@nlpjs/lang-it
v4.26.1
Published
Core
Downloads
31,293
Readme
@nlpjs/lang-it
TABLE OF CONTENTS
- Installation
- Normalization
- Tokenization
- Identify if a word is an italian stopword
- Remove stopwords from an array of words
- Change the stopwords dictionary
- Stemming word by word
- Stemming an array of words
- Normalizing, Tokenizing and Stemming a sentence
- Remove stopwords when stemming a sentence
- Sentiment Analysis
- Example of usage on a classifier
- Contributing
- Contributors
- Code of Conduct
- Who is behind it
- License
Installation
You can install @nlpjs/lang-it:
npm install @nlpjs/lang-it
Normalization
Normalization of a text converts it to lowercase and remove decorations of characters.
const { NormalizerIt } = require('@nlpjs/lang-it');
const normalizer = new NormalizerIt();
const input = 'Questo dòvrebbe essere normalizzato';
const result = normalizer.normalize(input);
console.log(result);
// output: questo dovrebbe essere normalizzato
Tokenization
Tokenization splits a sentence into words.
const { TokenizerIt } = require('@nlpjs/lang-it');
const tokenizer = new TokenizerIt();
const input = 'Questo dovrebbe essere tokenizzato';
const result = tokenizer.tokenize(input);
console.log(result);
// output: [ 'Questo', 'dovrebbe', 'essere', 'tokenizzato' ]
Tokenizer can also normalize the sentence before tokenizing, to do that provide a true as second argument to the method tokenize
const { TokenizerIt } = require('@nlpjs/lang-it');
const tokenizer = new TokenizerIt();
const input = 'Questo dovrebbe essere tokenizzato';
const result = tokenizer.tokenize(input, true);
console.log(result);
// output: [ 'questo', 'dovrebbe', 'essere', 'tokenizzato' ]
Identify if a word is an italian stopword
Using the class StopwordsIt you can identify if a word is an stopword:
const { StopwordsIt } = require('@nlpjs/lang-it');
const stopwords = new StopwordsIt();
console.log(stopwords.isStopword('uno'));
// output: true
console.log(stopwords.isStopword('sviluppatore'));
// output: false
Remove stopwords from an array of words
Using the class StopwordsIt you can remove stopwords form an array of words:
const { StopwordsIt } = require('@nlpjs/lang-it');
const stopwords = new StopwordsIt();
console.log(stopwords.removeStopwords(['ho', 'visto', 'uno', 'sviluppatore']));
// output: [ 'visto', 'sviluppatore' ]
Change the stopwords dictionary
Using the class StopwordsIt you can restart it dictionary and build it from another set of words:
const { StopwordsIt } = require('@nlpjs/lang-it');
const stopwords = new StopwordsIt();
stopwords.dictionary = {};
stopwords.build(['ho', 'visto']);
console.log(stopwords.removeStopwords(['ho', 'visto', 'uno', 'sviluppatore']));
// output: [ 'uno', 'sviluppatore' ]
Stemming word by word
An stemmer is an algorithm to calculate the stem (root) of a word, removing affixes.
You can stem one word using method stemWord:
const { StemmerIt } = require('@nlpjs/lang-it');
const stemmer = new StemmerIt();
const input = 'svilupp';
console.log(stemmer.stemWord(input));
// output: program
Stemming an array of words
You can stem an array of words using method stem:
const { StemmerIt } = require('@nlpjs/lang-it');
const stemmer = new StemmerIt();
const input = ['ho', 'visto', 'uno', 'sviluppatore'];
console.log(stemmer.stem(input));
// outuput: [ 'ho', 'vist', 'uno', 'svilupp' ]
Normalizing, Tokenizing and Stemming a sentence
As you can see, stemmer does not do internal normalization, so words with uppercases will remain uppercased. Also, stemmer works with lowercased affixes, so sviluppatore will be stemmed as svilupp but SVILUPPATORE will not be changed.
You can tokenize and stem a sentence, including normalization, with the method tokenizeAndStem:
const { StemmerIt } = require('@nlpjs/lang-it');
const stemmer = new StemmerIt();
const input = 'Ho visto uno SVILUPPATORE';
console.log(stemmer.tokenizeAndStem(input));
// output: [ 'ho', 'vist', 'uno', 'svilupp' ]
Remove stopwords when stemming a sentence
When calling tokenizeAndStem method from the class StemmerIt, the second parameter is a boolean to set if the stemmer must keep the stopwords (true) or remove them (false). Before using it, the stopwords instance must be set into the stemmer:
const { StemmerIt, StopwordsIt } = require('@nlpjs/lang-it');
const stemmer = new StemmerIt();
stemmer.stopwords = new StopwordsIt();
const input = 'Ho visto uno sviluppatore';
console.log(stemmer.tokenizeAndStem(input, false));
// output: [ 'vist', 'svilupp' ]
Sentiment Analysis
To use sentiment analysis you'll need to create a new Container and use the plugin LangIt, because internally the SentimentAnalyzer class try to retrieve the normalizer, tokenizer, stemmmer and sentiment dictionaries from the container.
const { Container } = require('@nlpjs/core');
const { SentimentAnalyzer } = require('@nlpjs/sentiment');
const { LangIt } = require('@nlpjs/lang-it');
(async () => {
const container = new Container();
container.use(LangIt);
const sentiment = new SentimentAnalyzer({ container });
const result = await sentiment.process({
locale: 'it',
text: 'amore per i gatti',
});
console.log(result.sentiment);
})();
// output:
// {
// score: 0.25,
// numWords: 4,
// numHits: 2,
// average: 0.0625,
// type: 'pattern',
// locale: 'it',
// vote: 'positive'
// }
The output of the sentiment analysis includes:
- score: final score of the sentence.
- numWords: total words of the sentence.
- numHits: total words of the sentence identified as having a sentiment score.
- average: score divided by numWords
- type: type of dictionary used, values can be afinn, senticon or pattern.
- locale: locale of the sentence
- vote: positive if score greater than 0, negative if score lower than 0, neutral if score equals 0.
Example of usage on a classifier
const { containerBootstrap } = require('@nlpjs/core');
const { Nlp } = require('@nlpjs/nlp');
const { LangIt } = require('@nlpjs/lang-it');
(async () => {
const container = await containerBootstrap();
container.use(Nlp);
container.use(LangIt);
const nlp = container.get('nlp');
nlp.settings.autoSave = false;
nlp.addLanguage('it');
// Adds the utterances and intents for the NLP
nlp.addDocument('it', 'Addio per ora', 'greetings.bye');
nlp.addDocument('it', 'arrivederci e stai attento', 'greetings.bye');
nlp.addDocument('it', 'molto bene a dopo', 'greetings.bye');
nlp.addDocument('it', 'devo andare', 'greetings.bye');
nlp.addDocument('it', 'ciao', 'greetings.hello');
// Train also the NLG
nlp.addAnswer('it', 'greetings.bye', 'fino alla prossima volta');
nlp.addAnswer('it', 'greetings.bye', 'A presto!');
nlp.addAnswer('it', 'greetings.hello', 'Ciao, come stai');
nlp.addAnswer('it', 'greetings.hello', 'Saluti!');
await nlp.train();
const response = await nlp.process('it', 'devo andare');
console.log(response);
})();
Contributing
You can read the guide of how to contribute at Contributing.
Contributors
Made with contributors-img.
Code of Conduct
You can read the Code of Conduct at Code of Conduct.
Who is behind it?
This project is developed by AXA Group Operations Spain S.A.
If you need to contact us, you can do it at the email opensource@axa.com
License
Copyright (c) AXA Group Operations Spain S.A.
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.