GamerRegalia/tajweed-analyzer

Fix qalqalah kubra/sughra detection to respect per-verse waqf boundaries

Enforce line-by-line 1-4 Malay feedback format

Format AI feedback in Malay and concise 4-point style

Fix GPT-5.4-mini max completion parameter

improve: robust Arabic word accuracy alignment

chore: switch OpenAI feedback model to gpt-5.4-mini

fix: transcription and tajweed detection bugs

fix: switch to sdk:docker to eliminate double-server conflict

fix: remove duplicate uvicorn server start in HF Spaces

fix: upgrade transformers and accelerate for huggingface-hub compatibility

fix: Qalqalah Sughra/Kubra classification by waqf context

Fix: Resolve uvicorn dependency conflict

Fix: Resolve python-multipart dependency conflict

Fix: Resolve dependency conflict between gradio and fastapi

Migrate from Gemini to OpenAI API

Update tajweed_checker and other files

Update tajweed_checker and other files

Update tajweed_checker and other files

Update accuracy calculator

Update accuracy calculator

Update accuracy calculator

feat: enforce diacritics check for expected_text in both Gradio and API endpoints for robust Tajweed analysis

Fix Whisper: use forced_decoder_ids for Arabic transcription

Switch to Tarteel Whisper model for better Quran transcription accuracy

Fix word accuracy: use fuzzy matching instead of substring method

Update accuracy calculator

Fix word accuracy: fuzzy matching with similarity threshold

Apply user edits

User requested push: sync local changes

User requested push: sync changes

User requested push (repeat)

Sync: user requested push

Apply user edits

Apply user edits

Apply user edits

Apply user edits

Apply user edits

Apply user edits

Fix word accuracy calculation - include in results and add autojunk=False

Fix accuracy calculation: add autojunk=False and swap parameter order

Fix audio preprocessing to return path only

Remove BOM character

Fix Gemini model and accuracy field mapping

Fix Gemini model name to gemini-1.5-flash

Fix audio format for transcription

Add generate_feedback wrapper method

Add check_rule method

Replace Flask with FastAPI for Gradio compatibility

Add accelerate library

Fix Gemini import error