What is this?
Natural is a general-purpose natural language processing library for Node.js that bundles tokenizers, stemmers, classifiers and other NLP utilities into one package.
Natural is a general-purpose natural language processing library for Node.js that bundles tokenizers, stemmers, classifiers and other NLP utilities into one package.Is this a framework or just a library I import?It's a plain library you require into your Node.js code, not a standalone app or a framework that dictates your project structure.
It's a plain library you require into your Node.js code, not a standalone app or a framework that dictates your project structure.How do I add it to a Node.js project?Install the 'natural' package with npm and require it in your code; the project's own build uses TypeScript compiled to CommonJS and ESM bundles via Rollup.
Install the 'natural' package with npm and require it in your code; the project's own build uses TypeScript compiled to CommonJS and ESM bundles via Rollup.Do I need anything besides Node.js to use it?No external services are required for core use; the package.json only requires Node.js 0.4.10 or newer, and it optionally depends on MongoDB, PostgreSQL, Redis or Memcached clients if you want to persist trained classifiers.
No external services are required for core use; the package.json only requires Node.js 0.4.10 or newer, and it optionally depends on MongoDB, PostgreSQL, Redis or Memcached clients if you want to persist trained classifiers.Are there any licensing quirks to know about?While the core code is MIT, the bundled WordNet data carries Princeton's own license and the German Porter stemmer is under a separate BSD license, so mixed terms apply depending on which parts you use.
While the core code is MIT, the bundled WordNet data carries Princeton's own license and the German Porter stemmer is under a separate BSD license, so mixed terms apply depending on which parts you use.What NLP tasks does it support out of the box?Based on its bundled dependencies it covers sentiment scoring via the AFINN-165 lexicon, text classification, multilingual stopword removal, and vector-math utilities for similarity calculations, alongside standard tokenizing and stemming.
Based on its bundled dependencies it covers sentiment scoring via the AFINN-165 lexicon, text classification, multilingual stopword removal, and vector-math utilities for similarity calculations, alongside standard tokenizing and stemming.Can trained classifiers be saved and reused?Yes, the library ships with optional storage adapters for MongoDB, PostgreSQL, Redis and Memcached so classifiers can be persisted and reloaded across runs.
Yes, the library ships with optional storage adapters for MongoDB, PostgreSQL, Redis and Memcached so classifiers can be persisted and reloaded across runs.Does it include WordNet for word lookups?It bundles the WordNet 3.0 lexical database under Princeton's license, giving it dictionary-style word relationship lookups alongside its own NLP tools.
It bundles the WordNet 3.0 lexical database under Princeton's license, giving it dictionary-style word relationship lookups alongside its own NLP tools.How old can my Node.js version be?The package allows Node.js 0.4.10 or newer, so it runs on virtually any modern Node release without version headaches.
The package allows Node.js 0.4.10 or newer, so it runs on virtually any modern Node release without version headaches.What are similar projects?nltk, mongoose, HanLP, spaCy, doccano, compromise, flexsearch, tiktoken
nltk, mongoose, HanLP, spaCy, doccano, compromise, flexsearch, tiktokenWhich topics is it in?sentiment-analysis, tokenization, integration, text-processing, classification, npm-package, natural-language-processing, mongodb, node, redis, nlp, postgresql
sentiment-analysis, tokenization, integration, text-processing, classification, npm-package, natural-language-processing, mongodb, node, redis, nlp, postgresql