Fix: TerminalPunctuationGuard for short phrases and comprehensive Nasb/Jazm grammar rules fixes

Refactor: Remove obsolete IVtoOOV filter to prevent False Negatives

Fix: Add missing dual and plural IVtoOOV bypass rules in app.py

Fix: Reorder Jaccard filter in app.py to run after evaluating grammar bypass rules

Fix: Strip trailing punctuation in grammar bypass rule to handle words at the end of sentences

Fix: Allow gender prefix swap (ي/ت) during feminine plural present tense bypass

Fix: Add bypass rule for present tense feminine plural to prevent IVtoOOV rejection

Phase 11.5: Adversarial diagnostic + 4 critical pipeline fixes

Phase 11: Hierarchical StageLocker — grammar overrides spelling locks

Phase 13: Fix punctuation terminal injection filter (context-aware Rule 0)

Phase 13: Fix spelling filter blind spots + eager model loading

Phase 12: Spelling pipeline integration + Benchmark integrity fixes

Phase 11: Inline telemetry in API response

Phase 11: Telemetry instrumentation + OffsetMapper tests

FIX-27: Grammar structured data + hallucination protection

FIX-25: Block grammar punct spacing (حالك؟→حالك ؟)

FIX-24: Block grammar model punct stripping (البلاد.→البلاد)

FIX-23: Block grammar model tanween removal + HF log fetcher

CRITICAL FIX: _DIRECTIONAL_BLOCKS was local scope → NameError in grammar diff filter

Hotfix: Revert pronoun-h guard to ته-only (was breaking ta_marbuta), add directional blocks for عمله/لسانه/بيته/كتابه

Phase 13: Grammar SV/gender whitelist, preposition/nasb regex, pronoun-h guard

Phase 12: Fix 6 batches — grammar pipeline bypass, religious/structured protection, punct rearrangement, SV/gender agreement

Phase 11: Master Bug Fix — 23 fixes across 8 files

Phase 8: Fix all critical bugs from adversarial validation

feat: dialect-to-MSA conversion — backend API + frontend + Docker model cache

fix: better quran import error logging to diagnose HF failure

feat: Quran verification & translation endpoint + modal UI (13 languages)

revert: restore backend to 1a76471, keep all UI/UX improvements

revert: restore backend/model files to cf07939 state, keep UI/UX changes only

fix: accept ص↔س/ض↔ظ/ذ↔ز spelling corrections (المدرصة→المدرسة), fix punctuation dup regex

fix: grammar/quran icons different, tooltip no strikethrough, punctuation no duplicate dots

UI: nav logo, bayyinah hover, quran icon, hero autocomplete shadow, clean tooltip, fix punctuation dup + hamza كان

fix: block grammar كان→كأن + punctuation mid-word split (الدفت.ر)

feat: P1-P3 all 8 items - ta marbuta fix, UI feedback, caching, rate limiting, unit tests

fix: block grammar model from removing/normalizing punctuation marks

feat: P1-P3 improvements + auto re-analyze after apply

fix: Pipeline hardening - 37 bugs + 10 edge cases (Round 1 + Round 2)

fix(spelling): strict ortho-only filter prevents word corruption

debug(punctuation): enhanced lock logging to show which owner blocks punctuation

fix(punctuation): allow pure punctuation changes through stage lock

feat(autocomplete): threshold=0.05 + max 3 suggestions

fix(autocomplete): true context awareness - GPT-2 now adds own candidates + cache key uses 5 words

feat(NLP-4): AutoComplete integration — hybrid bigram + GPT-2

fix(critical): stop spelling from corrupting correct words + fix pronoun agreement + reject hallucinations

fix: re-label grammar spelling fixes + handle prefix in hamza whitelist

fix: all model bugs — S1 S2 S3 G1 P1 (6 fixes across 4 files)

fix: QA sweep — 6 issues resolved (sidebar chips, corrected field, input sanitization, punc cap, docs, tests)

feat: Pipeline Hardening v3.3 — stage ownership, punctuation safety, UUID migration

feat: NLP-3.5 Hardening Sprint — Performance + Overlap Resolver TASK 2: Per-stage timing instrumentation (spelling_ms, grammar_ms, punctuation_ms, total_ms) TASK 3: Smart text processing — skip AraSpell for texts > 300 chars TASK 4+5: Global overlap resolver with priority system grammar(3) > punctuation(2) > spelling(1) > autocomplete(0) Full span collision detection (exact + partial overlaps) TASK 6: AutoComplete priority hook prepared (priority=0)

feat: NLP-3 PuncAra-v1 Integration — Local punctuation model - Created src/nlp/punctuation/ package (rules + service) - Extracted from PuncAra.py: preprocessing, postprocessing, chunking logic - PunctuationChecker: lazy-loaded EncoderDecoderModel from bayan10/PuncAra-v1 - Wired into /api/analyze as Step 3 (Spelling -> Grammar -> Punctuation) - Updated /api/punctuation standalone endpoint - Updated /api/health to report punctuation status - Dockerfile pre-downloads PuncAra-v1 weights during build - Increased gunicorn timeout to 300s for full pipeline