javanese-analyzer-core

Javanese Krama syntax analyzer with Unggah-Ungguh (politeness level) validation

A TypeScript library for analyzing Javanese language sentences, validating syntax structure (S-P-O-K), and checking politeness levels (Unggah-Ungguh) compliance.

🚀 Installation

npm install javanese-analyzer-core

📖 Quick Start

import { analyze } from "javanese-analyzer-core";

const result = analyze("Kula nedha sekul");

console.log(result);
// {
//   tokens: [
//     { token: "kula", label: "SUBJEK", keterangan: "pronomina diri sendiri" },
//     { token: "nedha", label: "PREDIKAT", keterangan: "verba krama lugu" },
//     { token: "sekul", label: "OBJEK_NOUN", keterangan: "nomina" }
//   ],
//   analisis: {
//     jenis_kalimat: "kalimat tunggal",
//     validitas_sintaksis: "VALID",
//     validitas_unggah_ungguh: "SESUAI"
//   },
//   structure: { /* Parse Tree */ },
//   derivations: [ /* Derivation Steps */ ]
// }

✨ Features

✅ Syntax Analysis (S-P-O-K)

• Tokenization of Javanese Krama/Ngoko words
• Context-Free Grammar (CFG) based parsing
• Support for simple and compound sentences
• Implicit subject detection

✅ Unggah-Ungguh Validation

• Politeness level validation
• Detection of Krama Inggil/Lugu usage errors
• Automatic correction suggestions

✅ Visualizations

• Parse Tree: Hierarchical syntax structure
• Derivation Trace: Step-by-step leftmost derivation
• Token Analysis: Word-by-word breakdown with annotations

✅ Fuzzy Matching

• Automatic typo correction using Levenshtein Distance
• Adaptive threshold based on word length

📚 API Reference

analyze(sentence: string): AnalysisResult

Analyzes a Javanese sentence and returns comprehensive results.

Parameters:

• sentence (string): Javanese sentence to analyze

Returns: AnalysisResult object containing:

• tokens: Array of tokenized words with labels
• analisis: Syntax and Unggah-Ungguh validation results
• structure: Parse tree structure
• derivations: Leftmost derivation steps

Type Definitions

interface AnalysisResult {
  tokens: Token[];
  analisis: {
    jenis_kalimat: string;
    validitas_sintaksis: "VALID" | "TIDAK VALID";
    validitas_unggah_ungguh: "SESUAI" | "TIDAK SESUAI";
    jenis_kesalahan?: string[];
    kalimat_terkoreksi?: string;
    penjelasan?: string;
  };
  structure: ParseNode;
  derivations: string[];
}

interface Token {
  token: string;
  label: string;
  keterangan: string;
}

🌐 Full Application

For a complete web application with interactive UI, check out:

npx create-javanese-analyzer my-app
cd my-app
npm run dev

Or visit the live demo: https://javanese-ai.vercel.app

📦 Dictionary

The package includes a comprehensive dictionary of 1000+ Javanese words with:

• Word classifications (nouns, verbs, pronouns, etc.)
• Politeness levels (Krama Inggil, Krama Lugu, Ngoko)
• Subject classifications (SELF, HONORED, NEUTRAL)

🤝 Contributing

Contributions are welcome! Please visit the main repository for contribution guidelines.

📄 License

MIT © Surya

🔗 Links

• Documentation
• GitHub Repository
• Report Issues
• Live Demo

Made with ❤️ for preserving Javanese language and culture