parse-latin

parse-latin

Natural language parser, for Latin-script languages, that produces nlcst.

Contents

• What is this?
• When should I use this?
• Install
• Use
• API
  • ParseLatin()
• Algorithm
• Types
• Compatibility
• Security
• Related
• Contribute
• License

What is this?

This package exposes a parser that takes Latin-script natural language and produces a syntax tree.

When should I use this?

If you want to handle natural language as syntax trees manually, use this.

Alternatively, you can use the retext plugin retext-latin, which wraps this project to also parse natural language at a higher-level (easier) abstraction.

Whether Old-English (“þā gewearþ þǣm hlāforde and þǣm hȳrigmannum wiþ ānum penninge”), Icelandic (“Hvað er að frétta”), French (“Où sont les toilettes?”), this project does a good job at tokenizing it.

For English and Dutch, you can instead use parse-english and parse-dutch.

You can somewhat use this for Latin-like scripts, such as Cyrillic (“привет”), Georgian (“გამარჯობა”), Armenian (“Բարեւ”), and such.

Install

This package is ESM only. In Node.js (version 16+), install with npm:

npm install parse-latin

In Deno with esm.sh:

import {ParseLatin} from 'https://esm.sh/parse-latin@7'

In browsers with esm.sh:

<script type="module">
  import {ParseLatin} from 'https://esm.sh/parse-latin@7?bundle'
</script>

Use

import {ParseLatin} from 'parse-latin'
import {inspect} from 'unist-util-inspect'

const tree = new ParseLatin().parse('A simple sentence.')

console.log(inspect(tree))

Yields:

RootNode[1] (1:1-1:19, 0-18)
└─0 ParagraphNode[1] (1:1-1:19, 0-18)
    └─0 SentenceNode[6] (1:1-1:19, 0-18)
        ├─0 WordNode[1] (1:1-1:2, 0-1)
        │   └─0 TextNode "A" (1:1-1:2, 0-1)
        ├─1 WhiteSpaceNode " " (1:2-1:3, 1-2)
        ├─2 WordNode[1] (1:3-1:9, 2-8)
        │   └─0 TextNode "simple" (1:3-1:9, 2-8)
        ├─3 WhiteSpaceNode " " (1:9-1:10, 8-9)
        ├─4 WordNode[1] (1:10-1:18, 9-17)
        │   └─0 TextNode "sentence" (1:10-1:18, 9-17)
        └─5 PunctuationNode "." (1:18-1:19, 17-18)

API

This package exports the identifier ParseLatin. There is no default export.

ParseLatin()

Create a new parser.

ParseLatin#parse(value)

Turn natural language into a syntax tree.

Parameters

• value (string, optional) — value to parse

Returns

Tree (RootNode).

Algorithm

👉 Note: The easiest way to see how parse-latin parses, is by using the online parser demo, which shows the syntax tree corresponding to the typed text.

parse-latin splits text into white space, punctuation, symbol, and word tokens:

• “word” is one or more unicode letters or numbers
• “white space” is one or more unicode white space characters
• “punctuation” is one or more unicode punctuation characters
• “symbol” is one or more of anything else

Then, it manipulates and merges those tokens into a syntax tree, adding sentences and paragraphs where needed.

• some punctuation marks are part of the word they occur in, such as non-profit, she’s, G.I., 11:00, N/A, &c, nineteenth- and…
• some periods do not mark a sentence end, such as 1., e.g., id.
• although periods, question marks, and exclamation marks (sometimes) end a sentence, that end might not occur directly after the mark, such as .), ."
• …and many more exceptions

Types

This package is fully typed with TypeScript. It exports no additional types.

Compatibility

Projects maintained by me are compatible with maintained versions of Node.js.

When I cut a new major release, I drop support for unmaintained versions of Node. This means I try to keep the current release line, parse-latin@^7, compatible with Node.js 16.

Security

This package is safe.

Related

• parse-english — English (natural language) parser
• parse-dutch — Dutch (natural language) parser

Contribute

Yes please! See How to Contribute to Open Source.

License

MIT © Titus Wormer