feat: add new day-centric results API endpoint

Implements new /results endpoint with day-centric data structure:
- Returns starting_position, daily_metrics, trades, and final_position
- Supports reasoning levels: none (default), summary, full
- Uses database helper methods from trading_days schema
- Replaces old positions-based endpoint

Changes:
- Created api/routes/results_v2.py with new endpoint
- Registered router in api/main.py
- Removed old /results endpoint (positions table)
- Added comprehensive integration tests

All tests pass.

api/routes/__init__.py

@@ -0,0 +1 @@
+"""API routes package."""

api/routes/results_v2.py

@@ -0,0 +1,112 @@
+"""New results API with day-centric structure."""
+
+from fastapi import APIRouter, Query, Depends
+from typing import Optional, Literal
+import json
+
+from api.database import Database
+
+router = APIRouter()
+
+
+def get_database() -> Database:
+    """Dependency for database instance."""
+    return Database()
+
+
+@router.get("/results")
+async def get_results(
+    job_id: Optional[str] = None,
+    model: Optional[str] = None,
+    date: Optional[str] = None,
+    reasoning: Literal["none", "summary", "full"] = "none",
+    db: Database = Depends(get_database)
+):
+    """Get trading results grouped by day.
+
+    Args:
+        job_id: Filter by simulation job ID
+        model: Filter by model signature
+        date: Filter by trading date (YYYY-MM-DD)
+        reasoning: Include reasoning logs (none/summary/full)
+        db: Database instance (injected)
+
+    Returns:
+        JSON with day-centric trading results and performance metrics
+    """
+
+    # Build query with filters
+    query = "SELECT * FROM trading_days WHERE 1=1"
+    params = []
+
+    if job_id:
+        query += " AND job_id = ?"
+        params.append(job_id)
+
+    if model:
+        query += " AND model = ?"
+        params.append(model)
+
+    if date:
+        query += " AND date = ?"
+        params.append(date)
+
+    query += " ORDER BY date ASC, model ASC"
+
+    # Execute query
+    cursor = db.connection.execute(query, params)
+
+    # Format results
+    formatted_results = []
+
+    for row in cursor.fetchall():
+        trading_day_id = row[0]
+
+        # Build response object
+        day_data = {
+            "date": row[3],
+            "model": row[2],
+            "job_id": row[1],
+
+            "starting_position": {
+                "holdings": db.get_starting_holdings(trading_day_id),
+                "cash": row[4],  # starting_cash
+                "portfolio_value": row[5]  # starting_portfolio_value
+            },
+
+            "daily_metrics": {
+                "profit": row[6],  # daily_profit
+                "return_pct": row[7],  # daily_return_pct
+                "days_since_last_trading": row[14] if len(row) > 14 else 1
+            },
+
+            "trades": db.get_actions(trading_day_id),
+
+            "final_position": {
+                "holdings": db.get_ending_holdings(trading_day_id),
+                "cash": row[8],  # ending_cash
+                "portfolio_value": row[9]  # ending_portfolio_value
+            },
+
+            "metadata": {
+                "total_actions": row[12] if row[12] is not None else 0,
+                "session_duration_seconds": row[13],
+                "completed_at": row[16] if len(row) > 16 else None
+            }
+        }
+
+        # Add reasoning if requested
+        if reasoning == "summary":
+            day_data["reasoning"] = row[10]  # reasoning_summary
+        elif reasoning == "full":
+            reasoning_full = row[11]  # reasoning_full
+            day_data["reasoning"] = json.loads(reasoning_full) if reasoning_full else []
+        else:
+            day_data["reasoning"] = None
+
+        formatted_results.append(day_data)
+
+    return {
+        "count": len(formatted_results),
+        "results": formatted_results
+    }