Describe your use case

Users need flexible ways to sort search results based on different criteria. Currently, PowerMem returns results with relevance scores, but users might want to:

• Sort by recency (newest first)
• Sort by importance score
• Sort by access frequency
• Sort by Ebbinghaus retention score
• Combine multiple sorting criteria (e.g., relevance first, then by date)

Describe the solution you'd like

Extend the search method to support multiple sorting options:

1. Sorting modes:

   • relevance (default, current behavior)
   • date_asc / date_desc (by creation/update time)
   • importance_asc / importance_desc (by importance score)
   • access_count_desc (by number of times accessed)
   • retention_desc (by Ebbinghaus retention score)
   • custom (user-defined sorting function)

2. Multi-criteria sorting:

   • Primary and secondary sort keys
   • Example: Sort by relevance, then by date for ties

3. API enhancement:

   results = memory.search(
       "query",
       user_id="user123",
       sort_by="date_desc",  # or ["relevance", "date_desc"]
       limit=20
   )

The solution should:

• Maintain backward compatibility (default to relevance sorting)
• Be efficient (use database-level sorting when possible)
• Support all existing search features (filters, scopes, etc.)
• Document sorting behavior clearly

Describe alternatives you've considered

No response

Additional context

No response