diff options
author | j-james | 2022-11-15 08:24:05 +0000 |
---|---|---|
committer | j-james | 2022-11-15 08:24:05 +0000 |
commit | ab451b3466b7f5151db9556fdf326934f86bd246 (patch) | |
tree | d8cdd9c49bc11cc05110e20b787ceb9cde997962 /lalge | |
parent | ef73dc1db2f0389912e442b0f6cd5fa6712d4026 (diff) |
Add my linear algebra library
Diffstat (limited to 'lalge')
-rw-r--r-- | lalge/README.md | 46 | ||||
-rw-r--r-- | lalge/src/matrices.nim | 171 | ||||
-rw-r--r-- | lalge/src/types.nim | 8 | ||||
-rw-r--r-- | lalge/src/vectors.nim | 57 | ||||
-rw-r--r-- | lalge/tests/nim.cfg | 1 | ||||
-rw-r--r-- | lalge/tests/tests.nim | 144 |
6 files changed, 427 insertions, 0 deletions
diff --git a/lalge/README.md b/lalge/README.md new file mode 100644 index 0000000..640654d --- /dev/null +++ b/lalge/README.md @@ -0,0 +1,46 @@ +# lalge (name tbd) + +A functional, dead simple linear algebra library written in Nim. Mostly for learning purposes. + +This is also serving as my finals review, haha. + +If you're looking for an actually robust linear algebra library for Nim, check out [Neo](https://github.com/andreaferretti/neo). + +## What's this language? + +Nim is a compiled whitespace-sensitive systems language with a strong type system and interesting memory management capabilities. + +It doesn't quite beat Python's usability or Rust's features, but I like it quite a lot, and think it's a good complement to other languages I use. + +Nim's [UFCS](https://en.wikipedia.org/wiki/Uniform_Function_Call_Syntax) (TL;DR: `a.add(b) == add(a, b)`) and operator overloading support is also a major draw for this type of project. + +## Architecture + +lalge uses Nim's `seqs` - dynamic arrays, equivalent to Rust's `Vec`, Go's `slice`, or Java's `ArrayList` - to implement matrices. This may have some performance overhead. On the other hand, it has the benefit of making the code much simpler. + +This also has the benefit of being able to declare matrices as such: +```nim +example: Matrix = @[ + @[1, 0, 0], + @[0, 1, 0], + @[0, 0, 1], +] +``` + +## Goals +- [ ] Cover the basic operations associated with the MATH 221 / typical introductory linear algebra curriculum + - [x] Addition / Subtraction + - [x] (Scalar / Vector) Multiplication + - [x] Inverse + - [x] Transpose + - [ ] Eigenvalues / Eigenvectors + - [ ] everything that i am forgetting +- [ ] Helpful characteristic functions + - [ ] Dimension + - [ ] Rank / Nullity + - [ ] Spans + - [ ] Determinant +- [x] Support arbitrary-sized matrices with internal seq representation +- [x] Support specialized "vectors", as opposed to Nx1 matrices + +After I get to a reasonable point of completeness, I'd love to go through and compare my library to Eigen (generally praised as a Very Good library) diff --git a/lalge/src/matrices.nim b/lalge/src/matrices.nim new file mode 100644 index 0000000..8e3e680 --- /dev/null +++ b/lalge/src/matrices.nim @@ -0,0 +1,171 @@ +import std/math, std/sugar +import types + +# A few notes about Nim: + +# "result" lets us operate on the return type +# without (necessarily) constructing it. +# this is very useful, and as such i use it a lot. + +# nim also supplies many functional tools which +# i also made liberal use