inditrans 0.17.2

A functional transliterator library for the most common Indic/Brahmic scripts.

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

0.17.2

• Add DevanagariTraditional output option to map Dravidian short e/o and consonants to standard Devanagari forms.

0.17.1

• Use readable Roman vowel combinations in Indlish, fixing misplaced breve marks in Urdu output.

0.17.0

• Add generic pronunciation lexicons with bundled Urdu wordlist support, reusable lexicon handles, build tooling and cross-runtime APIs.

0.16.5

• Accept common IPA vowel and consonant variants, breathy aspirates, affricates, diphthongs and syllabic liquids.
• Handle IPA syllable and stress marks, consonant gemination and nasal vowels, including precomposed forms.

0.16.4

• Expand approximate script coverage, including Burmese, Thai, Khmer, Gurmukhi, Bengali, Assamese, Telugu and Kannada, and document remaining unmapped characters.
• Render Indlish velar and palatal nasals as ng and ny outside consonant clusters.

0.16.3

• Support native punctuation, Tibetan syllable separators and abbreviation signs across scripts.
• Use supported demo samples, verify sample and keyboard coverage against WASM, and document remaining symbol limitations.
• Preserve separate edits for each demo input script and restore its sample independently.

0.16.2

• Improve the demo keyboard with curated script keys and preserve existing ASCII shortcuts.
• Support Burmese medial signs and fix Tamil superscript Vedic accent attachment.
• Improve Indic Vedic accent font fallback in the demo.

0.16.1

• Bundle native engine sources in the Flutter package so Android and desktop/Apple plugin builds work outside the monorepo (fixes #64).

0.16.0

• Use lightweight same-script copying with configurable marker, Tamil annotation, accent, XML tag, and protected-span handling.
• Disable Vedic accent output and demo accent keys for WX and SLP1.
• Disable Vedic accent output and demo accent keys for Grantha.
• Group scripts as Indic, Roman, and Perso-Arabic, with explicit Vedic accent output support; keep Tamil in Indic with its specialized transliteration handlers.
• Use Roman terminology consistently across engine code, generated data, tooling, demo, and documentation.
• Replace the Malayalam-based tamilExtended profile with Unicode Grantha (grantha) across the engine, Dart/TypeScript/JavaScript APIs, and demo keyboard. The old script ID is removed.

0.15.0

• Specialize common readers once per call to omit Thai and Urdu checks from Vedic, Tamil and Roman conversion loops.
• Focus native performance comparisons on Vedic scripts, Tamil and ISO; promote this engine to the main performance baseline.
• Keep MinGW support runtimes static when rebuilding the Windows example DLL.
• Add Urdu/Perso-Arabic transliteration (urdu, alias nastaliq), with explicit vowels, aspirates, shadda and limited hamza handling; configure Nastaliq fonts and RTL panes in the demo.
• Add missing ISO-profile notation inputs and native affricate, alveolar-stop, nasal and visarga-allophone identities; keep Sinhala ae distinct from candra e.
• Replace synthetic vowel/separator and consonant-marker approximations with native forms; make native reader choices explicit and reject unexplained generator collisions.
• Preserve Tamil written word-final stops with pulli instead of deleting them.
• Distinguish Odia ଵ (v) and ୱ (w), and emit Gurmukhi adhak for simple geminates after vowels.
• Correct Thai vowel order and Tibetan/Burmese/Khmer stacking; preserve additional real accents and diacritics rather than overwriting them.
• Preserve literal/protected/XML spans without applying writer-marker filtering.
• Display, copy and save raw engine output, deploy matching demo engine assets, and run actual Flutter tests from make testall.

0.14.1

• Indlish Romanization:
  • Rename readableLatin to indlish across native C++ engine, Flutter/Dart (Script.indlish), and TypeScript/JavaScript packages (Script.indlish).
  • Update vowel and diacritic notation for intuitive English readability.
  • Implement contextual anuswara assimilation before varga consonants: velar, palatal, and dental stops assimilate to n, retroflex stops assimilate to N, and labial stops assimilate to m.
  • Fallback to explicit ṁ when anuswara occurs before non-varga consonants (sibilants, semivowels, aspirates), vowels, symbols, raw-string skip boundaries, or at word endings.
  • Honor DisableInference option to bypass contextual assimilation and consistently output explicit ṁ.
  • Preserve chandrabindu as m̐ without nasal assimilation.

0.14.0

• Add DisableInference across native, Flutter, Node.js, and standalone JavaScript APIs to disable anuswara assimilation and contextual Tamil pronunciation and letter substitutions.
• Automatically disable inference when reading Tamil with TamilSuperscripted, including case-insensitive source names; Tamil output formatting alone continues to infer pronunciation.
• Rebuild bundled WASM engines and cover inference options in native, shared fixture, and standalone WASM tests.

