"""Retrieval quality tracking tools: mark_memory_used, retrieval_quality_stats."""
from memory_mcp.responses import success_response
from memory_mcp.server.app import mcp, storage
@mcp.tool
def mark_memory_used(
memory_id: int,
feedback: str | None = None,
) -> dict:
"""Mark a memory as actually used/helpful after recall.
Call this when a recalled memory was useful in your response.
Helps improve ranking by tracking which memories are valuable.
Args:
memory_id: ID of the memory that was useful
feedback: Optional feedback (e.g., "helpful", "partially_helpful")
Returns:
Success response with update count
"""
updated = storage.mark_retrieval_used(memory_id, feedback=feedback)
if updated > 0:
return success_response(
f"Marked memory {memory_id} as used",
updated_count=updated,
)
return success_response(
"No retrieval event found to update (tracking may be disabled)",
updated_count=0,
)
@mcp.tool
def retrieval_quality_stats(
memory_id: int | None = None,
days: int = 30,
) -> dict:
"""Get retrieval quality statistics.
Shows which memories are frequently retrieved and actually used.
Helps identify high-value and low-utility memories.
Args:
memory_id: Get stats for specific memory (None for global)
days: How many days back to analyze (default 30)
Returns:
Statistics on retrieval and usage patterns
"""
stats = storage.get_retrieval_stats(memory_id=memory_id, days=days)
return success_response("Retrieval quality stats", **stats)
