Fix Douyin pipeline path and WSL guidance

feat(v2.2.0): batch analysis protocol + anti-sampling policy + model compatibility notes

docs: add missing output structure to Chinese README

docs: update model names to Gemini 3.0 / Kimi 2.5, remove GPT-4o

docs: update README to v2.1, draft release notes

feat: v2.1 - align with video-copy-analyzer, cleanup legacy files

Add Douyin video download support

Add references for analysis methodology (bilingual)

Initial commit: Video Expert Analyzer v1.3.0 - AI-powered video analysis with Walter Murch's Six Rules