Skip to content

????????????????????? #257

@yeyixiang2007

Description

@yeyixiang2007

What to build

Replace per-request homepage recommendation recomputation with a cheaper path based on precomputation and caching. Anonymous homepage traffic should not force full-material candidate loading and full Python scoring on every request. The recommendation experience can remain useful while using bounded candidate sets and cached outputs.

Acceptance criteria

  • Anonymous homepage recommendations are served from a precomputed or cached result instead of being recomputed on every request.
  • Recommendation freshness and correctness remain acceptable for the homepage experience.
  • The implementation defines cache invalidation or refresh cadence clearly enough for maintainers to operate it.
  • Verification compares the before/after cost of serving repeated anonymous homepage visits.

Blocked by

Metadata

Metadata

Assignees

No one assigned

    Labels

    AFKAway-from-keyboard: 可独立完成enhancementNew feature or request

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions