feat(context): implement context scoring and ranking (Phase 3)

Add comprehensive scoring system with three strategies:
- RelevanceScorer: Semantic similarity with keyword fallback
- RecencyScorer: Exponential decay with type-specific half-lives
- PriorityScorer: Priority-based scoring with type bonuses

Implement CompositeScorer combining all strategies with configurable
weights (default: 50% relevance, 30% recency, 20% priority).

Add ContextRanker for budget-aware context selection with:
- Greedy selection algorithm respecting token budgets
- CRITICAL priority contexts always included
- Diversity reranking to prevent source dominance
- Comprehensive selection statistics

68 tests covering all scoring and ranking functionality.

Part of #61 - Context Management Engine

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

backend/app/services/context/scoring/recency.py

@@ -0,0 +1,141 @@
+"""
+Recency Scorer for Context Management.
+
+Scores context based on how recent it is.
+More recent content gets higher scores.
+"""
+
+import math
+from datetime import UTC, datetime, timedelta
+from typing import Any
+
+from .base import BaseScorer
+from ..types import BaseContext, ContextType
+
+
+class RecencyScorer(BaseScorer):
+    """
+    Scores context based on recency.
+
+    Uses exponential decay to score content based on age.
+    More recent content scores higher.
+    """
+
+    def __init__(
+        self,
+        weight: float = 1.0,
+        half_life_hours: float = 24.0,
+        type_half_lives: dict[ContextType, float] | None = None,
+    ) -> None:
+        """
+        Initialize recency scorer.
+
+        Args:
+            weight: Scorer weight for composite scoring
+            half_life_hours: Default hours until score decays to 0.5
+            type_half_lives: Optional context-type-specific half lives
+        """
+        super().__init__(weight)
+        self._half_life_hours = half_life_hours
+        self._type_half_lives = type_half_lives or {}
+
+        # Set sensible defaults for context types
+        if ContextType.CONVERSATION not in self._type_half_lives:
+            self._type_half_lives[ContextType.CONVERSATION] = 1.0  # 1 hour
+        if ContextType.TOOL not in self._type_half_lives:
+            self._type_half_lives[ContextType.TOOL] = 0.5  # 30 minutes
+        if ContextType.KNOWLEDGE not in self._type_half_lives:
+            self._type_half_lives[ContextType.KNOWLEDGE] = 168.0  # 1 week
+        if ContextType.SYSTEM not in self._type_half_lives:
+            self._type_half_lives[ContextType.SYSTEM] = 720.0  # 30 days
+        if ContextType.TASK not in self._type_half_lives:
+            self._type_half_lives[ContextType.TASK] = 24.0  # 1 day
+
+    async def score(
+        self,
+        context: BaseContext,
+        query: str,
+        **kwargs: Any,
+    ) -> float:
+        """
+        Score context based on recency.
+
+        Args:
+            context: Context to score
+            query: Query (not used for recency, kept for interface)
+            **kwargs: Additional parameters
+                - reference_time: Time to measure recency from (default: now)
+
+        Returns:
+            Recency score between 0.0 and 1.0
+        """
+        reference_time = kwargs.get("reference_time")
+        if reference_time is None:
+            reference_time = datetime.now(UTC)
+        elif reference_time.tzinfo is None:
+            reference_time = reference_time.replace(tzinfo=UTC)
+
+        # Ensure context timestamp is timezone-aware
+        context_time = context.timestamp
+        if context_time.tzinfo is None:
+            context_time = context_time.replace(tzinfo=UTC)
+
+        # Calculate age in hours
+        age = reference_time - context_time
+        age_hours = max(0, age.total_seconds() / 3600)
+
+        # Get half-life for this context type
+        context_type = context.get_type()
+        half_life = self._type_half_lives.get(context_type, self._half_life_hours)
+
+        # Exponential decay
+        decay_factor = math.exp(-math.log(2) * age_hours / half_life)
+
+        return self.normalize_score(decay_factor)
+
+    def get_half_life(self, context_type: ContextType) -> float:
+        """
+        Get half-life for a context type.
+
+        Args:
+            context_type: Context type to get half-life for
+
+        Returns:
+            Half-life in hours
+        """
+        return self._type_half_lives.get(context_type, self._half_life_hours)
+
+    def set_half_life(self, context_type: ContextType, hours: float) -> None:
+        """
+        Set half-life for a context type.
+
+        Args:
+            context_type: Context type to set half-life for
+            hours: Half-life in hours
+        """
+        if hours <= 0:
+            raise ValueError("Half-life must be positive")
+        self._type_half_lives[context_type] = hours
+
+    async def score_batch(
+        self,
+        contexts: list[BaseContext],
+        query: str,
+        **kwargs: Any,
+    ) -> list[float]:
+        """
+        Score multiple contexts.
+
+        Args:
+            contexts: Contexts to score
+            query: Query (not used)
+            **kwargs: Additional parameters
+
+        Returns:
+            List of scores (same order as input)
+        """
+        scores = []
+        for context in contexts:
+            score = await self.score(context, query, **kwargs)
+            scores.append(score)
+        return scores