This is a small and fast library for nlp sentiment analysis which supports any language, or any custom vocabulary. It can also use intensifiers with positive or negative words adding more contrast.
Sentiment uses simple algorithms and basic optimisations to run these tasks in a fast and easy manner. It does not have any dependencies or other libraries and is fully tested to give you confidence.
Test coverage is always kept at > 99%
.
This project is targeting the Node v18 LTS
. Target for ts
is set to ES3
to support all browsers with lower versions.
- Have
node
installed with at leastv18+
- Install required packages with
npm ci
Installation is simple:
npm i nlp-sentiment
You have access to all modules exported from the project, but in reality you probably need only one of them.
Here is an example how to use this library in your project:
// import it
import nlpSentiment from "nlp-sentiment";
// create a sentiment object with vocabulary
const sentiment = nlpSentiment()
.withPositiveWords(["good", "brilliant", "amazing"])
.withNegativeWords(["bad", "gross", "horrible"])
.withIntensifierWords(["very"])
.withNegationWords(["not"]);
// and now you are ready to use it for your sentences
const sentencesSentiment = sentiment.analyse([sentence]);
The input interfaces are:
export interface Sentence {
index: number;
tokens: Token[];
}
and for token:
export interface Token {
index: number;
lemma: string;
depType?: string;
depIndex?: number;
}
Then the output would be very similar:
export interface SentenceResult extends Omit<Sentence, "tokens"> {
tokens: TokenResult[];
sentiment: number;
attitude: "positive" | "negative";
}
and token is:
export interface TokenResult extends Token {
sentiment: number;
}
This project has just been made public, thus the documentation will be updated soon to cover all the aspects.