js-math-tools is a little library of (you guessed it) math tools for JS. It was built completely from scratch and has no other dependencies.
npm install --save https://github.com/jrc03c/js-math-toolsFor client-side use, attach the dist/js-math-tools.js file to your web page:
<script src="path/to/js-math-tools.js"></script>For use in Node:
let tools = require("js-math-tools")You can either pull individual functions out, like:
let { add } = require("js-math-tools")
add(3, 4) // 7Or, for easier access, you can "dump" all of the functions into the global scope:
require("js-math-tools").dump()
add(3, 4) // 7Most of the typical math functions (like sine, cosine, etc.) can take individual numeric values, arrays of numbers, or a combination of numbers and arrays of numbers. For example, all of the following are acceptable:
add(3, 4) // 7
add([2, 3, 4], [5, 6, 7]) // [7, 9, 11]
add([5, 10, 15], 10) // [15, 20, 25]
add(-7, [10, 11, 12]) // [3, 4, 5]There's a handy helper function called vectorize that can turn any arbitrary function into a function that operates on these mixed types of parameters (individual values or arrays of the same types of values). For example:
require("js-math-tools").dump()
function divide(a, b){
return a / b
}
divide = vectorize(divide)
divide(10, 2) // 5
divide([20, 30, 40], 10) // [2, 3, 4]
divide(24, [6, 8, 12]) // [4, 3, 2]
divide([20, 30, 40], [40, 30, 20]) // [0.5, 1, 2]Unlike some other JS math libraries, this library does not provide a wrapper around standard JS arrays. Therefore, pretty much everything is handled through functions, not objects. Here are some examples of things you might want to do with (n-dimensional) arrays.
require("js-math-tools").dump()
// make an empty array
let arr1 = ndarray(5)
// [undefined, undefined, undefined, undefined, undefined]
// make an empty array of a specific shape (e.g., with 3 rows and 2 columns)
let arr2 = ndarray([3, 2])
// [[undefined, undefined], [undefined, undefined], [undefined, undefined]]
// make an array of zeros
let arr3 = zeros(10)
// [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
// make an array of ones
let arr4 = ones(10)
// [1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
// make an array of 10,000 normally distributed values
let arr5 = normal(10000)
// make an array of random values (between 0 and 1) with 1,000 rows and 20 columns
let arr6 = random([1000, 20])
// add arrays together
add(normal([20, 30]), normal([20, 30]))
// get the dot-product of two 2-dimensional arrays
// NOTE: (n > 2)-dimensional arrays are not supported for the `dot` function!
dot(random([10, 20]), random([20, 5]))
// get statistics about an array of numbers
let x = normal(10000)
mean(x)
median(x)
mode(x)
std(x)
variance(x)
min(x)
max(x)
// get statistics about two arrays
let a = normal(10000)
let b = normal(10000)
correl(a, b)
cohensd(a, b)
covariance(a, b)
distance(a, b)
// get a range of numbers by step
range(4, 8)
// [4, 5, 6, 7]
range(0, 3, 0.5)
// [0, 0.5, 1, 1.5, 2, 2.5]
// get the shape of an array
shape(normal([4, 3, 2]))
// [4, 3, 2]
// transpose a 2-dimensional array
// NOTE: (n > 2)-dimensional arrays are not supported for the `transpose` function!
shape(transpose(normal([50, 10])))
// [10, 50]
// get the sum of all values in an array
sum(normal([20, 20]))
// get the set of values in an array
set([2, 2, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 3, 3, 3, 2, 2])
// [2, 3, 4]
// shuffle an array
shuffle(range(0, 10))
// [3, 2, 6, 5, 4, 9, 8, 1, 0, 7]
// sort an array by a function
sort(["b", "c", "a"], function(a, b){
if (a < b) return -1
if (a > b) return 1
return 0
})
// ["a", "b", "c"]
// get a slice of an array
let arr7 = [
[2, 3, 4],
[5, 6, 7]
]
let which = [null, 1]
slice(arr7, which)
// [3, 6]
which = [0, null]
slice(arr7, which)
// [2, 3, 4]
which = [1, 2]
slice(arr7, which)
// [7]
// flatten an array
flatten([[2, 3, 4], [5, 6, 7]])
// [2, 3, 4, 5, 6, 7]The PRNG (pseudo-random number generator) implemented in this library is technically a linear congruential generator, in case that matters to you. To seed the PRNG, pass a positive integer into the seed function. The random, normal, and shuffle functions can all be seeded. For example:
require("js-math-tools").dump()
seed(230498349)
random(5)
// [0.5713683813810349, 0.36495184898376465, 0.04452776908874512, 0.015275746583938599, 0.23413866385817528]
seed(230498349)
random(5)
// [0.5713683813810349, 0.36495184898376465, 0.04452776908874512, 0.015275746583938599, 0.23413866385817528]I recently added some helper classes to mimic the behavior of Python's pandas library. Note that these classes do not fully implement the functionality of the pandas DataFrames and Series!
