Repository navigation
lipimala v1.0.0 — Initial Release
lipimala is a cross-language Indic transliteration library with synchronized implementations for Dart, Python, JavaScript/Node.js, and PHP.
This is the first stable release of the canonical multi-runtime project.
Install
| Platform | Command |
|---|---|
| Python | pip install lipimala |
| JavaScript / Node.js | npm install lipimala |
| Dart / Flutter | dart pub add lipimala |
| PHP | composer require jayeshmepani/lipimala |
What's included
- IAST / extended Indic → Devanagari with profile control:
strictIastiso15919CoreextendedIndic
- IAST / extended Indic → Gujarati
- Devanagari ↔ Gujarati direct canonical conversion
- Canonical reverse conversion
- Devanagari → IAST
- Gujarati → IAST
- Plain English transcription
- Hunterian transcription
- Exact source recovery through a checksummed invisible Unicode Tag trailer
- Lossless round-trip envelope with integrity verification
- NFC / NFD and combining-mark preservation
- Vedic mark preservation using Unicode-recommended storage ordering
- Shared cross-runtime verification corpus
Exact source recovery
The exact-recovery mode preserves the original Latin source, including:
- case
- NFC versus NFD representation
- precomposed versus decomposed forms
- alias choice
- combining-mark order
- punctuation
- whitespace
- supplementary Unicode characters
This allows metadata-backed round trips to recover the exact original source key even where visible transliteration mappings are intentionally many-to-one.
Supported scripts / systems
- Devanagari
- Gujarati
- IAST
- ISO 15919 core profile
- Extended Indic
- Hunterian
- Plain English
- Vedic accent marks
Runtime support
- Dart: 3.4+
- Python: 3.12+
- JavaScript / Node.js: Node.js 20+ / ES2023+
- PHP: 8.3+
Verification
The four implementations are maintained against the same transliteration behavior and verification corpus.
Verified areas include:
- Latin / IAST → Devanagari
- Latin / IAST → Gujarati
- Devanagari → canonical IAST
- Gujarati → canonical IAST
- Devanagari ↔ Gujarati
- Plain English / Hunterian transcription
- exact metadata-backed round trips
- NFC / NFD handling
- combining-mark edge cases
- Vedic fixtures
The current shared corpus contains 497 transliteration cases plus 22 Vedic fixtures.
Repository model
This repository is the canonical source repository for all four implementations.
The PHP package is additionally published from a generated PHP-only mirror repository so that composer.json is available at the repository root for Packagist. Development remains centralized in this canonical repository.
Release status
This is the first stable synchronized release of lipimala.