of, occasionally to the +# detriment of readability. sorry. + +# the "*" operator on procs is the public marker. +# oh, and functions with "side effects" are called "procs". + +# TODO: implement a better error handling system +const not_square = "The matrix is not square" +const mismatch = "Mismatched rows and columns" + +## Overrides the echo() proc to provide more human-readable output +proc echo*(A: Matrix) = + echo "@[" + for row in A: + echo " ", row + echo "]" + +## Generate a new, filled XxX Matrix +func gen*(rows, columns: Natural, element: float = 0.0): Matrix = + result.setLen(rows) + for i in 0 ..< rows: + result[i].setLen(columns) + for j in 0 ..< columns: + result[i][j] = element + +## Returns the number of rows in a matrix +func rows*(A: Matrix): int = + return A.len() + +## Returns the number of columns in a matrix +func cols*(A: Matrix): int = + if A.rows() == 0: + return 0 + else: + return A[0].len() + +## Simple rows iterator for uniformity +iterator rows*(A: Matrix): RowVector = + for row in A: + yield row + +## Simple column iterator for matrices +iterator cols*(A: Matrix): Vector = + for i in 0 ..< A.cols(): + var row: Vector + for j in 0 ..< A.rows(): + row.add(A[i][j]) + yield row + +## Apply an arbitrary function to every element of a matrix +func map*(A: Matrix, oper: proc(i, j: int): float): Matrix = + result = gen(A.rows(), A.cols()) + for i in 0 ..< A.rows(): + for j in 0 ..< A.cols(): + # Note that the position is _actively_ provided + result[i][j] = oper(i, j) + +# TODO: check if these meet assert requirements + +## Matrix addition +func `+`*(A, B: Matrix): Matrix = + return map(A, (i, j) => A[i][j] + B[i][j]) + +## Matrix subtraction +func `-`*(A, B: Matrix): Matrix = + return map(A, (i, j) => A[i][j] - B[i][j]) + +## Scalar-Matrix multiplication +func `*`*(a: float, B: Matrix): Matrix = + return map(B, (i, j) => a * B[i][j]) + +## Scalar-Matrix multiplication +# func `*`*(A: Matrix, b: int): Matrix = +# return b * A + +## Matrix multiplication +func `*`*(A, B: Matrix): Matrix = + assert A.rows() == B.cols(), mismatch + result = gen(A.rows(), B.cols()) + for i in 0 ..< A.rows(): + for j in 0 ..< B.cols(): + for k in 0 ..< B.rows(): + result[i][j] += A[i][k] * B[k][j] + +## Absolute value of a matrix +func abs*(A: Matrix): Matrix = + return map(A, (i, j) => abs(A[i][j])) + +## Simple row function for uniformity +func row*(A: Matrix, r: int): RowVector = + return A[r] + +## Returns a "column vector" of a matrix as a Xx1 Matrix +# func col*(A: Matrix, j: int): Vector = +# result = gen(A.rows(), 1) +# for i in 0 ..< A.rows(): +# result[i] = @[A[i][j]] + +## Alternate implementation of col, returns a Vector +func col*(A: Matrix, c: int): Vector = + for row in A: + result.add(@[row[c]]) + +## Produce a vector of the diagonal entries of a matrix +func diag*(A: Matrix): Vector = + assert A.rows() == A.cols(), not_square + for i in 0 ..