Here are some of the things that you can do with Series:
require("js-math-tools").dump()
let series = new Series([1, 2, 3, 4, 5])
console.log(series.values) // [1, 2, 3, 4, 5]
console.log(series.name) // data
console.log(series.index) // ["row0", "row1", "row2", "row3", "row4"]
console.log(series.shape) // [5]
console.log(series.isEmpty()) // false
console.log(series.clear().values) // [undefined, undefined, undefined, undefined, undefined]
console.log(series.getSubsetByNames(["row0", "row2", "row4"]).values) // [1, 3, 5]
console.log(series.loc(["row0", "row2", "row4"]).values) // [1, 3, 5]
console.log(series.getSubsetByIndices([0, 2, 4]).values) // [1, 3, 5]
console.log(series.iloc([0, 2, 4]).values) // [1, 3, 5]
console.log(series.reverse().values) // [5, 4, 3, 2, 1]
console.log(series.reverse().index) // ["row4", "row3", "row2", "row1", "row0"]
console.log(series.reverse().resetIndex().index) // ["row0", "row1", "row2", "row3", "row4"]
console.log(series.copy().values) // [1, 2, 3, 4, 5]
console.log(series.apply((index, x) => x * 2).values) // [2, 4, 6, 8, 10]
series.values[0] = null
series.values[3] = null
console.log(series.values) // [null, 2, 3, null, 5]
console.log(series.dropMissing().values) // [2, 3, 5]
console.log(series.dropMissing().index) // ["row1", "row2", "row4"]And here are some of the things you can do with DataFrame:
require("js-math-tools").dump()
let x = new DataFrame([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
x.print()
/*
| col0 | col1 | col2
---- | ---- | ---- | ----
row0 | 1 | 2 | 3
row1 | 4 | 5 | 6
row2 | 7 | 8 | 9
*/
x = new DataFrame({a: range(0, 5), b: range(5, 10), c: range(10, 15)})
x.print()
/*
| a | b | c
---- | - | - | --
row0 | 0 | 5 | 10
row1 | 1 | 6 | 11
row2 | 2 | 7 | 12
row3 | 3 | 8 | 13
row4 | 4 | 9 | 14
*/
console.log(x.shape) // [5, 3]
console.log(x.isEmpty()) // false
console.log(x.clear().isEmpty()) // true
console.log(x.clear().shape) // [5, 3]
let subset = x.getSubsetByNames(["row2", "row4"], ["b", "a"]) // or x.loc(...)
subset.print()
/*
| b | a
---- | - | -
row2 | 7 | 2
row4 | 9 | 4
*/
subset = x.getSubsetByIndices([0, 2, 4], null) // or x.iloc(...)
subset.print()
/*
| a | b | c
---- | - | - | --
row0 | 0 | 5 | 10
row2 | 2 | 7 | 12
row4 | 4 | 9 | 14
*/
x.transpose().print() // or x.T.print()
/*
| row0 | row1 | row2 | row3 | row4
- | ---- | ---- | ---- | ---- | ----
a | 0 | 1 | 2 | 3 | 4
b | 5 | 6 | 7 | 8 | 9
c | 10 | 11 | 12 | 13 | 14
*/
let xCopy = x.copy()
console.log(isEqual(x, xCopy)) // true
console.log(x === xCopy) // false
x = x.assign({d: range(15, 20)})
x.print()
/*
| a | b | c | d
---- | - | - | -- | --
row0 | 0 | 5 | 10 | 15
row1 | 1 | 6 | 11 | 16
row2 | 2 | 7 | 12 | 17
row3 | 3 | 8 | 13 | 18
row4 | 4 | 9 | 14 | 19
*/
x.apply(col => reverse(col), 0) // axis = 0
x.apply(row => reverse(row), 1) // axis = 1
x.dropMissing() // drop any row that contains null / undefined values
x.dropMissing(0) // same
x.dropMissing(1) // drop any column that contains null / undefined values
x.dropMissing(0, "any") // drop any row that contains null / undefined values
x.dropMissing(0, "all") // drop a row only if all of its values are null / undefined
x.dropMissing(0, null, 5) // drop a row only if it contains at least 5 null / undefined values
x.dropColumns(["a", "c"])
x.dropRows(["row4", "row2"])
console.log(x.toObject())
/*
{
row0: { a: 0, b: 5, c: 10, d: 15 },
row1: { a: 1, b: 6, c: 11, d: 16 },
row2: { a: 2, b: 7, c: 12, d: 17 },
row3: { a: 3, b: 8, c: 13, d: 18 },
row4: { a: 4, b: 9, c: 14, d: 19 }
}
*/
x.toCSV("data.csv")Finally, you can read CSV files like this:
DataFrame.fromCSV("path/to/some.csv", {
hasHeaderRow: true,
hasIndexColumn: false,
}).then(data => {
// do something with `data`
})Honestly, this library's ability to read CSVs isn't super robust, so if you need something more configurable or flexible, try csvtojson!