winkjs / wink-nlp-utils Goto Github PK

1. Link to GRAYPE
2. Documentation link text updation
3. Make wink as wink
4. Make npm as npm

Code:

const nlp = require('wink-nlp-utils');

nlp.string.extractPersonsName('George Washington was an American military officer, statesman, and Founding Father who served as the first president of the United States from 1789 to 1797.')

Expected behavior: return George Washington

Actual Behavior: returns George Washington was an American military officer statesman and Founding Father who served as the first president of the United States from to

Reproduction: https://runkit.com/embed/glyj5qvv8rgi

1. mocha

The current list of abbreviations in the sentence boundary detection code is quite limited. Could we have the option of passing in other abbreviations to be considered so that we do not have sentences broken up in incorrect places?

Like this:

composeCorpus( [
    ['I'], 
    ['am having', 'have'], 
    ['a'], 
    ['problem', 'question']
]);

This allows to feed it with a pre cached dictionary without having to format strings to feed it.

using winkNLP' SBD

using winkNLP's tokenizer

Noticed some function names are shortened in the API, like http://winkjs.org/wink-nlp-utils/#tokensbow
Could be more readable to write out the full name, eg bag_of_words

I'm referring to Word Embeddings.

The ngram is for string, while the bigram is for tokens. Wondered if tokens could have a more generic ngram function where you can feed in the maximum gram as a parameter?

http://winkjs.org/wink-nlp-utils/#tokensbigrams

So we could have sg like this https://github.com/nlp-compromise/nlp-ngram

https://badge.fury.io/for/js/wink-nlp-utils

https://david-dm.org/decisively/wink-nlp-utils