JSPM

react-optimization-tools

Set of the fastest tools for optimizing the work of a React application

Install

npm install --save react-optimization-tools

Features

The set includes five algorithms: three functions (memoDeep, memoizeDeep, compareDeep) and two hooks (useMemoDeep, useCallbackDeep). They use deep comparison for equality. They are all based on a modified version of the qcompare function, which according to the author is the fastest comparison function in the industry.

1. memoDeep

If the props are not changed, the component is not rendered. Analog React.memo but with deep comparison.

Example of use:

import { memoDeep } from 'react-optimization-tools';

const Component = (props) => {...};
export default memoDeep(Component);

2. memoizeDeep

Remembers the last result of the function. Analog memoize-one but with deep comparison and better performance. Based on project deepMemoizeOnce.

Example of use:

• 2.1.

import { memoizeDeep } from 'react-optimization-tools';

const memoFunc = memoizeDeep((arg) => {...});// heavy computation function

const mapStateToProps = ({ data }) => {
    return {
    propValue: memoFunc(data)
  }}

• 2.2. (reselect)

import { createSelectorCreator } from 'reselect';
import { memoizeDeep } from 'react-optimization-tools';

const customSelectorCreator = createSelectorCreator(memoizeDeep);
// or
const customSelectorCreator = createSelectorCreator(memoizeDeep, {
  circular: true,
  strict: true
});

const selector = customSelectorCreator(
  state => state.a,
  state => state.b,
  (a, b) => {...}
);

3. compareDeep

Function of fast deep comparison of two values.

Example of use:

• 3.1. (reselect)

import { createSelectorCreator, defaultMemoize } from 'reselect';
import { compareDeep } from 'react-optimization-tools';

const createDeepEqualSelector = createSelectorCreator(
  defaultMemoize,
  compareDeep
)

const mySelector = createDeepEqualSelector(
  state => state.a,
  values => {...}
)

• 3.2. (React.memo)

import React from 'react';
import { compareDeep } from 'react-optimization-tools';

const MemoComponent = React.memo(Component, compareDeep);
// or
export default React.memo(Component, compareDeep);

• 3.3. (shouldComponentUpdate)

import React from 'react';
import { compareDeep } from 'react-optimization-tools';

class Component extends React.Component {
  shouldComponentUpdate(nextProps) {
    return !compareDeep(this.props, nextProps);
  }
  render() {...}
}

4. useMemoDeep

Hook for memoizing the result of the function with unchanged values of the input parameters. Similar to useMemo, but with a deep comparison.

Example of use:

import React, { useState } from 'react';
import { compareDeep } from 'react-optimization-tools';

const Component = () => {
const [ count, setCount ] = useState(0);

// ! attention is memoized result of the function !
const res = useMemoDeep(() => {...}, [count]);

return(...)
}

5. useCallbackDeep

Hook for memoizing a reference to a function with unchanged values of input parameters. Similar to useCallback, but with a deep comparison.

Example of use:

import React, { useState } from 'react';
import { useCallbackDeep } from 'react-optimization-tools';

const Component = () => {
const [ count, setCount ] = useState(0);

// ! attention is memoized function reference !
const funcRef = useCallbackDeep(() => {...}, [count]);

return(...)
}

Benchmarks

The algorithms are based on the speed of the modified algorithm of the qcompare project. The modification was related to the ability to work with React components. Therefore, the performance of this algorithm was compared with similar projects. Performance tests were used from these projects.

Nodejs - v14.9.0
Date - 14.10.2020

react-fast-compare project test result

--- speed tests: generic usage ---
qcompare x 140,332 ops/sec ±0.44% (92 runs sampled)
react-fast-compare x 97,891 ops/sec ±1.27% (88 runs sampled)
fast-deep-equal x 94,725 ops/sec ±1.04% (90 runs sampled)
lodash.isEqual x 16,635 ops/sec ±24.04% (91 runs sampled)
nano-equal x 93,555 ops/sec ±0.27% (90 runs sampled)
shallow-equal-fuzzy x 54,279 ops/sec ±0.43% (89 runs sampled)
  fastest: qcompare

--- speed tests: generic and react ---
qcompare x 64,123 ops/sec ±0.59% (88 runs sampled)
react-fast-compare x 44,744 ops/sec ±0.70% (88 runs sampled)
fast-deep-equal x 33,168 ops/sec ±0.37% (89 runs sampled)
lodash.isEqual x 3,085 ops/sec ±10.04% (87 runs sampled)
  fastest: qcompare

License

MIT Yuriy Khomenko