0.13.1

• Transliteration features:
  • Adopt uniform phonemic mapping for ITRANS vowels adhering to the ITRANS specification: e/o represent short vowels (Kannada ಎ/ಒ, Devanagari ऎ/ऒ), E/O represent long vowels (ಏ/ಓ, ए/ओ), with ^e/^o supported as aliases (fixes #37).

0.13.0

• Transliteration features:
  • Add skipStart and skipEnd delimiter support to allow passthrough of protected regions (defaults to ##...##).
  • Add Harvard-Kyoto (Script.hk) and additional notation support across runtimes.
  • Handle empty, single-character, and oversized protected-span delimiters safely.
• Engine architecture & performance:
  • Upgrade engine to C++23 standard across native, Flutter, Apple, Android, Windows, and WASM builds.
  • Optimize memory usage: streaming chunk readers with reusable token chunks, non-owning character maps, stack-allocated readers and writers, and allocation-free sinks.
  • Replace dynamic runtime map allocations with compact, static lookup tables and specialized script type conversions.
  • Match transliteration buffer allocation and deallocation across boundaries (releaseBuffer).
• Toolchains & platforms:
  • Reproducible Linux-to-Windows x86-64 DLL cross-compilation with MinGW-w64 (make dll) exporting strict C ABI symbols (transliterate, isScriptSupported, releaseBuffer).
  • Statically link MinGW C++ runtime libraries to eliminate external DLL runtime dependencies.
  • Upgrade and pin Emscripten to 6.0.10 for standalone Flutter WASM and Node.js WASM builds with deterministic parity checks.
  • Upgrade Node.js package to TypeScript 6.x and Node 24 support.
  • Raise minimum Flutter SDK constraint to 3.38.10 / Dart 3.10.8 across declared platforms.
• Release engineering & repository reorganization:
  • Reorganize repository into canonical root structure: move shared C++ engine sources, CLI, and test suite from flutter/native/ to root native/.
  • Establish automated, tag-driven release workflow (.github/workflows/release.yml) for vX.Y.Z Git tags.
  • Enable GitHub Actions OIDC trusted publishing for both pub.dev and npm (@vm75/inditrans).
  • Move repository code-generation and maintenance tooling to Python.
  • Add comprehensive release regression test suite (scripts/tests/test_release_engineering.py) and validator alignment.
  • Expand project documentation: consolidated root CHANGELOG.md, ARCHITECTURE.md, AGENTS.md, and distribution guides.

0.12.1

• Fixed tamil supersctipt encoding

0.12.0

• Fix toScript

0.11.1

• Update dart Option

0.11.0

• Fixed error conditions during transliterate. Fixed android build.

0.10.1

• Update topics

0.10.0

• Updated universal_ffi to version 1.1.0

0.9.0

• flutter: use universal_ffi

0.8.0

• upgraded wasm_ffi for flutter and added vanilla dart example

0.7.1

• Fixed Windows compile issues

0.7.0

• Fixing many issues in flutter. Adding indic to scripts.

0.6.0

• Refactoring and constants generation

0.5.0

• Added support for Gurmukhi adhak

0.4.6

• Fixed Zero Width Joiner and Non-Joiner

0.4.5

• Support compound alternates and Malayalam Chillus

0.4.4

• Fixed ITRANS and unit tests

0.4.3

• added support for xml tag skip

0.4.2

• Added case insensitive support for iast, iso and ipa. Added support for extended roman chars with composition

0.4.1

• Split symbols, Zero Width chars and Vedic symbols

0.4.0

• Lot of refactoring.

0.3.6

• Refactored nodejs interface. Updated readme.

0.3.5

• Fixed wasm, tests and nodejs support

0.3.4

• Fixed some corner case

0.3.3

• Added support for Tamil prefix

0.3.2

• Renamed indicroman to readablelatin

0.3.1

• Improved Tamil transliteration based on rules

0.2.0

• Fixed Sinhara symbols

0.1.9

• modified easyroman and renamed to readablelatin

0.1.8

• Made InferAnuswara automatic. Fixed Diacritic for Vowels.

0.1.7

• fixed bug for indic

0.1.6

• exported indic, added 2 scripts, fixed roman

0.1.5

• Added fromString to Script enum

0.1.4

• re-publish from correct branch.

0.1.3

• Fixed build error

0.1.2

• Added option to use ASCII numerals

0.1.1

• fixed inferAnuswara for consonants after ma

0.1.0

• Removed unnecessary Inditrans class. Added helpful scripts.

0.0.6

• fixed example.

0.0.5

• fixed android build issue.

0.0.4

• removed unnecessary code and files.

0.0.3

• Added comments and fixed formatting and example.

0.0.2

• Fixed ffi_lib for non-web

0.0.1

• Initial release