Skip to content

Recency-Weighted Ranking — Cải thiện briefing surface scoring #89

@magicpro97

Description

@magicpro97

Mô tả

briefing.py hiện surfaces entries by updated_at order. OpenWolf Hippocampus dùng exponential decay scoring với half-life 30 days.

Học hỏi từ OpenWolf Hippocampus

Recall scoring formula:

score = location(55%) + recency(20%) + valence(15%) + intensity(10%)
recency = exp(-days_since / 30)  // half-life 30 days
location = exact:1.0, glob:0.9, prefix:0.8, parent:0.7, sibling:0.6

Đề xuất

Cải tiến briefing.py surface ranking:

recency_score = math.exp(-days_since / 30)  # half-life 30 days
surface_score = confidence * 0.6 + recency_score * 0.25 + intensity * 0.15

Ưu tiên

Trung

References

Acceptance Criteria

  • Exponential decay scoring hoạt động
  • Configurable half-life (default 30 days)
  • briefing.py dùng new scoring
  • Kết quả relevant hơn so với simple timestamp sort

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions