feat(ai): ADR-036 NVIDIA Nemotron Tool Calling 整合

Phase 20 - 提升 Tool Calling 精準度 50% → 83.3%

新增:
- src/models/nvidia.py: Pydantic Schema
- src/services/nvidia_provider.py: NvidiaProvider 類別
- tests/test_nvidia_provider.py: 15 項單元測試 (全部通過)

修改:
- ai_router.py: AIProvider.NVIDIA + route_tool_calling()
- ai_rate_limiter.py: NVIDIA 限制 (5 RPM, 100/day)
- models.json: NVIDIA 配置
- cd.yaml: Secrets 注入 NVIDIA_API_KEY

路由策略:
- Tool Calling: Nemotron → Gemini → Claude
- 一般對話: Ollama → Gemini → Claude (不變)

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

apps/api/src/models/__init__.py

@@ -6,6 +6,7 @@ AWOOOI Models Package
 - Approval: 簽核相關模型 (Phase 2 HITL)
 - Incident: 事件相關模型 (Phase 6 認知覺醒)
 - AI: AI 相關模型
+- NVIDIA: Nemotron Tool Calling 模型 (ADR-036)
 """
 
 # Approval Models (Phase 2)
@@ -39,6 +40,16 @@ from src.models.incident import (
     Signal,
 )
 
+# NVIDIA Models (ADR-036 - Nemotron Tool Calling)
+from src.models.nvidia import (
+    NvidiaProviderResult,
+    NvidiaResponse,
+    NvidiaUsage,
+    ToolCall,
+    ToolCallValidationResult,
+    ToolDefinition,
+)
+
 __all__ = [
     # Approval
     "ApprovalRequest",
@@ -65,4 +76,11 @@ __all__ = [
     "IncidentUpdate",
     "Severity",
     "Signal",
+    # NVIDIA (ADR-036)
+    "NvidiaProviderResult",
+    "NvidiaResponse",
+    "NvidiaUsage",
+    "ToolCall",
+    "ToolCallValidationResult",
+    "ToolDefinition",
 ]