Fix AI Analyst: use Qwen2.5-7B as default HF model

Fix HF 410: always use router.huggingface.co for all models

Fix HF routing: use router for third-party models, not model-specific endpoint

Restore SYSTEM_PROMPT and finetuned model default for AI Analyst

Add dynamic LLM provider/model selection to AI Analyst

Fix HF Inference API: correct router URL + Qwen2.5 model

Fix Generate Now: HF 503 retry + 120s countdown timer in UI

AI Analyst: add Generate Now button for on-demand insights

Add AI Analyst service: Ollama (local) + HuggingFace (HF Spaces)