@syllst/glost

Remark plugin that enriches Syllst nodes with GLOST word-level annotations.

Installation

pnpm add @syllst/glost

Peer Dependencies

For full GLOST integration, install the glost packages:

pnpm add glost-core glost-processor

Usage

Basic Usage

import { createMDXParser } from '@syllst/processor';
import { remarkSyllabusDirectives } from '@syllst/processor/plugins';
import { remarkSyllstGlost } from '@syllst/glost';

// Create a transcription provider (example)
const thaiProvider = {
  getTranscription: (word: string, scheme: string) => {
    // Return transcription for the word
    return transcriptions[word]?.[scheme];
  },
  getDefaultScheme: () => 'paiboon+',
};

// Configure the processor
const processor = createMDXParser()
  .use(remarkSyllabusDirectives)
  .use(remarkSyllstGlost, {
    languages: [
      {
        lang: 'th',
        transcriptionProvider: thaiProvider,
        defaultScheme: 'paiboon+',
      },
    ],
    defaultLang: 'th',
    targets: ['vocabularyItem', 'dialogueTurn'],
  });

// Process MDX content
const result = await processor.process(mdxContent);

With @laeng Language Packages

import { remarkSyllstGlost } from '@syllst/glost';
import { createThaiTranscriptionProvider } from '@laeng/th';

const processor = createMDXParser()
  .use(remarkSyllabusDirectives)
  .use(remarkSyllstGlost, {
    languages: [
      {
        lang: 'th',
        transcriptionProvider: createThaiTranscriptionProvider(),
        defaultScheme: 'paiboon+',
      },
    ],
    defaultLang: 'th',
  });

API

remarkSyllstGlost(options?)

Remark plugin that enriches syllst nodes with transcription data.

Options

| Option | Type | Default | Description | |--------|------|---------|-------------| | languages | LanguageProviderConfig[] | [] | Language provider configurations | | defaultLang | string | 'en' | Default language for content without explicit lang | | targets | EnrichTarget[] | ['all'] | Node types to enrich: 'vocabularyItem', 'dialogueTurn', 'example', 'all' | | skipExisting | boolean | true | Skip nodes that already have transcription | | schemes | string[] | Provider default | Transcription schemes to generate | | wordLevelDialogue | boolean | false | Enable word-level GLOST annotations for dialogue | | sentenceLevelDialogue | boolean | false | Enable sentence-level GLOST annotations for dialogue |

LanguageProviderConfig

interface LanguageProviderConfig {
  lang: string;                              // Language code (e.g., 'th', 'ja')
  transcriptionProvider?: TranscriptionProvider;
  defaultScheme?: string;                    // Default transcription scheme
}

TranscriptionProvider

interface TranscriptionProvider {
  getTranscription(word: string, scheme: string): string | undefined;
  getTranscriptions?(word: string): Record<string, string> | undefined;
  getDefaultScheme?(): string;
}

createProviderRegistry(configs)

Creates a language provider registry from an array of configurations.

import { createProviderRegistry } from '@syllst/glost/providers';

const registry = createProviderRegistry([
  { lang: 'th', transcriptionProvider: thaiProvider },
  { lang: 'ja', transcriptionProvider: japaneseProvider },
]);

registry.get('th');      // Returns Thai provider config
registry.has('ja');      // true
registry.getLanguages(); // ['th', 'ja']

LanguageProviderRegistry

Class for managing language providers.

import { LanguageProviderRegistry } from '@syllst/glost/providers';

const registry = new LanguageProviderRegistry();
registry.register({ lang: 'th', transcriptionProvider });
registry.get('th-TH'); // Falls back to 'th' if 'th-TH' not found

Enrichers

For advanced usage, individual enrichers are available:

import {
  enrichVocabularyItem,
  enrichDialogueTurn,
  enrichExample,
} from '@syllst/glost/enrichers';

How It Works

1. The plugin runs after remarkSyllabusDirectives in the unified pipeline
2. It visits all nodes matching the configured targets
3. For each node, it:
   • Detects the language (from node, parent, or default)
   • Looks up the appropriate language provider
   • Calls the transcription provider to get transcription data
   • Enriches the node's transcription field

Transcription Output

Single scheme:

node.transcription = "sawatdee";

Multiple schemes:

node.transcription = {
  primary: "sawatdee",
  paiboon: "sà-wàt-dee",
  ipa: "/sà.wàt.diː/",
  rtgs: "sawatdi"
};

Metadata

After processing, metadata is attached to the vfile:

const result = await processor.process(content);
const meta = result.data.syllstGlost;

console.log(meta.nodesEnriched);   // Number of nodes enriched
console.log(meta.processingTime);  // Time in ms
console.log(meta.results);         // Detailed per-node results

License

MIT