feat: add Chinese variant conversion via OpenCC

[fzlins]

Summary

Add automatic Chinese variant conversion (Simplified ↔ Traditional) applied after extraction, before saving metadata.json and import.csv.

Changes

• requirements.txt — Add opencc-python-reimplemented>=0.1.7 (pure Python, no C++ build required)
• config.ini — Add chinese_convert option
• tmdb-import/util/chinese_convert.py (new) — Conversion module with lazy-singleton OpenCC converter and punctuation tables
• tmdb-import/extractor.py — Hook conversion into pipeline; extract_from_url now returns metadata
• README.md / docs/README.zh-CN.md — Document new config option

Configuration

# Leave empty to disable. Options: zh-CN, zh-TW, zh-HK
chinese_convert =

How it works

1. After any extractor returns a Metadata object, checks metadata.language
2. If language starts with zh (covers zh-CN, zh-TW, zh-HK, zh-SG, etc.) and chinese_convert is set, runs conversion in-place
3. Both metadata.json and import.csv are saved with the converted text

OpenCC config mapping:

Target	OpenCC config	Description
zh-CN	t2s	Traditional → Simplified
zh-TW	s2twp	Simplified → Taiwan Traditional (with phrase conversion)
zh-HK	s2hk	Simplified → HK Traditional

Punctuation conversion (handled separately since OpenCC doesn't convert punctuation by default):

Target	Conversion
zh-CN	「」『』 → ""''
zh-TW / zh-HK	""'' → 「」『』