lipimala 1.2.0

Premium exact round-trip Indic transliteration library for Devanagari, Gujarati, plain English, and canonical reverse conversion.

Changelog

1.2.0

• Added opt-in Brahmic naturalization for the currently supported Devanagari and Gujarati scripts while keeping all 1.1 defaults unchanged.
• Added a three-way Brahmic vowel-length policy: COLLAPSE, DOUBLE, and contextual SMART (Dart: collapse, doubled, smart).
• SMART is syllable spelling for modern Gujarati and Hindi. Closed one-syllable words keep length (aaj, chaar, naam, hoon, swaroop). Open particles and heavy clusters shorten (ka, marg, dharan, karine). An open ending keeps stem length (swaroopama, paasethi, shwetdweepthi). A nasal coda keeps a peripheral vowel (hoonga). Sanskrit SMART preserves phonemic length (shvetadveepa, svaroopa).
• Natural schwa between p and unaspirated t stays, so Hindi राष्ट्रपति is rashtrapati. Aspirated થ / थ still contracts (શાપથી → shaapthi). Legacy गांधीनगर stays gaandheengar; natural Hindi is gandhinagar.
• Plain-English defaults to doubled. Hunterian defaults to collapse when vowel length is omitted. With natural + smart on both, Gujarati Plain and Hunterian match. Gujarati, and modern natural Hindi, do not apply blanket consonant-plus-v → w.
• Named Hunterian as simplified ASCII Hunterian. A long vowel is one letter (jan) and consonant dots are removed. The 1954 macron form (jān) is not emitted.
• Generalized natural schwa/compound handling across the Devanagari/Hindi and Gujarati paths, including श्वेतद्वीप/શ્વેતદ્વીપ → shwetdweep, स्वरूप/સ્વરૂપ → swaroop, and सतयुग/સતયુગ → satyug.
• Added Gujarati ending patterns (જેટલાં → jetla, બ્રહ્માંડમાં → brahmandma, અક્ષરધામમાં → akshardhamma) and independent ઋ → ru in Gujarati-readable mode.
• Added controlled modern sv/śv/ṣv/dv → sw/shw/.../dw readable clusters without blanket v → w leakage into words such as manvantar, vasant, or ravivar.
• Added opt-in sentence-start capitalization and exact readable-word overrides (for example shwetdweep → Shwetdweep) without semantic proper-name guessing.
• Canonical Brahmic → IAST, exact-source recovery, legacy Plain-English/Hunterian defaults, and 1.1 constructor behavior remain unchanged.
• Added natural-mode scalar/list regression coverage for Gujarati, Devanagari/Hindi, Sanskrit-conservative SMART behavior, and the full Gujarati prose vector.

1.1.0

• Added direct Devanagari/Gujarati → Plain-English and Hunterian APIs, list helpers, and exact-source result-envelope variants.
• Added source-aware inherent-schwa tracking with Gujarati/Hindi/Sanskrit policies.
• Added direct-transcription controls for vowel length and terminal anusvara.
• Kept legacy IAST → Plain-English/Hunterian behavior backward-compatible.
• Added Brahmic naturalization regression tests.
• Hardened Unicode 18 Devanagari/Gujarati reverse coverage for historic prishthamatra sequences, short-AI AY, Devanagari Extended/Extended-A, Vedic segmental/nasal signs, and Common Indic number forms.
• Added semantic Gujarati Arabic-transliteration handling for sukun, shadda, maddah, special nuktas, and configurable ʿayn rendering.
• Corrected Sindhi implosive identities (ɠ, ʄ, ɗ, ɓ) and added canonical forward aliases where applicable.
• Plain-English/Hunterian now renders semantic OM as om and no longer leaks specialized Vedic/Devanagari nasal marks by default.
• Fixed historic long-E/OOE readable rendering (ॕ → ee, ऻ/ॴ → ooe) consistently across runtimes.
• Preserved pre-existing Latin/IAST text, Unicode diacritics, and smart apostrophes/quotes exactly in direct Brahmic → Plain-English/Hunterian mixed-script input.
• Added Konkani च़ → ċ → tsa, audible candrabindu handling, dangling-modifier cleanup, and Gujarati Hunterian visarga handling without rewriting original Latin w.

1.0.2

• Added bulk list transliteration extensions and top-level functions (toDevanagariFromIastList, toGujaratiFromIastList, toPlainEnglishFromIastList, toCanonicalGujaratiFromDevanagariList, toCanonicalDevanagariFromGujaratiList, toExactDevanagariFromGujaratiList, toExactGujaratiFromDevanagariList, toDevanagariList, toGujaratiList, toPlainEnglishList).
• Added primary package example example/example.dart and complete dartdoc comments to achieve 160/160 pub points score on pub.dev.
• Cleaned up obsolete Flutter references in package documentation and badges.

1.0.1

• Re-exported all public converters, models, options, and utilities from primary barrel file package:lipimala/lipimala.dart.
• Added comprehensive HTML API documentation and usage examples.

1.0.0

• Added canonical Devanagari/Gujarati script-to-script conversion options and policies.
• Preserved Vedic mark reattachment on canonical script transliterations.
• Standardized package manifests, author details, and release actions across all runtimes.
• Added the complete Node.js 20+/ES2023 port.
• Preserved extendedIndic as the default profile.
• Ported Devanagari and Gujarati forward conversion mappings and policies.
• Ported plain-English and Hunterian transcription policies.
• Ported canonical Devanagari/Gujarati reverse conversion.
• Ported Vedic accent handling and accent reattachment.
• Ported Unicode normalization controls.
• Ported the exact round-trip result envelope and JSON schema.
• Ported the invisible exact-source Unicode Tag trailer byte-for-byte.
• Added all 497 index-aligned corpus cases in five conversion directions.
• Added all 22 Vedic fixtures.
• Added five CLI-compatible output generators.
• Added Node built-in test-runner coverage with no runtime dependencies.