ndarray-ops

What is ndarray-ops?

The ndarray-ops npm package provides a collection of high-performance operations for manipulating ndarrays (N-dimensional arrays). It is designed to work efficiently with large datasets and supports a variety of mathematical and logical operations.

What are ndarray-ops's main functionalities?

Element-wise Operations

This feature allows you to perform element-wise operations such as addition, subtraction, multiplication, and division on ndarrays. The code sample demonstrates how to add two 2x2 matrices element-wise.

const ndarray = require('ndarray');
const ops = require('ndarray-ops');

let a = ndarray(new Float32Array([1, 2, 3, 4]), [2, 2]);
let b = ndarray(new Float32Array([5, 6, 7, 8]), [2, 2]);
let c = ndarray(new Float32Array(4), [2, 2]);

ops.add(c, a, b);
console.log(c.data); // Output: Float32Array [ 6, 8, 10, 12 ]

Scalar Operations

This feature allows you to perform operations involving a scalar value and an ndarray. The code sample demonstrates how to multiply each element of a 2x2 matrix by a scalar value (2).

const ndarray = require('ndarray');
const ops = require('ndarray-ops');

let a = ndarray(new Float32Array([1, 2, 3, 4]), [2, 2]);

ops.mulseq(a, 2);
console.log(a.data); // Output: Float32Array [ 2, 4, 6, 8 ]

Logical Operations

This feature allows you to perform logical operations such as less than, greater than, and equal to on ndarrays. The code sample demonstrates how to compare two 2x2 matrices element-wise to check if elements in the first matrix are less than those in the second matrix.

const ndarray = require('ndarray');
const ops = require('ndarray-ops');

let a = ndarray(new Float32Array([1, 2, 3, 4]), [2, 2]);
let b = ndarray(new Float32Array([4, 3, 2, 1]), [2, 2]);
let c = ndarray(new Float32Array(4), [2, 2]);

ops.lt(c, a, b);
console.log(c.data); // Output: Float32Array [ 1, 1, 0, 0 ]

Other packages similar to ndarray-ops

mathjs

Math.js is an extensive math library for JavaScript and Node.js. It provides a wide range of mathematical functions and supports multi-dimensional arrays. Compared to ndarray-ops, math.js offers a broader set of functionalities but may not be as optimized for performance in specific ndarray operations.

numeric

Numeric is a library for numerical computations in JavaScript. It provides support for matrix and vector operations, similar to ndarray-ops. However, ndarray-ops is more specialized for high-performance operations on ndarrays, while Numeric offers a more general approach to numerical computations.

ndarray

Ndarray is a module for creating and manipulating N-dimensional arrays in JavaScript. While it provides the basic structure for ndarrays, it does not include the extensive set of operations that ndarray-ops offers. Ndarray-ops can be seen as a complementary package to ndarray, providing the necessary operations to manipulate the arrays efficiently.

README

ndarray-ops

A collection of common mathematical operations for ndarrays. Implemented using cwise

Usage

First, install the library using npm:

npm install ndarray-ops

Then you can import the library by doing:

var ops = require("ndarray-ops")

Then you can use the functions as in the following example:

//First, import libraries
var ndarray = require("ndarray")
  , ops = require("ndarray-ops")


//Next, create some arrays
var a = ndarray.zeros([128,128])
  , b = ndarray.zeros([128,128])
  , c = ndarray.zeros([128,128])

//Initialize b with some random numbers:
ops.random(b)

//Set c to a constant 1
ops.assigns(c, 1.0)

//Add b and c, store result in a:
ops.add(a, b, c)

//Multiply a by 0.5 in place
ops.mulseq(a, 0.5)

//Print some statistics about a:
console.log(
  "inf(a) = ", ops.inf(a),
  "sup(a) = ", ops.sup(a),
  "argmin(a) = ", ops.argmin(a),
  "argmax(a) = ", ops.argmax(a),
  "norm1(a) = ", ops.norm1(a))

Conventions

This library implements component-wise operations for all of the operators and Math.* functions in JS, along with a few commonly used aggregate operations. Most of the functions in the library work by applying some symmetric binary operator to a pair of arrays. You call them like this:

ops.add(dest, arg1, arg2)

Which translates into code that works (approximately) like this:

for(var i=0; i<dest.shape[0]; ++i) {
  dest[i] = arg1[i] + arg2[i]
}

It is up to you to specify where the result gets store. This library does not create new arrays for you to avoid performing expensive intermediate allocations. There are also a few other variations:

ops.addeq(dest, arg1)

