concepts-parser
Nodejs module for extracting concepts from text.
A Concept is a part of a text that may be a Named entity. We use Concepts for learning new named-entities, for searching known entities, for identifying entity names(synonyms, abbreviations), etc.
Usage
JavaScript:
const parser = require('concepts-parser');
const concepts = parser.parse({ text: 'Some text', lang: 'ru', country: 'ru' });
TypeScript:
import { parse } from 'concepts-parser';
const concepts = parse({ text: 'Some text', lang: 'ru', country: 'ru '});
API
parse(context, options)
Finds concepts in a context.
-
context
(Object) required - Context-
text
(String) required - Text to find concepts; -
lang
(String) required - Text language, 2 chars code:en
,ru
; -
country
(String) optional - Context country:ru
,it
;
-
-
options
(Object) optional:-
mode
(String) optional - Can be identity or collect. Default: identity. identity mode excludes filters:start_word
,duplicate
andpartial
; -
filters
(String[]) optional - Ordered list of filters;
-
Valid filters
-
invalid_prefix
- deletes invalid prefixes; -
invalid
- exclude invalid concepts; -
partial
- exclude partial concepts; -
prefix
- add prefixes to concepts; -
suffix
- add suffixes to concepts; -
start_word
- exclude sentence start words; -
abbr
- finds concepts abbreviations; -
known
- finds known concepts; -
duplicate
- exclude duplicates; -
quote
- concats concepts in quotes:Teatrul Național "Mihai Eminescu"
;