (m × p)

diagonalmul(A, diagonal=None)

Diagonal Matrix Multiplication - O(mn) instead of O(mn²)

Parameter	Type	Description
A	numpy array	2D matrix
diagonal	array/scalar	Diagonal elements (None = extract from A)
Returns	dict	Result matrix and operation details

rankm(A)

Matrix Rank - Using efficient SVD method

Parameter	Type	Description
A	numpy array	Any 2D matrix
Returns	int	Rank of the matrix

eigenval(A)

Eigenvalues - For square or singular values for rectangular matrices

Parameter	Type	Description
A	numpy array	Any 2D matrix
Returns	dict	Type and values

A = np.array([[1, 2], [3, 4]])
result = eigenval(A)
# {'type': 'eigenvalues', 'values': array([5.372, -0.372])}

eigenvec(A)

Eigenvectors - Unique, normalized eigenvectors or singular vectors

Parameter	Type	Description
A	numpy array	Any 2D matrix
Returns	dict	Eigenvectors with eigenvalues

---

Linear System Solving

jacobi_iter(A, b, n=10, x0=None, tolerance=1e-6)

Jacobi Iteration - Solves Ax = b using iterative method

Parameter	Type	Description
A	numpy array	Coefficient matrix (n × n, square)
b	numpy array	Constants vector
n	int	Maximum iterations
x0	array	Initial guess (default zero vector)
tolerance	float	Convergence tolerance
Returns	dict	Solution, convergence history, residual

A = np.array([[4, -1], [-1, 3]], dtype=float)
b = np.array([7, 5], dtype=float)
result = jacobi_iter(A, b, n=20, tolerance=1e-6)
print(result['solution'])  # [2., 1.]
print(result['converged'])  # True

---

Vector Analysis

depvec(vectors)

Dependent Vectors - Check linear dependence/independence

Parameter	Type	Description
vectors	numpy array	2D array where each row is a vector
Returns	dict	Dependent flag (1=dependent, 0=independent)

v = np.array([[1, 0], [0, 1]])
result = depvec(v)
# {'dependent': 0, 'rank': 2, 'explanation': 'Linearly independent'}

---

Differential Equations

Lde(p_coeffs, q_coeffs, y0=0)

Linear Differential Equation Solver - Solves dy/dx + p(x)y = q(x)

Parameter	Type	Description
p_coeffs	list/array	Coefficients of p(x) polynomial
q_coeffs	list/array	Coefficients of q(x) polynomial
y0	float	Initial condition y(0)
Returns	dict	Particular and general solutions

p_coeffs = [0, 2]  # 0 + 2x
q_coeffs = [0, 1]  # 0 + x
result = Lde(p_coeffs, q_coeffs, y0=1)

---

Taylor Series Functions

taylor_1var(f_expr, x0, n, x_eval=None)

Taylor Series Expansion - Single Variable

Computes the Taylor polynomial of degree n for a function around point x₀.

Parameter	Type	Description
f_expr	str or sympy expr	Function expression (e.g., "sin(x)", "exp(x)")
x0	float	Center point for Taylor expansion
n	int	Degree of Taylor polynomial
x_eval	float	Optional point to evaluate Taylor approximation
Returns	dict	Symbolic series, value, error, terms

# Taylor expansion of sin(x) around x=0
result = taylor_1var("sin(x)", 0, 5)
print(result['series'])  # x**5/120 - x**3/6 + x

# Evaluate at x=0.1
result = taylor_1var("sin(x)", 0, 5, x_eval=0.1)
print(f"Approximation: {result['value']}")
print(f"Actual value: {result['actual_value']}")
print(f"Error: {result['error']}")

# Exponential function
result = taylor_1var("exp(x)", 0, 4, x_eval=0.5)

taylor_2var(f_expr, x0, y0, n, x_eval=None, y_eval=None)

Taylor Series Expansion - Two Variables

Computes the Taylor polynomial of degree n for a two-variable function around point (x₀, y₀).

Parameter	Type	Description
f_expr	str or sympy expr	Function of x,y (e.g., "x2 + y2", "sin(x)*cos(y)")
x0	float	Center point x-coordinate
y0	float	Center point y-coordinate
n	int	Degree of Taylor polynomial
x_eval	float	Optional x-coordinate for evaluation
y_eval	float	Optional y-coordinate for evaluation
Returns	dict	Symbolic series, partial derivatives, value, error

# Taylor expansion of x² + y² + xy around (0,0)
result = taylor_2var("x**2 + y**2 + x*y", 0, 0, 2)
print(result['series'])  # x**2 + x*y + y**2

# With evaluation
result = taylor_2var("x**2 + y**2 + x*y", 0, 0, 2, x_eval=0.1, y_eval=0.1)
print(f"Taylor value: {result['value']}")  # 0.03
print(f"Partial derivatives: {result['partial_derivatives']}")

# Two-variable trigonometric function
result = taylor_2var("sin(x)*cos(y)", 0, 0, 3, x_eval=0.2, y_eval=0.2)
print(f"Error: {result['error']}")

---

Requirements

numpy>=1.20
pandas>=1.3
scipy>=1.7
sympy>=1.9

Performance Notes

• All functions use NumPy with BLAS/LAPACK backends
• Matrix operations: O(n³) for multiplication
• Interpolation: O(n) per evaluation
• Jacobi iteration: O(n²) per iteration
• Simpson's rule: ~100x more accurate than trapezoidal

Error Handling

Comprehensive validation included:

• Dimension checks
• Singular matrix detection
• Invalid parameter ranges
• Convergence monitoring

License

MIT License - Free to use and modify

Contact & Support

For issues, contributions, and feedback, please submit to the repository.

Citation

@software{avlgmath2026,
  title = {AVLGMath: Advanced Linear Algebra and Numerical Methods Library},
  year = {2026}
}