新增角色管理、世界观设定、章节管理和AI工具调用功能

- 角色管理：CRUD API + 卡片网格UI，每个角色约200字描述
- 世界观设定：单篇长文编辑器，建议500字以上
- 章节管理：CRUD + 排序 API，左右分栏编辑器，单章5000字上限
- AI工具调用：支持工具审批流程和结构化工具调用
- 小说详情页新增章节、角色、世界观入口按钮

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend/app/schemas.py

@@ -60,6 +60,82 @@ class OutlineReorder(BaseModel):
     ordered_ids: list[int]
 
 
+# ── 角色 ──
+
+
+class CharacterCreate(BaseModel):
+    name: str = Field(min_length=1, max_length=100)
+    description: str = Field(default="", max_length=1000)
+
+
+class CharacterUpdate(BaseModel):
+    name: str | None = Field(default=None, min_length=1, max_length=100)
+    description: str | None = Field(default=None, max_length=1000)
+
+
+class CharacterRead(BaseModel):
+    id: int
+    novel_id: int
+    name: str
+    description: str
+    created_at: datetime
+    updated_at: datetime
+
+
+class CharacterList(BaseModel):
+    items: list[CharacterRead]
+    total: int
+
+
+# ── 章节 ──
+
+
+class ChapterCreate(BaseModel):
+    title: str = Field(min_length=1, max_length=200)
+    content: str = Field(default="", max_length=5000)
+
+
+class ChapterUpdate(BaseModel):
+    title: str | None = Field(default=None, min_length=1, max_length=200)
+    content: str | None = Field(default=None, max_length=5000)
+    sort_order: int | None = None
+
+
+class ChapterRead(BaseModel):
+    id: int
+    novel_id: int
+    sort_order: int
+    title: str
+    content: str
+    created_at: datetime
+    updated_at: datetime
+
+
+class ChapterList(BaseModel):
+    items: list[ChapterRead]
+    total: int
+
+
+class ChapterReorder(BaseModel):
+    """排序请求：按顺序传入章节id列表"""
+    ordered_ids: list[int]
+
+
+# ── 世界观 ──
+
+
+class WorldSettingUpdate(BaseModel):
+    content: str = Field(default="", max_length=5000)
+
+
+class WorldSettingRead(BaseModel):
+    id: int
+    novel_id: int
+    content: str
+    created_at: datetime
+    updated_at: datetime
+
+
 # ── AI配置 ──
 
 
@@ -87,9 +163,21 @@ class ChatMessageInput(BaseModel):
     content: str
 
 
+class PendingToolCall(BaseModel):
+    """待审批的工具调用"""
+    id: str
+    name: str
+    arguments: dict
+
+
 class ChatRequest(BaseModel):
     messages: list[ChatMessageInput]
     novel_id: int | None = None
+    page_context: str | None = None      # 用户当前所在页面，如 "大纲页" "世界观页"
+    tools_enabled: bool = True            # 是否启用工具调用
+    # 工具审批续传
+    pending_tool_calls: list[PendingToolCall] | None = None
+    assistant_text: str | None = None     # LLM 在工具调用前输出的文本
 
 
 class ChatMessageRead(BaseModel):