feat: adjust alert thresholds to match business logic for zoom recordings

fix: unbound local error for os module in background task

feat: combined audio/video analysis with reduced alert noise

feat: implement asynchronous polling system with background tasks

feat: remove API key authentication temporarily for testing

chore: add Hugging Face Spaces configuration metadata to README

refactor: simplify project by removing heavy PyTorch/MUSIQ dependency to fix HF build

fix: update Dockerfile for HF Spaces (non-root user, DNSMOS downloaded at build time)

feat: initial LectureLens API v1.0 — full audio/video quality analysis service