< A.rows(): + result.add(A[i][i]) + +## Transpose the rows and columns of a matrix +func transpose*(A: Matrix): Matrix = + result = gen(A.cols(), A.rows()) + return map(result, (i, j) => (A[j][i])) + +# func transpose*(a: Vector): Matrix = +# result = gen(a.len(), 1) +# return map(result, (i, j) => (a[j])) + +## Generate an arbitary sized identity matrix +func identity*(size: Natural): Matrix = + return map(gen(size, size), (i, j) => (if i==j: 1.0 else: 0.0)) + +## Oft-used identity matrices +let + I1*: Matrix = identity(1) + I2*: Matrix = identity(2) + I3*: Matrix = identity(3) + I4*: Matrix = identity(4) + I5*: Matrix = identity(5) + +## Calculates the determinant of a matrix through Laplace expansion +func det*(A: Matrix): float = + assert A.rows() == A.cols(), not_square + # Shortcut by formula + if A.rows() == 2: + return (A[0][0]*A[1][1]) - (A[0][1]*A[1][0]) + # Actual factual formula + for i in 0 ..< A.rows(): + var sub: Matrix + for a in 1 ..< A.rows(): + var row: RowVector + for b in 0 ..< A.cols(): + if b == i: continue + row.add(A[a][b]) + sub.add(row) + result += (-1.0)^i * A[0][i] * det(sub) + +# func spans(A, B: Matrix): bool = +# return false +# func subspace(a, b: Matrix): bool = +# return false + +# Matrix Properties + +## Calculate the dimension of a matrix by Gaussian Elimination +# func dim(a: Matrix): int = +# return 1 +# func rank(a: Matrix): int = +# return 1 +# func nullity(a: Matrix): int = +# return 1 diff --git a/lalge/src/types.nim b/lalge/src/types.nim new file mode 100644 index 0000000..321042a --- /dev/null +++ b/lalge/src/types.nim @@ -0,0 +1,8 @@ +type + # Any generic Vector is assumed to be a column vector. + Vector* = seq[float] + RowVector* = Vector # can this be distinct? + Matrix* = seq[RowVector] + +type + Radian* = float diff --git a/lalge/src/vectors.nim b/lalge/src/vectors.nim new file mode 100644 index 0000000..a9880d9 --- /dev/null +++ b/lalge/src/vectors.nim @@ -0,0 +1,57 @@ +import std/math, std/sequtils, std/sugar +import types + +const dim_mismatch = "Dimensional mismatch - check your vectors" + +## Vector addition +func `+`*(a, b: Vector): Vector = + assert a.len() == b.len(), dim_mismatch + result.setLen(a.len) + for i in 0 ..< a.len(): + result[i] = (a[i] + b[i]) + +## Vector subtraction +func `-`*(a, b: Vector): Vector = + assert a.len() == b.len(), dim_mismatch + result.setLen(a.len) + for i in 0 ..< a.len(): + result[i] = (a[i] - b[i]) + +## Vector dot product +func `*`*(a, b: Vector): float = + assert a.len() == b.len(), dim_mismatch + for i in 0 ..< a.len(): + result += a[i] * b[i] + +func dot*(a, b: Vector): float = + return a * b + +## Scalar-Vector multiplication +func `*`*(a: float, b: Vector): Vector = + return map(b, (x) => (a*x)) + +## Produce the length (in space) of a vector +func length*(a: Vector): float = + return sqrt(a * a) + +## Produce the number of elements in a vector +func size*(a: Vector): int = + return len(a) + +## Returns whether two vectors are orthagonal +func ortho*(a, b: Vector): bool = + return a * b == 0 + +## Produce the angle between two vectors, in radians +func angle*(a, b: Vector): Radian = + return arccos((a * b) / (a.length * b.length)) + +## Produce the cross product between two 3D vectors +func cross*(a, b: Vector): Vector = + assert a.len() == 3, "The first vector is not three-dimensional" + assert b.len() == 3, "The second vector is not three-dimensional" + return @[ + (a[1]*b[2]) - (b[1]*a[2]), + (a[2]*b[0]) - (b[2]*a[0]), + (a[0]*b[1]) - (b[0]*a[1]) + ] diff --git a/lalge/tests/nim.cfg b/lalge/tests/nim.cfg new file mode 100644 index 0000000..85bf6c4 --- /dev/null +++ b/lalge/tests/nim.cfg @@ -0,0 +1 @@ +--path:"../src/" diff --git a/lalge/tests/tests.nim b/lalge/tests/tests.nim new file mode 100644 index 0000000..e798c39 --- /dev/null +++ b/lalge/tests/tests.nim @@ -0,0 +1,144 @@ +import std/math +import types, vectors, matrices + +let A: Matrix = @[ + @[10.0, 2.0, 3.0, 5.0, 6.0], + @[8.0, 7.0, 6.0, 4.0, 3.0], + @[4.0, 6.0, 0.0, 9.0, 0.0], + @[6.0, 7.0, 9.0, 3.0, 9.0], + @[3.0, 0.0, 7.0, 9.0, 9.0], +] +let B: Matrix = @[ + @[1.0, 2.0, 3.0, 4.0, 5.0], + @[5.0, 3.0, 1.0, 22.0, 3.0], + @[5.0, 21.0, 4.0, 6.0, 3.0], + @[12.0, 1.0, 5.0, 0.0, 9.0], + @[6.0, 7.0, 1.0, 3.0, 5.0], +] + +let L: Matrix = @[ + @[1.0, 0.0, 0.0, 0.0, 0.0, 0.0], + @[-1.0, 1.0, 0.0, 0.0, 0.0, 0.0], + @[0.0, -1.0, 1.0, 0.0, 0.0, 0.0], + @[0.0, 0.0, -1.0, 1.0, 0.0, 0.0], + @[0.0, 0.0, 0.0, -1.0, 1.0, 0.0], + @[0.0, 0.0, 0.0, 0.0, -1.0, 1.0], +] +let U: Matrix = @[ + @[1.0, -1.0, 0.0, 0.0, 0.0, 0.0], + @[0.0, 1.0, -1.0, 0.0, 0.0, 0.0], + @[0.0, 0.0, 1.0, -1.0, 0.0, 0.0], + @[0.0, 0.0, 0.0, 1.0, -1.0, 0.0], + @[0.0, 0.0, 0.0, 0.0, 1.0, -1.0], + @[0.0, 0.0, 0.0, 0.0, 0.0, 1.0], +] + +let + P13: Matrix = @[ + @[0.0, 0.0, 1.0], + @[0.0, 1.0, 0.0], + @[1.0, 0.0, 0.0], + ] + P23: Matrix = @[ + @[1.0, 0.0, 0.0], + @[0.0, 0.0, 1.0], + @[0.0, 1.0, 0.0], + ] + +let foo = @[ + @[1.0, 1.0, 1.0], + @[3.0, 4.0, 5.0], + @[40.0, 51.0, 12.0] +] + +let eliminationMatrix: Matrix = @[ + @[1.0, 0.0, 0.0], + @[-2.0, 1.0, 0.0], + @[0.0, 0.0, 1.0] +] + +let + ones: Matrix = @[ + @[1.0, 2.0, 3.0], + @[4.0, 5.0, 6.0] + ] + twos: Matrix = @[ + @[7.0, 8.0], + @[9.0, 10.0], + @[11.0, 12.0] + ] + +# Addition and subtraction +assert I2 + I2 == 2 * I2 +assert I3-I3 == gen(3, 3, 0.0) + +# Num. of rows / columns +assert rows(I2) == 2 +assert I2.rows() == 2 +assert I2.rows == 2 +assert cols(I3) == 3 +assert I3.cols() == 3 +assert I3.cols == 3 + +# Identity Matrices +assert I1 == @[ + @[1.0], +] +assert I2 == @[ + @[1.0, 0.0], + @[0.0, 1.0], +] +assert I3 == @[ + @[1.0, 0.0, 0.0], + @[0.0, 1.0, 0.0], + @[0.0, 0.0, 1.0], +] +assert I4 == @[ + @[1.0, 0.0, 0.0, 0.0], + @[0.0, 1.0, 0.0, 0.0], + @[0.0, 0.0, 1.0, 0.0], + @[0.0, 0.0, 0.0, 1.0], +] +assert I5 == @[ + @[1.0, 0.0, 0.0, 0.0, 0.0], + @[0.0, 1.0, 0.0, 0.0, 0.0], + @[0.0, 0.0, 1.0, 0.0, 0.0], + @[0.0, 0.0, 0.0, 1.0, 0.0], + @[0.0, 0.0, 0.0, 0.0, 1.0], +] + +assert det(@[ + @[2.0, -1.0, 0.0, -1.0], + @[-1.0, 2.0, -1.0, 0.0], + @[0.0, -1.0, 2.0, -1.0], + @[-1.0, 0.0, -1.0, 2.0], +]) == 0 + +assert det(A * B) == (det(A) * det(B)) +assert det(I3) == 1 +assert det(2*I4) == 16 + +# Vectors + +let + alice: Vector = @[1.0, 1.0, 1.0] + bob = @[3.0, 4.0, 5.0] + carol = @[40.0, 51.0, 12.0] + dan: Vector = @[10.0, 20.0, 30.0, 40.0] + eve = @[2.0, 3.0, 4.0] + +assert alice + bob == bob + alice +assert alice + carol == carol + alice +assert carol - eve != eve - carol +assert alice.length() == sqrt(3.0) + +# echo transpose(ones) +# echo P13 * 8 +# echo P13 + P23 +# echo ones * twos +# assert col(eliminationMatrix, 0) == col(eliminationMatrix, 0) +# echo identity(0) +# echo identity(2) +# echo (1/0) +# echo L*U +# echo U*L |