string-sameness

Generate scores that represents how similar two strings are based on different string comparison algorithms.

Scores from all used algorithms are averaged and then weighted by the length of the content being compared (more weight for longer content).

The sameness is then given a score of 0 to 100.

• 0 => Totally unique pieces of content
• 100 => Identical content

Install/Usage

npm install @foxxmd/string-sameness
Copy

import {stringSameness} from '@foxxmd/string-sameness';

const result =  stringSameness('This is one sentence', 'This is another sentence');
console.log(result);
// {
//     "strategies": {
//         "dice": {
//             "rawScore": 0.6666,
//             "score": 66.66
//         },
//         "leven": {
//             "rawScore": 5,
//             "distance": 5,
//             "score": 79.16
//         },
//         "cosine": {
//             "rawScore": 0.75,
//             "score": 75
//         }
//     },
//     "highScore": 73.61,
//     "highScoreWeighted": 83.58
// }
Copy

Options

An optional third argument can be provided to stringSameness to customize how strings are normalized before comparison and what strategies are used for comparison.

Strategies

Pass a list of ComparisonStrategy objects using {strategies: []} to define which string comparisons should be performed on the given strings.

The average of the scores from all passed strategies is returned as highScore (and highScoreWeighted) from stringSameness()

When no strategies are explicitly passed a default set of strategies is used, found in import {defaultStrategies} from @foxxmd/string-sameness;:

• Dice's Coefficient in diceSimilarities.ts
• Cosine Similarity in cosineSimilarities.ts
• Levenshtein Distance in levenSimilarities.ts

Strategies can be accessed individually using import {strategies} from @foxxmd/string-sameness

Bring Your Own Strategy

Use your own strategy by creating an object that conforms to ComparisonStrategy:

export interface ComparisonStrategy {
    /**
     * The name of this strategy
     * */
    name: string
    /**
     * A function that accepts two string arguments and returns a number between 0 and 100 signifying how closely similar the strings are:
     * 0 => not similar at all
     * 100 => identical
     * */
    strategy: (strA: string, strB: string) => number
    /**
     * An optional function that accepts two string arguments and returns whether this strategy should be used
     * */
    isValid?: (strA: string, strB: string) => boolean
}
Copy

Example of using your own strategy with the defaults:

import {stringSameness} from "@foxxmd/string-sameness";
import {ComparisonStrategy, levenStrategy, cosineStrategy, diceStrategy} from "@foxxmd/string-sameness/strategies";

const myStrat: ComparisonStrategy = {
    name: 'MyCoolStrat',
    strategy: (valA: string, valB: string) => {
        const a = valA.concat(valB);
        return a.length;
    },
}
const strats = [
    levenStrategy,
    cosineStrategy,
    diceStrategy,
    myStrat
]

const result = stringSameness('This is one sentence', 'This is another sentence', {strategies: strats});
Copy

Normalization

Pass a list of functions using {transforms: []} to transform the strings before comparison. When not explicitly provided a default set of functions is applied to normalize the strings (to remove trivial differences):

• normalize unicode EX convert Ö => O

• convert to lowercase

• trim (remove whitespace at beginning/end)

• remove non-alphanumeric characters (punctuation and newlines)

• replace any instances of 2 or more consecutive whitespace with 1 whitespace

• The default set of transformer functions is exported as import {strDefaultTransforms} from @foxxmd/string-sameness;

• All built-in transformers can be found at import {transforms} from @foxxmd/string-sameness;

Example of supplying your own transform functions:

import {stringSameness, defaultStrCompareTransformFuncs} from '@foxxmd/string-sameness';

const myFuncs = [
    ...defaultStrCompareTransformFuncs,
    // replace all vowels with the letter e
    (str) => str.replace(/[aeiou]/ig, 'e')
]

const result =  stringSameness('This is one sentence', 'This is another sentence', {transforms: myFuncs});
Copy

Token Re-ordering

If tokens (word) ordering in the strings is not important you can choose to have string-sameness attempt to re-order all words before comparing sameness. This makes comparison scores much closer to "absolute sameness in all characters within string". EX:

• this is correct order
• order correct this is

Scores 60 without reordering

Scores 100 with reordering

Behavior caveats:

• The second string argument is reordered to match the first string argument
• If the second string is longer than the first than any non-matched words are concatenated to the end of the re-ordered string in the same order they were found

To use:

import {stringSameness} from '@foxxmd/string-sameness';

const res = stringSameness(strA, strB, {reorder: true});
Copy

Factory

For convenience, a factory function is also provided:

import {createStringSameness, strategies} from "@foxxmd/string-sameness";
import {myTransforms, myStrats} from './util';

const {levenStrategy} = strategies;

// sets the default object to used with the third argument for `stringSameness`
const myCompare = createStringSameness({transforms: myTransforms, strategies: [levenStrategy, ...myStrats]});

// uses myTransforms and myStrats
const plainResult = myCompare('This is one sentence', 'This is another sentence');

// override your defaults using the third argument like normal
const overrideResults = myCompare('This is one sentence', 'This is another sentence', {strategies: [levenStrategy]});
Copy