Operators with the -eq suffix perform an assignment.

for(var i=0; i<dest.shape[0]; ++i) {
  dest[i] += arg1[i]
}

ops.adds(dest, arg1, 1.0)

The -s suffix denotes scalar/broadcast operations; so the above would translate to:

for(var i=0; i<dest.shape[0]; ++i) {
  dest[i] = arg1[i] + 1.0
}

ops.addseq(dest, 1.0)

The -seq suffix is basically the combination of the above, and translates to:

for(var i=0; i<dest.shape[0]; ++i) {
  dest[i] += 1.0
}

The following operators follow this rule:

• add[,s,eq,seq] - Addition, +
• sub[,s,eq,seq] - Subtraction, -
• mul[,s,eq,seq] - Multiplication, *
• div[,s,eq,seq] - Division, /
• mod[,s,eq,seq] - Modulo, %
• band[,s,eq,seq] - Bitwise And, &
• bor[,s,eq,seq] - Bitwise Or, &
• bxor[,s,eq,seq] - Bitwise Xor, ^
• lshift[,s,eq,seq] - Left shift, <<
• rshift[,s,eq,seq] - Signed right shift, >>
• rrshift[,s,eq,seq] - Unsigned right shift, >>>
• lt[,s,eq,seq] - Less than, <
• gt[,s,eq,seq] - Greater than, >
• leq[,s,eq,seq] - Less than or equal, <=
• geq[,s,eq,seq] - Greater than or equal >=
• eq[,s,eq,seq] - Equals, ===
• neq[,s,eq,seq] - Not equals, !==
• and[,s,eq,seq] - Boolean And, &&
• or[,s,eq,seq] - Boolean Or, ||
• max[,s,eq,seq] - Maximum, Math.max
• min[,s,eq,seq] - Minimum, Math.min

Special Cases

There are a few corner cases that follow slightly different rules. These can be grouped using the following general categories:

Assignment

There are two assignment operators:

• assign
• assigns

op.assign(dest, src) copies one array into another, while op.assigns(dest, val) broadcasts a scalar to all elements of an array.

Copying

You can also copy ndarrays using the clone() method. For example:

var n = ops.clone(a)

Nullary operators

Nullary operators only take on argument for the array they are assigning to, and don't have any variations. Currently there is only one of these:

• random - Sets each element of an array to a random scalar between 0 and 1, Math.random()

Unary operators

Unary operators have one of two forms, they can be written as either:

op.abs(dest, arg)

Or:

op.abseq(dest)

The former version sets dest = |arg|, while in the latter the operation is applied in place. ndarray-ops exposes the following unary operators:

• not[,eq] - Boolean not, !
• bnot[,eq] - Bitwise not, ~
• neg[,eq] - Negative, -
• recip[,eq] - Reciprocal, 1.0/
• abs[,eq] - Absolute value, Math.abs
• acos[,eq] - Inverse cosine, Math.acos
• asin[,eq] - Inverse sine, Math.asin
• atan[,eq] - Inverse tangent, Math.atan
• ceil[,eq] - Ceiling, Math.ceil
• cos[,eq] - Cosine, Math.cos
• exp[,eq] - Exponent, Math.exp
• floor[,eq] - Floor, Math.floor
• log[,eq] - Logarithm, Math.log
• round[,eq] - Round, Math.round
• sin[,eq] - Sine, Math.sin
• sqrt[,eq] - Square root, Math.sqrt
• tan[,eq] - Tangent, Math.tan

Non-symmetric binary operators

There are also a few non-symmetric binary operators. These operators have an extra suffix op which flips the order of the arguments. There are only two of these:

• atan2[,s,eq,seq,op,sop,opeq,sopeq]
• pow[,s,eq,seq,op,sop,opeq,sopeq]

Map-reduce (aggregate) operators

Finally, there are aggregate operators that take an array as input and compute some aggregate result or summary. These functions don't have any special suffixes and all of them take a single array as input.

• any - Check if any element of the array is truthy
• all - Checks if any element of the array is falsy
• sum - Sums all elements of the array
• prod - Multiplies all elements of the array
• norm2squared - Computes the squared L2 norm
• norm2 - Computes the L2 norm
• norminf - Computes the L-infinity norm
• norm1 - Computs the L1 norm
• sup - Max element in array
• inf - Min element in array
• argmin - Index of min element
• argmax - Index of max element

Credits

(c) 2013 Mikola Lysenko. BSD