feat(memory): Day M.1 complete - importance scoring system

- Add FrequencyTracker: increment(), get_frequency_score(), get_recency_score(), get_time_decay()
- Add EmotionAnalyzer: EMOTION_KEYWORDS dict, extract(), calculate_score(), get_emotion_profile()
- Add ImpactEvaluator: evaluate(), get_topic_overlap(), rank_by_impact()
- Add ImportanceScorer: composite scoring (freq 35% + recency 20% + emotion 25% + impact 20%)
- Update UserMemory model: frequency_count, emotion_tags, importance_score, importance_level, associated_topics
- Integrate ImportanceScorer into memory_service.py (recall + importance update)
- Add 37 tests for all memory scoring components
- Fix urgency patterns: remove overly broad '今天' that matched neutral text
- Update memory-update checklist: mark all M.1 tasks complete

backend/app/services/memory/impact_evaluator.py

@@ -0,0 +1,52 @@
+"""
+ImpactEvaluator
+
+Evaluates the breadth of impact a memory has based on associated topics.
+"""
+
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from app.models.memory import UserMemory
+
+
+class ImpactEvaluator:
+    """Evaluate the impact breadth of a memory"""
+
+    # Threshold for maximum impact score
+    IMPACT_THRESHOLD = 5  # Number of associated topics for max impact
+
+    def evaluate(self, memory: "UserMemory") -> float:
+        """Calculate impact score (0.0 - 1.0)
+
+        The more associated topics a memory has, the higher its impact.
+        Topics represent "what this memory is about" — if it touches
+        many aspects of the user's life, it has high impact.
+        """
+        associated_topics = memory.associated_topics or []
+        if not associated_topics:
+            return 0.0
+
+        # Normalize: IMPACT_THRESHOLD topics = full impact (1.0)
+        raw_score = len(associated_topics) / self.IMPACT_THRESHOLD
+        return min(1.0, raw_score)
+
+    def get_topic_overlap(self, memory_a: "UserMemory", memory_b: "UserMemory") -> float:
+        """Calculate topic overlap between two memories (0.0 - 1.0)
+
+        Used for finding related memories.
+        """
+        topics_a = set(memory_a.associated_topics or [])
+        topics_b = set(memory_b.associated_topics or [])
+
+        if not topics_a or not topics_b:
+            return 0.0
+
+        intersection = topics_a & topics_b
+        union = topics_a | topics_b
+
+        return len(intersection) / len(union) if union else 0.0
+
+    def rank_by_impact(self, memories: list["UserMemory"]) -> list["UserMemory"]:
+        """Rank memories by impact score (descending)"""
+        return sorted(memories, key=lambda m: self.evaluate(m), reverse=True)