feat(xhs): add "collect" crawler type to fetch collected/收藏 notes

[dake6767]

What

Adds a new CRAWLER_TYPE="collect" for Xiaohongshu that crawls the
logged-in user's collected (收藏) notes. Previously the xhs crawler only
supported search / detail / creator, so there was no way to export
your own 收藏.

How

• client.py: get_self_user_id() resolves the numeric user_id via
  /api/sns/web/v2/user/me (note: selfinfo only returns red_id, not the
  numeric user_id the collect API needs), plus get_notes_by_collect() /
  get_all_notes_by_collect() paging /api/sns/web/v2/note/collect/page.
• core.py: new dispatch branch + get_collected_notes(), reusing the
  existing note-detail / media / comment pipeline — so ENABLE_GET_COMMENTS,
  CRAWLER_MAX_NOTES_COUNT, and all storage backends work unchanged.
• config: XHS_COLLECT_USER_ID_LIST (empty = own collection; or list
  bare user_ids / full profile URLs with xsec_token for others' public
  collections) and CrawlerTypeEnum.COLLECT in cmd_arg (otherwise the
  config value silently falls back to search).

Usage

# config/base_config.py
PLATFORM = "xhs"
CRAWLER_TYPE = "collect"
# config/xhs_config.py
XHS_COLLECT_USER_ID_LIST = []   # empty = the logged-in user's own collection

python main.py            # or: python main.py --platform xhs --type collect

Output goes to the usual data/xhs/... sink (jsonl/csv/db/...), written as
collect_contents_* / collect_comments_*.

Testing

Verified end-to-end against a live logged-in account: resolves the user_id,
pages the collection, fetches note details + comments, and writes the output
files. Only xhs implements collect; other platforms treat the type as a
no-op via the existing else branch, so nothing else is affected.