@apto-space/moving-ngrams v0.0.1

moving-ngrams

A utility package for generating all possible n-gram combinations from a sequence of words, useful for search engine query analysis and natural language processing tasks.

Description

moving-ngrams helps you analyze text by generating all possible meaningful combinations of consecutive words. This is particularly useful for:

• Search engine query analysis
• Automatic keyphrase extraction
• Faceted search implementation
• Natural language understanding

Installation

# Using bun
bun add moving-ngrams

# Using npm
npm install moving-ngrams

# Using yarn
yarn add moving-ngrams

Usage

import { movingNgrams } from 'moving-ngrams';

// Example 1: Simple phrase analysis
const result = movingNgrams(['hello', 'world']);
console.log(result);
/* Output:
[
  [['hello'], ['world']],      // Individual words
  [['hello', 'world']]         // Combined phrase
]
*/

// Example 2: Search query analysis
const searchQuery = movingNgrams(['elon', 'musk', 'san', 'francisco']);
console.log(searchQuery);
/* Output will include all possible combinations like:
- [['elon'], ['musk'], ['san'], ['francisco']]
- [['elon', 'musk'], ['san'], ['francisco']]
- [['elon'], ['musk', 'san'], ['francisco']]
- [['elon'], ['musk'], ['san', 'francisco']]
... and more combinations
*/

Use Cases

1. Search Query Analysis

When users enter a search query like "Elon Musk San Francisco", the package helps you analyze all possible meaningful combinations:

• Individual terms: "Elon", "Musk", "San", "Francisco"
• Compound terms: "Elon Musk", "San Francisco"
• Mixed combinations: "Elon Musk", "San Francisco", "Elon", "Musk San Francisco", etc.

This is particularly useful for:

• Improving search relevance
• Understanding user intent
• Supporting multi-term matching

2. Keyphrase Extraction

When analyzing text content, you can use this to:

• Extract potential keyphrases
• Identify compound terms
• Generate search indices

API

function movingNgrams(words: string[]): string[][][]

• Input: Array of words
• Output: Array of all possible combinations where each combination is an array of word groups

License

MIT

This project was created using Bun v1.1.41.