fix: update QWEN_MODEL_ID default to 1.5B in main.py (missed in previous commit)

Swap default Query-Transform LLM to Qwen2.5-0.5B-Instruct to significantly reduce memory footprint and CPU generation latency on Hugging Face free tier

Optimize PyTorch resource usage: set CPU threads to 1 to prevent health check timeouts and reduce max generation tokens to 45

Optimize backend preloading: preload Qwen & Jina-v5 at startup, cache Qwen in QueryTransformService class attributes, and bypass Ollama connection attempts when not ready

fix: add Cache-Control headers to static files — no-cache for root icons, immutable for hashed assets

fix: serve all root-level static files (icons, favicon.svg, apple-touch-icons)

feat: implement Reference Lookup API and ReferencePopup integration

เตรียมความพร้อมสำหรับการ Deploy บน HF Spaces: เพิ่ม ONNX Reranker, แก้ไข Static Path และปรับปรุง Asset Downloader

chore: stabilization and cleanup audit

feat: ปรับปรุงระบบ Tipitaka AI Reader และ LLM Prompt

feat: health check + Discord webhook alert, non-streaming API endpoint

Expand POPULAR_TERMS: 40 -> 293 via 84000.org dictionary

fix: replace middleware with explicit catch-all route for mobile 404

fix: replace conflicting root route with middleware-based SPA catch-all

Deploy React+FastAPI webapp to HF Space