Spaces:
Running
Running
| # ── Core framework ──────────────────────────────────────────── | |
| fastapi | |
| uvicorn[standard] | |
| python-multipart # required for Form / UploadFile in FastAPI | |
| # ── AI / ML services ────────────────────────────────────────── | |
| groq # Groq Whisper transcription + LLM story generation | |
| google-genai # Google Gemini / Gemma translation | |
| openai # Required for Hugging Face Router (Qwen text generation V14.0) | |
| # ── Audio processing ────────────────────────────────────────── | |
| pydub # audio chunking + format conversion | |
| # Note: pydub calls ffmpeg/ffprobe — installed as system packages in Dockerfile | |
| # ── [IMMERSION] Text-to-speech, dataset upload, scheduler ───── | |
| edge-tts # Microsoft Edge neural TTS (free, no API key required) | |
| huggingface_hub # Upload story assets to a HF Dataset repo | |
| Pillow>=10.0.0 # Convert FLUX.1-schnell PNG output → JPEG before CDN upload | |
| apscheduler>=3.10.0 # Auto-publish Immersion story every 6 hours | |
| # ── [FIX-ARABIC] RTL text shaping for FFmpeg subtitle burn ──── | |
| arabic-reshaper # connects Arabic letters into proper ligature forms | |
| python-bidi # handles right-to-left visual ordering for LTR renderers | |
| # ── HTTP client ─────────────────────────────────────────────── | |
| httpx>=0.27.0 | |
| # ── [TITAN MENTOR] Firebase Admin SDK ───────────────────────── | |
| firebase-admin # Firestore integration for lesson status tracking | |
| # (immersion.py imports it with try/except — server | |
| # starts normally even if cred env var is missing) |