This commit is contained in:
Ben Calderon 2025-02-12 20:09:14 +00:00 committed by GitHub
commit 288c769377
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 62 additions and 1 deletions

View file

@ -325,7 +325,8 @@
"Unescape string",
"Pseudo-Random Number Generator",
"Sleep",
"File Tree"
"File Tree",
"N-gram"
]
},
{

View file

@ -0,0 +1,60 @@
/**
* @author benjcal [benj.calderon@gmail.com]
* @copyright Crown Copyright 2024
* @license Apache-2.0
*/
import Operation from "../Operation.mjs";
import {JOIN_DELIM_OPTIONS} from "../lib/Delim.mjs";
/**
* ngram operation
*/
class Ngram extends Operation {
/**
* Ngram constructor
*/
constructor() {
super();
this.name = "N-gram";
this.module = "Default";
this.description = "Extracts n-grams from the input text. N-grams are contiguous sequences of n characters from a given text sample.";
this.infoURL = "https://wikipedia.org/wiki/N-gram";
this.inputType = "string";
this.outputType = "string";
this.args = [
{
name: "N-gram size",
type: "number",
value: 3
},
{
"name": "Join delimiter",
"type": "editableOptionShort",
"value": JOIN_DELIM_OPTIONS
}
];
}
/**
* @param {string} input
* @param {Object[]} args
* @returns {string}
*/
run(input, args) {
const nGramSize = args[0],
joinDelim = args[1];
const ngrams = [];
for (let i = 0; i <= input.length - nGramSize; i++) {
ngrams.push(input.slice(i, i + nGramSize));
}
return ngrams.join(joinDelim);
}
}
export default Ngram;