feat(canvas): consolidate multi-image generation into a single image node with enhanced metadata structure

docs/content/docs/development/canvas-data-structure.mdx

@@ -48,9 +48,9 @@ type CanvasConnection = {
 };
 ```
 
-Node positions and sizes use canvas-world coordinates. Edges store only node IDs; their paths are calculated while rendering. Deleting a node removes connected edges, and deleting an image-batch root also removes its child nodes.
+Node positions and sizes use canvas-world coordinates. Edges store only node IDs; their paths are calculated while rendering. Deleting a node removes its connected edges.
 
-Common metadata includes content, prompt, status, generation mode, model, size, count, natural dimensions, batch relationships, input order, `storageKey`, MIME type, and byte size. Image and video `content` values are normally session-scoped `blob:` URLs; `storageKey` is the persistent identity.
+Common metadata includes content, prompt, status, generation mode, model, size, count, natural dimensions, image results, input order, `storageKey`, MIME type, and byte size. A multi-image generation remains one image node: `images` stores every result and `primaryImageId` selects the result mirrored by the node-level `content` and image metadata. Image and video `content` values are normally session-scoped `blob:` URLs; `storageKey` is the persistent identity.
 
 ## Assistant sessions
 

docs/content/docs/development/canvas-data-structure.zh-CN.mdx

@@ -77,6 +77,16 @@ type CanvasNodeData = {
 `metadata` 当前常用字段：
 
 ```ts
+type CanvasNodeImage = {
+  id: string;
+  content: string;
+  storageKey: string;
+  naturalWidth: number;
+  naturalHeight: number;
+  bytes: number;
+  mimeType: string;
+};
+
 type CanvasNodeMetadata = {
   content?: string;
   prompt?: string;
@@ -90,11 +100,8 @@ type CanvasNodeMetadata = {
   naturalWidth?: number;
   naturalHeight?: number;
   freeResize?: boolean;
-  isBatchRoot?: boolean;
-  batchRootId?: string;
-  batchChildIds?: string[];
+  images?: CanvasNodeImage[];
   primaryImageId?: string;
-  imageBatchExpanded?: boolean;
   inputOrder?: string[];
   storageKey?: string;
   mimeType?: string;
@@ -108,7 +115,7 @@ type CanvasNodeMetadata = {
 - 视频节点：`content` 是当前可播放的视频 URL，通常是 `blob:` URL；`storageKey` 指向本地视频 Blob；`bytes/mimeType` 保存文件信息。
 - 文本节点：`content` 保存文本内容；`fontSize` 保存字体大小；`prompt/status/errorDetails` 保存生成状态。
 - 生成配置节点：`generationMode/model/size/count/inputOrder` 保存生成配置；`generationMode` 可选择文本、图片或视频；上游输入通过 `connections` 计算。
-- 图片组节点：根节点用 `isBatchRoot/batchChildIds/primaryImageId/imageBatchExpanded` 记录批量生成结果；子图节点用 `batchRootId` 指回根节点。
+- 多图图片节点：批量生成仍只保存为一个图片节点，`images` 记录全部结果，`primaryImageId` 指定由节点级 `content` 和图片元数据展示的主图；展开状态仅属于当前界面，不写入项目数据。
 
 ## 连线结构
 

docs/content/docs/progress/pending-test.mdx

@@ -10,6 +10,7 @@ The current release needs manual verification in these areas:
 - English and Simplified Chinese switching across navigation, settings, titles, descriptions, Ant Design components, and persisted preferences.
 - Global Dropdown, Menu, Select, Cascader, and TreeSelect popup backgrounds, hover states, and selected states in both light and dark themes.
 - Canvas node resize stability, prompt scrolling, generated-prompt restoration, image editing, drag-and-drop references, and generation configuration.
+- Multi-image canvas generation: generating N images should create one image node with exactly N stored results and one incoming connection; the collapsed count control should remain clearly legible over any image, hide the floating toolbar when clicked, and expand the N images upward and to the right; every secondary image should keep a clearly legible primary-image action visible in both themes, and clicking the empty canvas should collapse the group.
 - Canvas Agent startup, MCP status, model and permission controls, approvals, version reporting, diagnostics, streaming responses, history consistency, multi-tab isolation, and local Skill management.
 - Agent header tabs in English and Chinese at different panel widths; labels should collapse to icons and counts before either edge is clipped.
 - Agent message layout, Markdown, code blocks, attachments, token statistics, progress timelines, command groups, and scroll-follow behavior.

docs/content/docs/progress/pending-test.zh-CN.mdx

@@ -48,6 +48,7 @@ description: 当前版本已实现但仍需人工验证的变更项
 - 画布节点提示词：图片等节点下方的提示词超过输入区域高度后，鼠标悬停在输入区域内滚动应查看提示词内容，不应缩放画布。
 - 画布节点提示词回显：单图或多图生成完成后，切换选择生成结果的根节点及任一子节点，下方输入区域都应继续显示生成时的提示词；切换到其他节点再选回来后也不应清空。
 - 画布生成配置：连接文本节点后连续点击生成（包含失败重试），组装提示词不应重复追加上游文本；在生图、文本、视频、音频之间切换时，实际请求应始终使用当前面板展示的对应类型模型。
+- 画布多图生成：生成 N 张图片后应只创建一个包含 N 张结果的图片节点，并且来源到结果只保留一条连线；收起状态下右上角应显示醒目的「N 张」数量提示，在任意图片背景上都清晰可辨；点击数量后，浮动工具条应暂时隐藏，主图与其余图片应从当前节点向右上方整齐展开且总数严格为 N；其余图片的「设为主图」应常驻显示，并在亮色、深色主题及不同图片背景上保持清晰；点击画布空白处应自动收起并恢复工具条。
 - 画布左侧元素列表：点击元素整行应平滑定位并选中对应节点；有内容的图片元素应显示预览按钮，点击后打开大图弹窗且不触发画布定位。
 - 配置与用户偏好：导出 JSON 后应包含渠道、默认模型、生成偏好、提示词来源和 WebDAV 配置；在修改当前配置后重新导入该文件，应恢复导出时的设置，错误 JSON 文件应提示格式不正确。配置文件包含 API Key 和 WebDAV 凭据，不应公开分享。
 - 模型渠道协议：渠道编辑可选择「火山方舟」并自动填入方舟接口地址；任意名称的生图模型应按方舟 JSON 格式提交参考图，任意名称的视频模型应按方舟任务格式提交和查询，不再依赖模型名包含 `doubao`、`seedream` 或 `seedance`；1080p 不应再因模型名包含 `fast` 被禁用，参考视频应允许最大 200MB、总像素 409600-8295044，并继续校验官方宽高、比例和时长限制。