@spec composite_score(float(), Nous.Memory.Entry.t(), keyword()) :: float()

Compute composite score combining relevance, importance, and recency.

Default weights: relevance: 0.5, importance: 0.3, recency: 0.2

@spec rrf_merge(
  [{Nous.Memory.Entry.t(), number()}],
  [{Nous.Memory.Entry.t(), number()}],
  keyword()
) ::
  [{Nous.Memory.Entry.t(), float()}]

Reciprocal Rank Fusion merge of two ranked result lists.

RRF formula: score(d) = sum(1 / (k + rank(d))) across all lists where d appears.

@spec temporal_decay(float(), Nous.Memory.Entry.t(), keyword()) :: float()

Apply temporal decay to a relevance score.

decay = exp(-lambda * hours_since_access) Returns original score if entry is evergreen.